rolling_out.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "f2498938-a0d3-485d-a83f-699ed6d8f8a8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "importing Jupyter notebook from rolling_data_AR.ipynb\n",
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import datetime as dt\n",
    "import pathlib\n",
    "import re\n",
    "import json\n",
    "import math\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import scipy as sp\n",
    "from statsmodels.tsa.ar_model import AutoReg, ar_select_order\n",
    "import statsmodels.api as sm\n",
    "from functools import partial\n",
    "\n",
    "%load_ext autoreload\n",
    "%autoreload 1\n",
    "from utils.hourly_volas_and_dt import hourly_volas_and_dt\n",
    "%aimport utils.hourly_volas_and_dt\n",
    "\n",
    "from utils.rolling import rolling\n",
    "%aimport utils.rolling\n",
    "\n",
    "from utils.train_test_split import train_test_split\n",
    "%aimport utils.train_test_split\n",
    "\n",
    "from utils.loss_functions import mean_squared, quasi_likelihood\n",
    "%aimport utils.loss_functions\n",
    "\n",
    "from utils.r_squared_daily import r_squared_daily\n",
    "%aimport utils.r_squared_daily\n",
    "\n",
    "from utils.mean_squared_daily import mean_squared_daily\n",
    "%aimport utils.mean_squared_daily\n",
    "\n",
    "from utils.q_likelihood_daily import q_likelihood_daily\n",
    "%aimport utils.q_likelihood_daily\n",
    "\n",
    "from itertools import product\n",
    "\n",
    "from utils.mean_squared_oos import mean_squared_oos\n",
    "%aimport utils.mean_squared_oos\n",
    "from utils.q_likelihood_oos import q_likelihood_oos\n",
    "%aimport utils.q_likelihood_oos\n",
    "from utils.r_squared_oos import r_squared_oos\n",
    "%aimport utils.r_squared_oos\n",
    "from utils.q_likelihood_oos_details import q_likelihood_oos_details\n",
    "%aimport utils.q_likelihood_oos_details\n",
    "from utils.squared_errors_oos import squared_errors_oos\n",
    "%aimport utils.squared_errors_oos\n",
    "\n",
    "import import_ipynb\n",
    "\n",
    "from rolling_data_AR import in_sample_AR, in_sample_HAR, in_sample_HAR_plus_I\n",
    "\n",
    "from statsmodels.regression.rolling import RollingOLS"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "id": "82e0ed57-d16f-4de2-aa3e-b55f69cb094d",
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils.diebold_mariano_test import diebold_mariano_variance_estimate, normalised_stat, normal_p_value, autocovariance\n",
    "%aimport utils.diebold_mariano_test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c3ffa2af-168c-404e-8de6-90d2a8802ae7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "7a66664d-8146-4ae9-986a-aa5d8ecc59b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "symbols_list_file_name = \"binance_BTC_from_2019_05_01_to_2022_04_30.json\"\n",
    "with open(symbols_list_file_name, \"r\") as f:\n",
    "    symbols_list = json.load(f)\n",
    "\n",
    "input_folder_name = \"full_returns_winsorised\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "a28ff886-57e1-456d-884c-251fed2ef8e1",
   "metadata": {},
   "outputs": [],
   "source": [
    "symbols_list.remove('scbtc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "ceae701f-3096-4859-88f4-026eff4af739",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_empty_oos_results():\n",
    "    model_names = ['AR(1)', 'AR(30)', 'HAR', 'HAR+I']\n",
    "    data_type_names = ['Calendar', 'Rolling']\n",
    "\n",
    "    summary_names = ['$R^2$', 'MSE', 'QLIKE']\n",
    "\n",
    "    column_iterable = [model_names, data_type_names]\n",
    "    column_iterable_name = [\"Model\", \"Data\"]\n",
    "    column_index = pd.MultiIndex.from_product(column_iterable, names=column_iterable_name)\n",
    "\n",
    "    row_index = summary_names\n",
    "\n",
    "    return pd.DataFrame(np.full([3, 8], np.nan), index=row_index, columns=column_index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "49997be3-8b34-4dbf-ae55-acc278f13ba4",
   "metadata": {},
   "outputs": [],
   "source": [
    "oos_results = make_empty_oos_results()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "dfc8ffc5-5f8e-4b79-954a-1ea999d05004",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_name_to_function = {\n",
    "    'AR(1)' : partial(in_sample_AR, lag_days=1),\n",
    "    'AR(30)' : partial(in_sample_AR, lag_days=30),\n",
    "    'HAR' : in_sample_HAR,\n",
    "    'HAR+I' : in_sample_HAR_plus_I\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "44542276-572b-4688-bf88-05577d514f7f",
   "metadata": {},
   "outputs": [],
   "source": [
    "oos_results_dict = dict()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "f8ea5819-2b86-41eb-8264-41da211404b6",
   "metadata": {},
   "outputs": [],
   "source": [
    "def daily_rolling_OLS(y, X, window_length, data_type):\n",
    "    if data_type == 'Calendar':\n",
    "        rols = RollingOLS(y, X, window=window_length)\n",
    "        rres = rols.fit()\n",
    "        beta = rres.params\n",
    "        beta = beta.shift(1)\n",
    "        return beta, y, X\n",
    "    if data_type == 'Rolling':\n",
    "        rols = RollingOLS(y, X, window=window_length * 24)\n",
    "        rres = rols.fit() # this does 23 wasteful computations a day. Consider optimising.\n",
    "        beta = rres.params\n",
    "        beta = beta.shift(1)\n",
    "        beta_daily = beta[beta.index.hour == 0]\n",
    "        y_daily = y[y.index.hour == 0]\n",
    "        X_daily = X[X.index.hour == 0]\n",
    "        return beta_daily, y_daily, X_daily"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "64441d07-6140-4305-9f0c-b0ac9599b104",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_names = ['AR(1)', 'AR(30)', 'HAR', 'HAR+I']\n",
    "data_type_names = ['Calendar', 'Rolling']\n",
    "summary_names = ['MSE', 'QLIKE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 413,
   "id": "28acfd1f-aacb-49e6-8e72-6ffb9a6d1215",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 414,
   "id": "698e68fa-916f-4580-8387-2de04e27b67f",
   "metadata": {},
   "outputs": [],
   "source": [
    "details_file_name = 'summary_details.csv'\n",
    "with open(details_file_name, 'wb') as f:\n",
    "    pickle.dump(summary_details_dict, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 415,
   "id": "305b4d84-e361-4837-af6e-ae239e09a528",
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import defaultdict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 423,
   "id": "e0d90c86-c36c-46c7-b0f1-362159c0110f",
   "metadata": {},
   "outputs": [],
   "source": [
    "from utils.prediction import prediction\n",
    "%aimport utils.prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 431,
   "id": "75b2c7ba-45e1-4e40-a39e-17eb5069faaa",
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions_file_name = 'predictions.csv'\n",
    "with open(predictions_file_name, 'wb') as f:\n",
    "    pickle.dump(predictions_dict, f)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b0602d6b-ec04-4061-ba4e-b08f3602654d",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 427,
   "id": "4943ab37-6c83-4d22-9e0f-91d8841724eb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with bnbbtc, which is index 0.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with adabtc, which is index 1.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with ethbtc, which is index 2.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with batbtc, which is index 3.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with btcusdc, which is index 4.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with btcusdt, which is index 5.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with neobtc, which is index 6.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with btctusd, which is index 7.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with etcbtc, which is index 8.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with ltcbtc, which is index 9.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with xrpbtc, which is index 10.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with trxbtc, which is index 11.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with eosbtc, which is index 12.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with xmrbtc, which is index 13.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with zecbtc, which is index 14.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with qtumbtc, which is index 15.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with dcrbtc, which is index 16.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with xlmbtc, which is index 17.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with iotabtc, which is index 18.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with manabtc, which is index 19.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with dashbtc, which is index 20.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with zrxbtc, which is index 21.\n",
      "Done with Calendar.\n",
      "Done with Rolling.\n",
      "Done with repbtc, which is index 22.\n"
     ]
    }
   ],
   "source": [
    "# store PREDICTIONS\n",
    "predictions_dict = dict()\n",
    "for i, symbol in enumerate(symbols_list):\n",
    "    predictions = dict()\n",
    "    for data_type in ['Calendar', 'Rolling']:\n",
    "        for model_name, function in model_name_to_function.items():\n",
    "            res, y, X = function(\n",
    "                input_folder_name=input_folder_name, \n",
    "                symbol=symbol, \n",
    "                data_type=data_type\n",
    "            )\n",
    "            window_length = 365\n",
    "            beta_daily, y_daily, X_daily = daily_rolling_OLS(y, X, window_length, data_type)\n",
    "            \n",
    "            predictions[model_name, data_type] = prediction(beta_daily, y_daily, X_daily, window_length)\n",
    "            \n",
    "        print(f\"Done with {data_type}.\")\n",
    "    # oos_results_dict[symbol] = oos_results\n",
    "    predictions_dict[symbol] = predictions\n",
    "    print(f\"Done with {symbol}, which is index {i}.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 430,
   "id": "9559d68b-58f2-4c93-a9fe-f5d367a5fc3a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{('AR(1)',\n",
       "  'Calendar'): array([0.00170336, 0.00110998, 0.00052763, 0.00068782, 0.00087208,\n",
       "        0.00070739, 0.00078719, 0.00132661, 0.00089619, 0.00094663,\n",
       "        0.00167391, 0.00137309, 0.00121911, 0.00082128, 0.00122187,\n",
       "        0.00103388, 0.00069923, 0.00070798, 0.00076059, 0.00078623,\n",
       "        0.00076977, 0.00082948, 0.00067523, 0.00055797, 0.00064167,\n",
       "        0.00069397, 0.000609  , 0.00060763, 0.00067386, 0.00067337,\n",
       "        0.0006672 , 0.00085803, 0.00082881, 0.0006886 , 0.00052686,\n",
       "        0.00053841, 0.00061109, 0.00049218, 0.00056107, 0.00046651,\n",
       "        0.00048115, 0.00053018, 0.00059681, 0.00048372, 0.00040103,\n",
       "        0.0003845 , 0.00059331, 0.00051073, 0.00055677, 0.00051876,\n",
       "        0.00049624, 0.00042182, 0.00045302, 0.000431  , 0.00041168,\n",
       "        0.00051788, 0.0004013 , 0.00045278, 0.00051291, 0.00047542,\n",
       "        0.00052869, 0.00041542, 0.00046475, 0.00052845, 0.00040613,\n",
       "        0.00043121, 0.00043552, 0.00052032, 0.00079742, 0.00094704,\n",
       "        0.00090547, 0.00078306, 0.00079513, 0.00109886, 0.00134425,\n",
       "        0.00109147, 0.00078564, 0.0008924 , 0.00081707, 0.00064477,\n",
       "        0.00097427, 0.00087875, 0.00073318, 0.00064201, 0.00074062,\n",
       "        0.00110819, 0.0008193 , 0.0010001 , 0.0019123 , 0.00183485,\n",
       "        0.001025  , 0.00072046, 0.00103898, 0.00098758, 0.00119423,\n",
       "        0.00115612, 0.00084515, 0.00112661, 0.00070302, 0.00088765,\n",
       "        0.00072891, 0.00065243, 0.00074742, 0.00093326, 0.00104472,\n",
       "        0.00099884, 0.00133025, 0.00109333, 0.00087425, 0.00089325,\n",
       "        0.00074211, 0.00071426, 0.00073117, 0.00077934, 0.00079517,\n",
       "        0.00061488, 0.00060586, 0.00072675, 0.00074145, 0.0008077 ,\n",
       "        0.00068661, 0.00064631, 0.00061948, 0.00061659, 0.00118632,\n",
       "        0.00198506, 0.00282062, 0.00230317, 0.00178213, 0.00325919,\n",
       "        0.00258165, 0.00242753, 0.00216424, 0.00205425, 0.00186809,\n",
       "        0.00284248, 0.00512513, 0.00417031, 0.00435472, 0.00329814,\n",
       "        0.00242964, 0.00276142, 0.0018381 , 0.00193288, 0.0032236 ,\n",
       "        0.00230603, 0.00208942, 0.00141668, 0.00127247, 0.00129419,\n",
       "        0.00130159, 0.00136471, 0.00140359, 0.00119536, 0.00154069,\n",
       "        0.00185931, 0.0010144 , 0.0009243 , 0.00092086, 0.00111762,\n",
       "        0.00084128, 0.00095674, 0.0007738 , 0.00088034, 0.00063074,\n",
       "        0.00151001, 0.00112051, 0.00072873, 0.00073237, 0.00099955,\n",
       "        0.00052401, 0.00056291, 0.00061374, 0.00089978, 0.00106269,\n",
       "        0.0009763 , 0.00069197, 0.00055629, 0.00060359, 0.0007311 ,\n",
       "        0.00112844, 0.0007409 , 0.00070003, 0.00132092, 0.00090213,\n",
       "        0.00061434, 0.00084692, 0.00129692, 0.00107532, 0.00155794,\n",
       "        0.00139764, 0.00119587, 0.00079853, 0.00083176, 0.00067883,\n",
       "        0.00073706, 0.00090197, 0.0007816 , 0.00058232, 0.00059148,\n",
       "        0.00074969, 0.00104678, 0.00127975, 0.00084168, 0.00102116,\n",
       "        0.00126022, 0.00107619, 0.00106568, 0.00368241, 0.00252825,\n",
       "        0.00234669, 0.00109724, 0.00091738, 0.00082984, 0.00128451,\n",
       "        0.0015038 , 0.00099783, 0.00075129, 0.00080049, 0.00066772,\n",
       "        0.00060018, 0.00056528, 0.00072754, 0.00076561, 0.00060999,\n",
       "        0.00061966, 0.00056302, 0.00067904, 0.00088522, 0.0006714 ,\n",
       "        0.00103099, 0.00203908, 0.00090766, 0.00179197, 0.00219145,\n",
       "        0.00185097, 0.00152411, 0.00181787, 0.00118335, 0.00105244,\n",
       "        0.00144585, 0.00228855, 0.00180169, 0.0026126 , 0.00180244,\n",
       "        0.00151393, 0.00145152, 0.00247978, 0.00323339, 0.00326864,\n",
       "        0.00251679, 0.00278686, 0.0031418 , 0.00244111, 0.00161301,\n",
       "        0.00197939, 0.00395286, 0.00209476, 0.00201281, 0.00161703,\n",
       "        0.00166331, 0.00171332, 0.00201866, 0.00159882, 0.0023103 ,\n",
       "        0.00153887, 0.00185605, 0.00204496, 0.00116997, 0.00098286,\n",
       "        0.00159084, 0.00142272, 0.00151561, 0.00150008, 0.00405657,\n",
       "        0.00203103, 0.00143148, 0.00339318, 0.00275774, 0.00179164,\n",
       "        0.00237882, 0.00293238, 0.00530964, 0.00332261, 0.00469091,\n",
       "        0.00760023, 0.01035873, 0.00384628, 0.00449371, 0.00227011,\n",
       "        0.00297987, 0.00271769, 0.00200809, 0.00558505, 0.00575981,\n",
       "        0.01141202, 0.00888264, 0.00757336, 0.00665227, 0.00744192,\n",
       "        0.00728267, 0.0039533 , 0.00377935, 0.00184919, 0.00253367,\n",
       "        0.00396198, 0.00212863, 0.00139877, 0.00125521, 0.00141578,\n",
       "        0.00090338, 0.0013429 , 0.00097523, 0.00370092, 0.00286892,\n",
       "        0.00263118, 0.00234217, 0.00157158, 0.00091108, 0.00123177,\n",
       "        0.00110785, 0.00088928, 0.00118622, 0.00085014, 0.00090229,\n",
       "        0.00079215, 0.00102762, 0.00078288, 0.00104643, 0.00121462,\n",
       "        0.00081477, 0.00116362, 0.00125572, 0.00082337, 0.00213524,\n",
       "        0.00195051, 0.00190764, 0.00245719, 0.00187235, 0.00146792,\n",
       "        0.00195516, 0.00285561, 0.0036832 , 0.00199485, 0.0019894 ,\n",
       "        0.00317573, 0.0021953 , 0.00599068, 0.00475278, 0.00315465,\n",
       "        0.00113012, 0.00303311, 0.00187859, 0.00352872, 0.00375558,\n",
       "        0.00401867, 0.00386183, 0.00353761, 0.00522405, 0.00195792,\n",
       "        0.00124935, 0.00150399, 0.00169578, 0.00125672, 0.0022038 ,\n",
       "        0.00187336, 0.00126759, 0.00099007, 0.00152327, 0.00183546,\n",
       "        0.00131817, 0.00093556, 0.00096019, 0.0012893 , 0.00144366,\n",
       "        0.00184545, 0.00153522, 0.0013749 , 0.00277239, 0.00149371,\n",
       "        0.00109662, 0.00125388, 0.00194073, 0.00127481, 0.00568492,\n",
       "        0.00504547, 0.00455921, 0.00373839, 0.00512419, 0.00494755,\n",
       "        0.00513446, 0.00324552, 0.00236901, 0.00272446, 0.00192531,\n",
       "        0.001952  , 0.00186663, 0.00157536, 0.00216941, 0.0017332 ,\n",
       "        0.00252407, 0.00168117, 0.00090332, 0.00135181, 0.00213023,\n",
       "        0.00157614, 0.00111482, 0.00114924, 0.0011185 , 0.00102388,\n",
       "        0.00099262, 0.00089276, 0.00086473, 0.00075692, 0.00082339,\n",
       "        0.00072384, 0.00085526, 0.00177085, 0.00342822, 0.0019951 ,\n",
       "        0.00141567, 0.00139186, 0.00117054, 0.0009649 , 0.00106038,\n",
       "        0.0012235 , 0.00110699, 0.0009118 , 0.00080698, 0.00077573,\n",
       "        0.00083947, 0.00082319, 0.00103152, 0.0010646 , 0.00094151,\n",
       "        0.00071351, 0.00069911, 0.0007015 , 0.00085329, 0.00071061,\n",
       "        0.00104572, 0.00126177, 0.0009335 , 0.00084081, 0.00072736,\n",
       "        0.00088313, 0.00134612, 0.00120204, 0.00089208, 0.00101806,\n",
       "        0.00077983, 0.00077447, 0.00135528, 0.00098644, 0.00101542,\n",
       "        0.00081687, 0.00093804, 0.00086381, 0.00088897, 0.00067932,\n",
       "        0.00070246, 0.0006435 , 0.00090176, 0.00095996, 0.00077335,\n",
       "        0.00077165, 0.00086195, 0.00097908, 0.00100603, 0.00110556,\n",
       "        0.0008683 , 0.00082313, 0.00070493, 0.00092561, 0.00097962,\n",
       "        0.00114433, 0.00095829, 0.00094424, 0.00084674, 0.00066657,\n",
       "        0.00154823, 0.00111977, 0.00120675, 0.00130828, 0.00093626,\n",
       "        0.00059173, 0.0006547 , 0.00070909, 0.00096324, 0.00081273,\n",
       "        0.00088496, 0.00079075, 0.00088924, 0.00069955, 0.0007802 ,\n",
       "        0.00217021, 0.00151152, 0.00111366, 0.00106533, 0.00069832,\n",
       "        0.0007134 , 0.0010078 , 0.00067826, 0.00077321, 0.00069463,\n",
       "        0.00073209, 0.00060117, 0.00054963, 0.00113908, 0.00101573,\n",
       "        0.00077266, 0.00069446, 0.00094809, 0.00066176, 0.00103695,\n",
       "        0.00070798, 0.00066403, 0.00102828, 0.00081746, 0.00091712,\n",
       "        0.0007993 , 0.00079299, 0.00080319, 0.00078396, 0.00101501,\n",
       "        0.00089289, 0.00099621, 0.00063249, 0.00082361, 0.00080435,\n",
       "        0.00145091, 0.00226913, 0.0012412 , 0.00171597, 0.00101707,\n",
       "        0.00086474, 0.00095854, 0.00092879, 0.00085549, 0.0010239 ,\n",
       "        0.00078177, 0.00059568, 0.00056226, 0.00061018, 0.00062298,\n",
       "        0.00093035, 0.00107605, 0.0010505 , 0.00082544, 0.00081194,\n",
       "        0.00095786, 0.00095131, 0.00090923, 0.00065516, 0.00118281,\n",
       "        0.00093846, 0.00116994, 0.00118698, 0.00075407, 0.00133482,\n",
       "        0.00078866, 0.00063436, 0.0007517 , 0.00072245, 0.00064461,\n",
       "        0.00123427, 0.00082485, 0.0009412 , 0.00082512, 0.00065239,\n",
       "        0.00059042, 0.00068111, 0.0007812 , 0.00066845, 0.0010288 ,\n",
       "        0.00160481, 0.00083787, 0.00108001, 0.00075224, 0.00082816,\n",
       "        0.00083021, 0.0007373 , 0.00074589, 0.0013884 , 0.00078965,\n",
       "        0.00109852, 0.00069439, 0.00074097, 0.00063216, 0.00072682,\n",
       "        0.00068753, 0.0005566 , 0.00067892, 0.00071962, 0.0007075 ,\n",
       "        0.00055483, 0.00068726, 0.00051398, 0.00050741, 0.00048291,\n",
       "        0.00054211, 0.00053005, 0.00052542, 0.00050101, 0.00040639,\n",
       "        0.00045266, 0.00050096, 0.00053375, 0.00050271, 0.00044678,\n",
       "        0.0004752 , 0.00045567, 0.00040095, 0.00044881, 0.00049585,\n",
       "        0.00070105, 0.00055213, 0.00056886, 0.00063747, 0.00042499,\n",
       "        0.00063588, 0.00078799, 0.00054961, 0.00056114, 0.00064186,\n",
       "        0.00049715, 0.00048377, 0.00056733, 0.00055103, 0.00051307,\n",
       "        0.00082264, 0.00115134, 0.00185694, 0.00101856, 0.00130176,\n",
       "        0.00078062, 0.00091635, 0.00101613, 0.00076149, 0.00060383,\n",
       "        0.00054884, 0.00065589, 0.00051308, 0.00045242, 0.00043145,\n",
       "        0.00064816, 0.00075224, 0.00047105, 0.00051865, 0.00061061,\n",
       "        0.00051942, 0.00069509, 0.00055261, 0.00041375, 0.00034845,\n",
       "        0.00038201, 0.00095372, 0.00046223, 0.00044005, 0.00039778,\n",
       "        0.00038016, 0.00044584, 0.00062514, 0.00065236, 0.00056059,\n",
       "        0.00109634, 0.00059167, 0.00053519, 0.00056906, 0.00075443,\n",
       "        0.001062  , 0.00058203, 0.00049457, 0.00056637, 0.00051163,\n",
       "        0.00043633, 0.00078671, 0.0005581 , 0.00077399, 0.00066437,\n",
       "        0.00049207, 0.00037109, 0.00037943, 0.00045169, 0.00037682,\n",
       "        0.00071597, 0.00041208, 0.00041589, 0.00036787, 0.0003232 ,\n",
       "        0.0003294 , 0.0004709 , 0.00033544, 0.00035799, 0.00030643,\n",
       "        0.00025821, 0.0004562 , 0.0004255 , 0.00036291, 0.00045569,\n",
       "        0.00050691, 0.00040969, 0.00040268, 0.00039781, 0.00033633,\n",
       "        0.00040999, 0.00041496, 0.00041525, 0.00038113, 0.00030353,\n",
       "        0.00029869, 0.00038331, 0.00030658, 0.00040952, 0.00026804,\n",
       "        0.00023773, 0.00020222, 0.00020863, 0.00037368, 0.00027748,\n",
       "        0.0002922 , 0.00034886, 0.00027445, 0.0002709 , 0.00025592,\n",
       "        0.00037981, 0.0002912 , 0.00026945, 0.00047205, 0.00029985]),\n",
       " ('AR(30)',\n",
       "  'Calendar'): array([ 4.04975297e-04,  7.06082635e-04,  7.77161028e-04,  6.66704326e-04,\n",
       "         6.07602629e-04,  5.55944590e-04,  5.35108031e-04,  4.88552006e-04,\n",
       "         4.41429529e-04,  4.61818271e-04,  3.76722717e-04,  3.36759306e-04,\n",
       "         4.55512951e-04,  4.74677405e-04,  3.31778897e-04,  3.25353021e-04,\n",
       "         3.13176055e-04,  4.16912523e-04,  3.10597313e-04,  3.97031495e-04,\n",
       "         3.62059356e-04,  3.46107436e-04,  3.32455339e-04,  4.05361249e-04,\n",
       "         2.89818011e-04,  2.41546546e-04,  4.19124426e-04,  2.54389229e-04,\n",
       "         3.28915190e-04,  3.57883871e-04,  3.30239999e-04,  3.92028685e-04,\n",
       "         3.25101317e-04,  3.33373097e-04,  3.72051443e-04,  2.33220036e-04,\n",
       "         3.27942622e-04,  3.26910034e-04,  3.82675726e-04,  5.90195887e-04,\n",
       "         8.00587208e-04,  7.55052801e-04,  7.44768177e-04,  7.84453056e-04,\n",
       "         1.02906836e-03,  1.26113708e-03,  1.05830189e-03,  8.35136865e-04,\n",
       "         9.35716023e-04,  7.40573540e-04,  6.53075579e-04,  1.08504604e-03,\n",
       "         6.97201801e-04,  6.32823159e-04,  6.99119411e-04,  6.69580440e-04,\n",
       "         9.40086760e-04,  7.12193085e-04,  1.00885751e-03,  1.85896534e-03,\n",
       "         1.55048911e-03,  1.01580100e-03,  1.06571322e-03,  1.05897159e-03,\n",
       "         8.81723418e-04,  1.52501304e-03,  1.17507585e-03,  6.56688471e-04,\n",
       "         1.17332873e-03,  7.23946940e-04,  1.11377966e-03,  8.66960795e-04,\n",
       "         7.06769958e-04,  6.06740316e-04,  8.92581983e-04,  8.40988183e-04,\n",
       "         9.74345246e-04,  1.33673311e-03,  9.36933237e-04,  1.12028267e-03,\n",
       "         1.10684047e-03,  6.38031467e-04,  6.56053285e-04,  9.44853773e-04,\n",
       "         9.83744787e-04,  8.02275392e-04,  6.35523080e-04,  5.80080091e-04,\n",
       "         7.48742060e-04,  7.23725703e-04,  7.97480582e-04,  7.20437204e-04,\n",
       "         5.88439892e-04,  5.84340089e-04,  5.92315336e-04,  1.14585255e-03,\n",
       "         1.85126187e-03,  2.64721438e-03,  2.23253039e-03,  2.07533431e-03,\n",
       "         3.70248717e-03,  2.45073024e-03,  3.04833277e-03,  2.69307926e-03,\n",
       "         2.08985056e-03,  1.75379159e-03,  3.26021354e-03,  4.97421104e-03,\n",
       "         3.98834620e-03,  5.18924095e-03,  3.62439172e-03,  2.83481384e-03,\n",
       "         3.28331750e-03,  2.24311850e-03,  2.01829479e-03,  3.65176732e-03,\n",
       "         2.32296672e-03,  2.58067262e-03,  1.86577949e-03,  1.16289079e-03,\n",
       "         1.37046889e-03,  1.92091795e-03,  9.91361269e-04,  1.48111764e-03,\n",
       "         1.07001380e-03,  1.76169856e-03,  2.09919267e-03,  9.85425587e-04,\n",
       "         1.10825675e-03,  7.05108322e-04,  1.10482450e-03,  1.33675769e-03,\n",
       "         1.57142425e-03,  9.18736535e-04,  1.00075622e-03,  6.00956031e-04,\n",
       "         1.69269076e-03,  7.75028214e-04,  9.07229049e-04,  8.27798941e-04,\n",
       "         1.18970769e-03,  2.86262539e-04,  8.42223070e-04,  5.03349573e-04,\n",
       "         8.16890479e-04,  9.19781090e-04,  1.06120515e-03,  5.89405994e-04,\n",
       "         6.43366028e-04,  5.91190342e-04,  8.39007473e-04,  1.03580235e-03,\n",
       "         5.73440052e-04,  7.53964886e-04,  1.31725675e-03,  6.53595709e-04,\n",
       "         7.39923646e-04,  9.35885688e-04,  1.13200213e-03,  8.35057913e-04,\n",
       "         1.87799646e-03,  1.19938848e-03,  1.21379855e-03,  8.65456798e-04,\n",
       "         1.16047688e-03,  5.79664261e-04,  8.57644335e-04,  8.62576665e-04,\n",
       "         8.37005294e-04,  5.97381130e-04,  6.51011364e-04,  5.87117837e-04,\n",
       "         8.74607089e-04,  1.05262413e-03,  7.10056988e-04,  1.04845380e-03,\n",
       "         1.05479591e-03,  9.55705639e-04,  1.25178181e-03,  3.68377361e-03,\n",
       "         1.70549997e-03,  2.91556436e-03,  1.50662759e-03,  1.36418665e-03,\n",
       "         9.32499066e-04,  2.12039222e-03,  1.54038252e-03,  1.19243464e-03,\n",
       "         8.20312888e-04,  7.88659619e-04,  5.55883872e-04,  6.68269060e-04,\n",
       "         4.13637543e-04,  5.67066396e-04,  3.66019476e-04,  5.48198866e-04,\n",
       "         3.80465816e-04,  7.65676346e-04,  4.38296980e-04,  1.06126129e-03,\n",
       "         5.69844323e-04,  8.49795582e-04,  1.67573056e-03,  1.15193923e-03,\n",
       "         2.02330386e-03,  2.18625868e-03,  1.56667710e-03,  1.78293649e-03,\n",
       "         2.06811819e-03,  1.07067101e-03,  1.52022760e-03,  1.67949997e-03,\n",
       "         2.20333918e-03,  1.68914887e-03,  2.83594971e-03,  1.49008600e-03,\n",
       "         1.68929662e-03,  1.51145239e-03,  2.57263517e-03,  2.82599441e-03,\n",
       "         3.43809306e-03,  2.49288560e-03,  3.06368159e-03,  3.16360087e-03,\n",
       "         2.71772703e-03,  2.26953703e-03,  2.75601634e-03,  3.61078242e-03,\n",
       "         2.16764800e-03,  2.85075185e-03,  2.05338798e-03,  1.37255170e-03,\n",
       "         2.07676499e-03,  2.19534722e-03,  1.36380932e-03,  2.07405549e-03,\n",
       "         1.66441910e-03,  1.66111234e-03,  2.40661307e-03,  1.11058444e-03,\n",
       "         1.25135526e-03,  1.67631947e-03,  1.22288404e-03,  1.75744943e-03,\n",
       "         1.62300244e-03,  3.73342373e-03,  1.42749001e-03,  2.45762797e-03,\n",
       "         3.18824112e-03,  2.25425212e-03,  1.91690382e-03,  3.84625054e-03,\n",
       "         2.74705601e-03,  4.61746594e-03,  3.38210793e-03,  5.44314699e-03,\n",
       "         6.39644397e-03,  9.66039046e-03,  4.98553057e-03,  7.55225427e-03,\n",
       "         3.78053737e-03,  2.38747565e-03,  5.16012427e-03,  3.27114695e-03,\n",
       "         6.00441561e-03,  5.13483061e-03,  1.03547073e-02,  8.34698633e-03,\n",
       "         7.64931120e-03,  6.19040720e-03,  5.92434140e-03,  5.39782149e-03,\n",
       "         5.86807227e-03,  6.61480516e-03,  5.30004912e-03,  2.56399050e-03,\n",
       "         6.34152149e-03,  6.06504331e-03,  1.31622254e-03,  3.67733159e-03,\n",
       "         1.45856306e-03,  9.31019878e-04,  5.10853033e-04,  5.86700205e-04,\n",
       "         5.17786662e-03,  2.56233621e-03,  2.26599986e-03,  6.51545015e-04,\n",
       "         1.75219470e-03, -7.45166392e-04,  1.18556122e-03,  5.47836203e-04,\n",
       "         5.50425147e-04,  2.36750528e-03,  2.06371252e-03,  1.57294676e-03,\n",
       "         8.52383973e-04,  1.37946901e-03,  6.74010825e-04,  1.02715246e-03,\n",
       "         1.37000859e-03,  9.66970263e-04,  1.01015933e-03,  1.38709115e-03,\n",
       "         1.01906756e-03,  1.66525555e-03,  1.67837183e-03,  1.47537172e-03,\n",
       "         2.44731402e-03,  1.88299328e-03,  1.40329358e-03,  1.87907178e-03,\n",
       "         2.71877145e-03,  3.63049437e-03,  2.42920851e-03,  2.32536461e-03,\n",
       "         3.12634740e-03,  2.12338079e-03,  5.10596231e-03,  5.38942959e-03,\n",
       "         3.76450735e-03,  1.67351279e-03,  3.08282558e-03,  1.63579427e-03,\n",
       "         3.04987139e-03,  4.42895059e-03,  4.80800954e-03,  3.58673052e-03,\n",
       "         3.63691793e-03,  4.59735690e-03,  2.92647557e-03,  1.59861517e-03,\n",
       "         2.00605096e-03,  1.71820349e-03,  7.07581895e-04,  2.77040437e-03,\n",
       "         2.65437802e-03,  1.82402180e-03,  9.28972862e-04,  1.25674142e-03,\n",
       "         9.32169103e-04,  8.73742145e-04,  1.37282992e-03,  8.00645765e-04,\n",
       "         1.32238018e-03,  1.12506008e-03,  2.08245838e-03,  1.39609278e-03,\n",
       "         9.00363766e-04,  2.47067164e-03,  1.33701208e-03,  5.87251174e-04,\n",
       "         1.41272455e-03,  2.26759735e-03,  1.37494662e-03,  4.96936640e-03,\n",
       "         5.12274819e-03,  4.77482909e-03,  4.14134365e-03,  4.92650787e-03,\n",
       "         4.38599319e-03,  5.24215181e-03,  3.80646556e-03,  3.12154304e-03,\n",
       "         3.35594380e-03,  2.89624160e-03,  1.97421605e-03,  2.36122327e-03,\n",
       "         2.26322421e-03,  2.59151245e-03,  1.79847941e-03,  1.85359813e-03,\n",
       "         1.86370119e-03,  1.07962748e-03,  1.60367242e-03,  1.71582970e-03,\n",
       "         1.16112317e-03,  1.03343433e-03,  5.71536182e-04,  7.94756862e-04,\n",
       "         7.30414912e-04,  8.38958070e-04,  4.02159991e-04,  6.65387331e-04,\n",
       "         1.12305227e-03,  8.50979059e-04,  8.27240927e-04,  8.58677117e-04,\n",
       "         1.86090155e-03,  3.34912214e-03,  2.00536446e-03,  1.44888086e-03,\n",
       "         1.52705252e-03,  9.64275239e-04,  7.36579661e-04,  1.08181293e-03,\n",
       "         1.33067463e-03,  1.35597466e-03,  1.06412466e-03,  9.97373492e-04,\n",
       "         6.13799246e-04,  8.47246311e-04,  1.00084713e-03,  8.23009140e-04,\n",
       "         9.21812094e-04,  7.87617478e-04,  8.21465998e-04,  6.98927240e-04,\n",
       "         6.13908142e-04,  7.73693345e-04,  4.69450829e-04,  6.26975642e-04,\n",
       "         8.26328609e-04,  8.53078841e-04,  9.21045928e-04,  6.99241320e-04,\n",
       "         8.50719409e-04,  1.34129279e-03,  1.12449210e-03,  8.41218498e-04,\n",
       "         1.04239507e-03,  8.48793493e-04,  7.19478752e-04,  1.17502464e-03,\n",
       "         9.99974960e-04,  9.53041039e-04,  8.75440694e-04,  7.98342039e-04,\n",
       "         7.37788526e-04,  8.54927316e-04,  7.32964846e-04,  6.84834949e-04,\n",
       "         5.78992591e-04,  7.84152632e-04,  7.19529394e-04,  6.98416209e-04,\n",
       "         7.63916063e-04,  8.38870610e-04,  8.31351048e-04,  8.01293628e-04,\n",
       "         9.53862545e-04,  8.08470784e-04,  8.13002167e-04,  6.83323498e-04,\n",
       "         7.89497238e-04,  8.84490173e-04,  1.02369744e-03,  9.47437398e-04,\n",
       "         9.34188261e-04,  7.78053991e-04,  6.20111010e-04,  1.32123037e-03,\n",
       "         1.14709154e-03,  1.10749941e-03,  1.38355962e-03,  9.00461803e-04,\n",
       "         5.18271805e-04,  6.16275589e-04,  5.85856264e-04,  9.07227790e-04,\n",
       "         7.40819927e-04,  8.93508948e-04,  6.60614235e-04,  7.65646746e-04,\n",
       "         7.25321295e-04,  5.71571313e-04,  1.86358722e-03,  1.44839112e-03,\n",
       "         9.72342928e-04,  1.12455403e-03,  5.73645623e-04,  6.22712639e-04,\n",
       "         9.08779733e-04,  6.16683066e-04,  7.23818285e-04,  6.65871709e-04,\n",
       "         7.84859818e-04,  3.14519647e-04,  5.37451310e-04,  1.23755956e-03,\n",
       "         8.76826010e-04,  7.11436830e-04,  5.67036388e-04,  6.98685343e-04,\n",
       "         6.14185234e-04,  8.31323525e-04,  7.60795321e-04,  5.57394781e-04,\n",
       "         8.26056525e-04,  6.21590860e-04,  5.59084389e-04,  8.86232475e-04,\n",
       "         7.67226691e-04,  8.72140797e-04,  7.46443809e-04,  8.09788682e-04,\n",
       "         8.37105001e-04,  8.90302283e-04,  7.45197423e-04,  6.38708613e-04,\n",
       "         7.88762615e-04,  1.14571578e-03,  1.91341150e-03,  1.32072899e-03,\n",
       "         1.55754079e-03,  1.31092180e-03,  5.83854750e-04,  9.29008626e-04,\n",
       "         7.93304821e-04,  7.70249375e-04,  1.20685940e-03,  7.29211167e-04,\n",
       "         7.22824854e-04,  3.77637916e-04,  6.92849902e-04,  5.30325876e-04,\n",
       "         6.81990236e-04,  1.00861217e-03,  6.99737921e-04,  8.50702506e-04,\n",
       "         7.29880201e-04,  8.60129270e-04,  8.91744613e-04,  6.68302981e-04,\n",
       "         4.93246905e-04,  7.55378851e-04,  8.74015700e-04,  1.03397878e-03,\n",
       "         1.19489114e-03,  9.63855354e-04,  1.22883002e-03,  8.72319603e-04,\n",
       "         4.21110564e-04,  7.21982114e-04,  6.36258951e-04,  7.19784313e-04,\n",
       "         1.12142907e-03,  8.10051093e-04,  9.03551453e-04,  6.65415546e-04,\n",
       "         5.99069522e-04,  3.27112664e-04,  5.50727395e-04,  7.11415640e-04,\n",
       "         5.08503947e-04,  9.37327346e-04,  1.45144705e-03,  8.14208130e-04,\n",
       "         9.62355978e-04,  7.80990527e-04,  5.58723537e-04,  6.72383849e-04,\n",
       "         6.05544813e-04,  6.97031170e-04,  1.28567618e-03,  9.99298777e-04,\n",
       "         1.07078108e-03,  6.40349043e-04,  7.26739544e-04,  4.73639769e-04,\n",
       "         4.76429285e-04,  6.52345312e-04,  4.29157830e-04,  7.38260874e-04,\n",
       "         7.68930162e-04,  5.07604011e-04,  5.75278066e-04,  4.84565262e-04,\n",
       "         3.18378376e-04,  2.84785637e-04,  2.49272762e-04,  5.21264115e-04,\n",
       "         3.54824506e-04,  5.99643466e-04,  4.25456792e-04,  2.97218283e-04,\n",
       "         3.04484005e-04,  2.80842478e-04,  3.25608853e-04,  4.11609378e-04,\n",
       "         3.56115698e-04,  4.36722155e-04,  3.61343952e-04,  2.80697423e-04,\n",
       "         3.32044189e-04,  3.64824056e-04,  5.32504362e-04,  4.39671562e-04,\n",
       "         4.19372303e-04,  5.31185325e-04,  3.34133511e-04,  4.76469385e-04,\n",
       "         6.67625783e-04,  4.26492876e-04,  4.92648658e-04,  5.52733742e-04,\n",
       "         4.12305340e-04,  3.52580248e-04,  4.69029683e-04,  4.89779115e-04,\n",
       "         3.62892602e-04,  7.14685341e-04,  9.59358361e-04,  1.51682485e-03,\n",
       "         1.13232389e-03,  1.18560585e-03,  8.77940224e-04,  6.78608818e-04,\n",
       "         8.07945377e-04,  6.48237001e-04,  5.38413203e-04,  8.06398055e-04,\n",
       "         5.84798305e-04,  6.09897889e-04,  2.80648266e-04,  4.51853418e-04,\n",
       "         6.15149172e-04,  5.80492879e-04,  5.09673116e-04,  2.61063826e-04,\n",
       "         6.37746361e-04,  5.10296892e-04,  5.68117811e-04,  4.34417639e-04,\n",
       "         2.67452961e-04,  2.33022325e-04,  1.22852838e-04,  6.91702437e-04,\n",
       "         5.72540942e-04,  4.39053135e-04,  4.87965480e-04,  3.20753406e-04,\n",
       "         4.78224272e-04,  4.74893661e-04,  4.98133365e-04,  6.15960048e-04,\n",
       "         9.66988276e-04,  6.81944545e-04,  4.28346025e-04,  5.30597025e-04,\n",
       "         5.93469121e-04,  9.79466592e-04,  6.20537029e-04,  3.54353497e-04,\n",
       "         5.21963210e-04,  4.20452863e-04,  6.73091215e-04,  3.91809850e-04,\n",
       "         5.02044051e-04,  7.65320883e-04,  6.05618290e-04,  3.99822619e-04,\n",
       "         2.71222866e-04,  3.26812672e-04,  4.91502233e-04,  4.04005546e-04,\n",
       "         5.01212443e-04,  2.79796334e-04,  2.16868526e-04,  4.39196407e-04,\n",
       "         1.57574787e-04,  2.86465733e-04,  3.07730124e-04,  2.44274294e-04,\n",
       "         2.30717891e-04,  9.89772484e-05,  2.91949929e-04,  3.09461669e-04,\n",
       "         4.22979475e-04,  3.50553440e-04,  2.54132249e-04,  3.45013711e-04,\n",
       "         3.39460088e-04,  2.21296466e-04,  3.49690106e-04,  3.42116243e-04,\n",
       "         3.80849359e-04,  4.04206617e-04,  2.66111674e-04,  2.52849786e-04,\n",
       "         2.02057743e-04,  2.56433312e-04,  2.69432040e-04,  2.60291878e-04,\n",
       "         2.98889380e-04,  1.83265988e-04,  1.62277372e-04,  1.45212532e-04,\n",
       "         6.60575640e-05,  2.90627546e-04,  1.88178046e-04,  1.60559884e-04,\n",
       "         2.62989932e-04,  1.52662183e-04,  1.81091170e-04,  1.48573982e-04,\n",
       "         2.69701642e-04,  2.54720228e-04,  1.92978402e-04,  3.61618009e-04,\n",
       "         2.05770458e-04]),\n",
       " ('HAR',\n",
       "  'Calendar'): array([0.00057934, 0.00057365, 0.00074258, 0.00071902, 0.00060876,\n",
       "        0.00047338, 0.00047574, 0.00053256, 0.00042215, 0.00046495,\n",
       "        0.00037118, 0.00036777, 0.00040005, 0.00045678, 0.00035886,\n",
       "        0.0002928 , 0.00027002, 0.00044673, 0.00038129, 0.00041717,\n",
       "        0.00037834, 0.00036831, 0.0003067 , 0.00033995, 0.00031522,\n",
       "        0.00029359, 0.00037911, 0.00029336, 0.00032841, 0.00037958,\n",
       "        0.00034974, 0.00039035, 0.00030116, 0.00034218, 0.00039682,\n",
       "        0.00029592, 0.00031055, 0.00030576, 0.00036969, 0.00061847,\n",
       "        0.0007742 , 0.00077402, 0.00069332, 0.0007389 , 0.00101293,\n",
       "        0.00125427, 0.00106709, 0.00080412, 0.00090475, 0.00085556,\n",
       "        0.0006864 , 0.00097519, 0.00086802, 0.00073303, 0.00064235,\n",
       "        0.0007199 , 0.00104505, 0.000802  , 0.00095837, 0.0017833 ,\n",
       "        0.0017695 , 0.00109031, 0.00082999, 0.001109  , 0.00107243,\n",
       "        0.00125986, 0.00120197, 0.0008999 , 0.00114614, 0.00078333,\n",
       "        0.00093262, 0.00078606, 0.00069835, 0.00076187, 0.00092321,\n",
       "        0.00101311, 0.00098792, 0.00129029, 0.00110792, 0.00093353,\n",
       "        0.00095698, 0.0008192 , 0.00077974, 0.00078184, 0.00079919,\n",
       "        0.00080012, 0.00063313, 0.00061156, 0.00071073, 0.00071987,\n",
       "        0.00077441, 0.00066554, 0.00062523, 0.00060099, 0.00059762,\n",
       "        0.00110112, 0.00184536, 0.00267901, 0.00228684, 0.00186489,\n",
       "        0.00328768, 0.00277925, 0.002707  , 0.00248558, 0.00235514,\n",
       "        0.00217055, 0.00307202, 0.00506042, 0.00438103, 0.00463784,\n",
       "        0.00382162, 0.00309063, 0.00340856, 0.00250381, 0.0024082 ,\n",
       "        0.00346109, 0.00260675, 0.0023702 , 0.0017387 , 0.00152519,\n",
       "        0.00150383, 0.00146816, 0.00142337, 0.00141392, 0.00118724,\n",
       "        0.00150528, 0.00183029, 0.00109669, 0.000997  , 0.00096714,\n",
       "        0.00111953, 0.00087138, 0.00093626, 0.00073005, 0.00081388,\n",
       "        0.00059312, 0.00135918, 0.00104755, 0.00073054, 0.00072528,\n",
       "        0.00095513, 0.00055771, 0.00058879, 0.00058592, 0.00080496,\n",
       "        0.00095669, 0.00090302, 0.00066321, 0.00055796, 0.00059852,\n",
       "        0.00070616, 0.00102915, 0.00071097, 0.00066587, 0.00118469,\n",
       "        0.00087924, 0.00066237, 0.00084856, 0.00120124, 0.00105173,\n",
       "        0.00146965, 0.00135116, 0.00121345, 0.00091958, 0.00094355,\n",
       "        0.00078818, 0.00081407, 0.00090292, 0.00077455, 0.00058615,\n",
       "        0.0005819 , 0.00069825, 0.00094671, 0.0011564 , 0.00081623,\n",
       "        0.00096781, 0.00118896, 0.00107576, 0.00108481, 0.00323295,\n",
       "        0.00242974, 0.00238124, 0.00142603, 0.00126147, 0.00117429,\n",
       "        0.00154001, 0.00157874, 0.0010988 , 0.000815  , 0.00083533,\n",
       "        0.00071985, 0.00065567, 0.00058765, 0.00066741, 0.00068333,\n",
       "        0.00055522, 0.00055207, 0.00050319, 0.00059767, 0.00077553,\n",
       "        0.0006062 , 0.00089852, 0.00175964, 0.00090352, 0.00165525,\n",
       "        0.00205372, 0.00184572, 0.00164067, 0.00191277, 0.00137558,\n",
       "        0.00128291, 0.00157022, 0.00222515, 0.00184517, 0.00253135,\n",
       "        0.00190532, 0.00170164, 0.00167455, 0.00252557, 0.0031557 ,\n",
       "        0.00327056, 0.00267918, 0.00294104, 0.00330839, 0.00282186,\n",
       "        0.00210545, 0.00230762, 0.00384945, 0.00242524, 0.00230796,\n",
       "        0.00190972, 0.00188476, 0.00193085, 0.00215637, 0.00168943,\n",
       "        0.00223728, 0.00162938, 0.0018803 , 0.00204752, 0.00135903,\n",
       "        0.00114255, 0.00159406, 0.00141204, 0.00148237, 0.00145202,\n",
       "        0.00350523, 0.00205851, 0.00164851, 0.00320314, 0.0028488 ,\n",
       "        0.00217313, 0.00266071, 0.00297947, 0.00491024, 0.00364902,\n",
       "        0.00470636, 0.00713156, 0.00993554, 0.00517946, 0.00573269,\n",
       "        0.00364128, 0.00411011, 0.00372417, 0.00282456, 0.0054041 ,\n",
       "        0.00569102, 0.01056366, 0.00898359, 0.00816597, 0.0076023 ,\n",
       "        0.00848705, 0.00841339, 0.00548317, 0.00499591, 0.00297594,\n",
       "        0.00340348, 0.00461535, 0.00280647, 0.0019091 , 0.00163907,\n",
       "        0.0016788 , 0.00113048, 0.00147674, 0.0010104 , 0.00359762,\n",
       "        0.00290143, 0.0027244 , 0.00247863, 0.00183339, 0.00116687,\n",
       "        0.00144968, 0.00126762, 0.00100922, 0.0012282 , 0.00088167,\n",
       "        0.00091278, 0.0008049 , 0.00100352, 0.00077567, 0.00100887,\n",
       "        0.00115481, 0.00080157, 0.00111164, 0.00120299, 0.00082073,\n",
       "        0.00199796, 0.00185905, 0.00183925, 0.00236064, 0.00186816,\n",
       "        0.00152032, 0.00197379, 0.0027786 , 0.00354803, 0.002075  ,\n",
       "        0.0020573 , 0.00312503, 0.00229012, 0.00568632, 0.00467863,\n",
       "        0.00328773, 0.0015035 , 0.00319229, 0.00216062, 0.00361824,\n",
       "        0.00374173, 0.00394267, 0.00383357, 0.00363552, 0.00513372,\n",
       "        0.00238609, 0.00169184, 0.00183062, 0.00191783, 0.00145722,\n",
       "        0.00222547, 0.00183865, 0.00129742, 0.00104898, 0.00150493,\n",
       "        0.0017767 , 0.00133536, 0.00096641, 0.00095811, 0.00123913,\n",
       "        0.00138581, 0.00174124, 0.00146662, 0.00133268, 0.00258552,\n",
       "        0.00151294, 0.00116861, 0.0012959 , 0.00188202, 0.00130913,\n",
       "        0.00518125, 0.00473415, 0.00442485, 0.00380824, 0.00512179,\n",
       "        0.00507048, 0.0053595 , 0.00363997, 0.00277945, 0.0030315 ,\n",
       "        0.0022731 , 0.00220086, 0.00203759, 0.001683  , 0.00217085,\n",
       "        0.00177156, 0.00245632, 0.00171291, 0.00100215, 0.00137777,\n",
       "        0.00207293, 0.00157108, 0.00114892, 0.00113792, 0.00109365,\n",
       "        0.00101271, 0.00097315, 0.00084851, 0.00080196, 0.00069565,\n",
       "        0.00074671, 0.00065047, 0.00076306, 0.00158986, 0.0031304 ,\n",
       "        0.00189965, 0.0014147 , 0.00141195, 0.00123172, 0.00105502,\n",
       "        0.001118  , 0.00119777, 0.00107077, 0.00088529, 0.00077745,\n",
       "        0.00073987, 0.00079423, 0.00077417, 0.00095183, 0.00097925,\n",
       "        0.00087237, 0.00066987, 0.00065577, 0.00065437, 0.00078798,\n",
       "        0.0006521 , 0.00094742, 0.00114861, 0.00086617, 0.00078949,\n",
       "        0.0006919 , 0.00082871, 0.00125463, 0.00113437, 0.00085505,\n",
       "        0.0009695 , 0.00075729, 0.0007555 , 0.00127937, 0.00093927,\n",
       "        0.00095803, 0.00078517, 0.0008912 , 0.00082789, 0.0008537 ,\n",
       "        0.00064889, 0.00066082, 0.00059907, 0.00082749, 0.0008809 ,\n",
       "        0.00071427, 0.00070974, 0.00079456, 0.00090508, 0.00093874,\n",
       "        0.00103107, 0.00082028, 0.000781  , 0.00067566, 0.00087128,\n",
       "        0.00091843, 0.00106737, 0.0009007 , 0.00089101, 0.00080612,\n",
       "        0.00064707, 0.00143737, 0.00106599, 0.00114377, 0.00124271,\n",
       "        0.00091641, 0.00060705, 0.00066122, 0.00068356, 0.00090133,\n",
       "        0.00075563, 0.00080183, 0.00070918, 0.00080441, 0.00064277,\n",
       "        0.00070936, 0.00195473, 0.00140352, 0.00106655, 0.00103366,\n",
       "        0.00071101, 0.00072003, 0.00097357, 0.0006454 , 0.00070537,\n",
       "        0.00062659, 0.00065028, 0.00053461, 0.00048576, 0.00099865,\n",
       "        0.00090627, 0.0006956 , 0.00062791, 0.00085312, 0.00060699,\n",
       "        0.0009455 , 0.00064893, 0.00060069, 0.00092248, 0.0007437 ,\n",
       "        0.00082962, 0.00073311, 0.00071868, 0.00072975, 0.00071644,\n",
       "        0.00091562, 0.00081102, 0.00090199, 0.00058272, 0.00074771,\n",
       "        0.00073042, 0.00130859, 0.00205351, 0.00117501, 0.0016057 ,\n",
       "        0.0010142 , 0.00088303, 0.00096794, 0.00092747, 0.00082412,\n",
       "        0.00096281, 0.000728  , 0.00055587, 0.00051829, 0.00054993,\n",
       "        0.00055199, 0.0008188 , 0.00094737, 0.00093304, 0.00074423,\n",
       "        0.00073988, 0.00087607, 0.00087945, 0.00084185, 0.00061116,\n",
       "        0.00106907, 0.00086154, 0.0010715 , 0.00109417, 0.00071575,\n",
       "        0.00122897, 0.0007618 , 0.00061157, 0.0007068 , 0.00066831,\n",
       "        0.00058582, 0.00110721, 0.00073976, 0.00084343, 0.0007495 ,\n",
       "        0.00059728, 0.0005394 , 0.00061858, 0.00069245, 0.00058658,\n",
       "        0.00090237, 0.00142121, 0.00076766, 0.00099057, 0.00071205,\n",
       "        0.00077868, 0.00078458, 0.00069724, 0.00068125, 0.00124934,\n",
       "        0.00072712, 0.00100209, 0.00065152, 0.00068816, 0.00059153,\n",
       "        0.00067117, 0.00061791, 0.00049909, 0.00059216, 0.00062749,\n",
       "        0.00061652, 0.0004822 , 0.00059255, 0.00044076, 0.0004325 ,\n",
       "        0.00040498, 0.00045114, 0.00043646, 0.00043147, 0.00040635,\n",
       "        0.00032235, 0.00035898, 0.00039922, 0.00042711, 0.00039794,\n",
       "        0.000349  , 0.00037286, 0.00035766, 0.00030811, 0.00034418,\n",
       "        0.00038202, 0.00056339, 0.0004383 , 0.00045418, 0.00051883,\n",
       "        0.00033888, 0.0005243 , 0.00066335, 0.00045716, 0.00046767,\n",
       "        0.00053937, 0.00041194, 0.00040207, 0.00047343, 0.00045386,\n",
       "        0.00042192, 0.00069547, 0.00099397, 0.00164104, 0.00092917,\n",
       "        0.00119273, 0.00075086, 0.00087799, 0.00096996, 0.00074157,\n",
       "        0.00058116, 0.00051965, 0.00059908, 0.0004716 , 0.00040918,\n",
       "        0.00037889, 0.00056781, 0.00066375, 0.00041518, 0.00045706,\n",
       "        0.00054114, 0.00045895, 0.00060607, 0.00050169, 0.00037345,\n",
       "        0.00031364, 0.00034201, 0.00085289, 0.0004189 , 0.00040359,\n",
       "        0.00036325, 0.00034763, 0.00041242, 0.0005687 , 0.00058337,\n",
       "        0.00051064, 0.00098033, 0.00055296, 0.00050839, 0.00053543,\n",
       "        0.0006862 , 0.00096076, 0.00056811, 0.0004689 , 0.00052137,\n",
       "        0.00047807, 0.00040926, 0.00068024, 0.00047959, 0.00067494,\n",
       "        0.00059993, 0.00046735, 0.0003696 , 0.00036956, 0.00040399,\n",
       "        0.00033975, 0.00059127, 0.00034771, 0.00034777, 0.00030943,\n",
       "        0.00027297, 0.00027115, 0.00038202, 0.0002626 , 0.00027703,\n",
       "        0.00023354, 0.00018992, 0.00034427, 0.00032575, 0.00027394,\n",
       "        0.00034952, 0.00039303, 0.00032482, 0.00032602, 0.00032002,\n",
       "        0.00027046, 0.00032818, 0.00032941, 0.00032427, 0.00029914,\n",
       "        0.00023632, 0.00022806, 0.00029547, 0.00023164, 0.00031386,\n",
       "        0.00020274, 0.00017412, 0.00014609, 0.00014873, 0.00028904,\n",
       "        0.00020905, 0.00021953, 0.00026827, 0.0002077 , 0.00020875,\n",
       "        0.00019887, 0.00029982, 0.000226  , 0.0002075 , 0.0003746 ,\n",
       "        0.00023954]),\n",
       " ('HAR+I',\n",
       "  'Calendar'): array([0.00055012, 0.0005351 , 0.00067776, 0.00095371, 0.00062843,\n",
       "        0.00044948, 0.00052196, 0.00057007, 0.00040113, 0.00052711,\n",
       "        0.00032148, 0.00032848, 0.0005431 , 0.00052008, 0.00031198,\n",
       "        0.00027696, 0.00026858, 0.00036007, 0.00030316, 0.00032395,\n",
       "        0.00044109, 0.00026315, 0.0002675 , 0.00022549, 0.00036158,\n",
       "        0.00026316, 0.00043855, 0.00025557, 0.00030104, 0.00051618,\n",
       "        0.00035185, 0.00040379, 0.00027631, 0.00031083, 0.00038161,\n",
       "        0.00023655, 0.00023406, 0.00029132, 0.00040508, 0.00070886,\n",
       "        0.00062538, 0.0006266 , 0.00071871, 0.00068915, 0.00091702,\n",
       "        0.00181208, 0.00086022, 0.00086683, 0.00088253, 0.00072594,\n",
       "        0.00063355, 0.0009223 , 0.00089009, 0.00070771, 0.00076559,\n",
       "        0.00066354, 0.00130694, 0.00073094, 0.00093889, 0.00242788,\n",
       "        0.00164637, 0.00113034, 0.00071366, 0.00100736, 0.00117598,\n",
       "        0.000671  , 0.0013916 , 0.00075948, 0.00087749, 0.00081275,\n",
       "        0.00121592, 0.0008548 , 0.00070224, 0.00060006, 0.0012555 ,\n",
       "        0.00090871, 0.00110412, 0.00144966, 0.00092722, 0.00081795,\n",
       "        0.00104239, 0.00085751, 0.00064415, 0.00093985, 0.00088749,\n",
       "        0.00057186, 0.00059677, 0.00054642, 0.00081232, 0.00060762,\n",
       "        0.00096899, 0.00062567, 0.00068094, 0.00062596, 0.00059553,\n",
       "        0.00163316, 0.00185923, 0.00352154, 0.00189593, 0.00254943,\n",
       "        0.00341339, 0.00247534, 0.00299851, 0.00207077, 0.00268064,\n",
       "        0.002013  , 0.00254152, 0.00518113, 0.00500547, 0.00398356,\n",
       "        0.00341566, 0.00264149, 0.00334503, 0.00218064, 0.00319591,\n",
       "        0.00423141, 0.00217039, 0.00251515, 0.00184699, 0.00141783,\n",
       "        0.00172009, 0.0013074 , 0.00137795, 0.00150827, 0.00111345,\n",
       "        0.00208122, 0.00155513, 0.00101074, 0.00094706, 0.0008147 ,\n",
       "        0.00123161, 0.00071733, 0.00090797, 0.00063134, 0.00054769,\n",
       "        0.00057188, 0.00127074, 0.00087631, 0.00067545, 0.00081082,\n",
       "        0.00054654, 0.00040101, 0.00051574, 0.00050375, 0.00091558,\n",
       "        0.00108431, 0.0009538 , 0.00065667, 0.00057   , 0.00044502,\n",
       "        0.00078942, 0.00129134, 0.00060262, 0.00084668, 0.00097059,\n",
       "        0.00074081, 0.00063206, 0.00075588, 0.00062491, 0.00129082,\n",
       "        0.00181279, 0.00115821, 0.00156655, 0.00091075, 0.00100448,\n",
       "        0.0006312 , 0.00093989, 0.00100212, 0.00069963, 0.00058215,\n",
       "        0.00057372, 0.00080869, 0.00119116, 0.00092333, 0.00080308,\n",
       "        0.00113811, 0.00136047, 0.00090585, 0.00096562, 0.00271963,\n",
       "        0.00201294, 0.0019963 , 0.00144813, 0.00137321, 0.00104871,\n",
       "        0.00180039, 0.00184276, 0.00094604, 0.00070432, 0.00085298,\n",
       "        0.00055333, 0.0006267 , 0.00056536, 0.00070618, 0.00064363,\n",
       "        0.00059203, 0.00051734, 0.00048814, 0.00057669, 0.00078084,\n",
       "        0.00044932, 0.00124761, 0.0017815 , 0.00080049, 0.00191324,\n",
       "        0.00270324, 0.00146631, 0.00161544, 0.00201285, 0.0011576 ,\n",
       "        0.00145002, 0.00185493, 0.00245051, 0.00185311, 0.00169278,\n",
       "        0.0014953 , 0.00155715, 0.0017191 , 0.00402533, 0.00374591,\n",
       "        0.00280304, 0.00302712, 0.00311491, 0.0038935 , 0.00308533,\n",
       "        0.00187299, 0.00263564, 0.00443666, 0.00235301, 0.00238601,\n",
       "        0.00177565, 0.00211466, 0.00186286, 0.00259721, 0.00144399,\n",
       "        0.00295561, 0.0014791 , 0.00240561, 0.00164602, 0.00124535,\n",
       "        0.00106921, 0.00176898, 0.00143513, 0.00182524, 0.00144846,\n",
       "        0.004285  , 0.00163744, 0.00157395, 0.00412899, 0.00257477,\n",
       "        0.00205164, 0.00276764, 0.00399112, 0.00431388, 0.00335768,\n",
       "        0.0056361 , 0.00908278, 0.00967908, 0.00551223, 0.00606218,\n",
       "        0.00355017, 0.00399082, 0.00247806, 0.00266447, 0.00967475,\n",
       "        0.00433901, 0.01023154, 0.00817966, 0.00744175, 0.00794942,\n",
       "        0.00784647, 0.0082213 , 0.00508289, 0.00449849, 0.00280124,\n",
       "        0.0035579 , 0.00443077, 0.00285643, 0.00155936, 0.00168207,\n",
       "        0.00142402, 0.00098329, 0.00138227, 0.00080831, 0.00402492,\n",
       "        0.00197116, 0.00284838, 0.00283197, 0.00168873, 0.00111199,\n",
       "        0.00138456, 0.00116215, 0.00116701, 0.0010846 , 0.00077401,\n",
       "        0.00085759, 0.00071644, 0.00120451, 0.00070596, 0.00118297,\n",
       "        0.00112371, 0.00075728, 0.00126733, 0.00108537, 0.00076574,\n",
       "        0.00221594, 0.00158575, 0.0021192 , 0.00219609, 0.00225728,\n",
       "        0.00155178, 0.0020304 , 0.00293563, 0.00329978, 0.00203965,\n",
       "        0.00216851, 0.00270788, 0.00279322, 0.00522071, 0.00394997,\n",
       "        0.00333891, 0.00156551, 0.00285295, 0.00203882, 0.00300976,\n",
       "        0.00339701, 0.00494288, 0.00366971, 0.00445427, 0.00416328,\n",
       "        0.00215316, 0.00218354, 0.00187558, 0.00190092, 0.0014723 ,\n",
       "        0.00203058, 0.00171274, 0.00104291, 0.00092824, 0.00159154,\n",
       "        0.00147   , 0.00127711, 0.00107554, 0.00094555, 0.00126763,\n",
       "        0.0012156 , 0.00227566, 0.00137141, 0.00161914, 0.00239271,\n",
       "        0.0014583 , 0.00131617, 0.00161021, 0.00162287, 0.00123885,\n",
       "        0.00517813, 0.00365719, 0.00590133, 0.00337644, 0.00548131,\n",
       "        0.00508143, 0.0047328 , 0.00343124, 0.00265244, 0.00292361,\n",
       "        0.00242362, 0.00221099, 0.00185946, 0.00153008, 0.00212666,\n",
       "        0.00156658, 0.00192304, 0.00160206, 0.00098636, 0.00178566,\n",
       "        0.00183014, 0.00156038, 0.00113382, 0.00113061, 0.0009397 ,\n",
       "        0.00130336, 0.00090959, 0.00085659, 0.0008685 , 0.00064447,\n",
       "        0.00073994, 0.00062047, 0.00075231, 0.00180871, 0.00300024,\n",
       "        0.00175699, 0.00122687, 0.00137208, 0.00115647, 0.0011752 ,\n",
       "        0.00106663, 0.00117943, 0.00110994, 0.00083792, 0.00073496,\n",
       "        0.00077747, 0.00087342, 0.00075028, 0.00087289, 0.00088305,\n",
       "        0.00086129, 0.00063451, 0.00061431, 0.00068599, 0.00080928,\n",
       "        0.00059299, 0.00081169, 0.00103434, 0.00078061, 0.00068446,\n",
       "        0.00068683, 0.00080197, 0.00103109, 0.00113711, 0.00080309,\n",
       "        0.00100849, 0.00076144, 0.00072213, 0.00133923, 0.0010515 ,\n",
       "        0.00101818, 0.00078497, 0.00111521, 0.00080814, 0.00083323,\n",
       "        0.00064115, 0.00064104, 0.00061491, 0.00093678, 0.00083235,\n",
       "        0.0006806 , 0.00072065, 0.00083483, 0.00089131, 0.00106065,\n",
       "        0.00096149, 0.00083424, 0.00072922, 0.00070736, 0.000891  ,\n",
       "        0.00104577, 0.00103997, 0.00088619, 0.00093045, 0.00076949,\n",
       "        0.00063517, 0.00132891, 0.0011306 , 0.001027  , 0.00105896,\n",
       "        0.00082601, 0.00059551, 0.00066587, 0.000775  , 0.00084604,\n",
       "        0.00073792, 0.00073955, 0.00062756, 0.00079857, 0.00069598,\n",
       "        0.00068114, 0.00210783, 0.00119389, 0.00107589, 0.00098762,\n",
       "        0.00069335, 0.00078911, 0.00088954, 0.00065181, 0.00066263,\n",
       "        0.00060623, 0.00064021, 0.00050923, 0.00052368, 0.0008963 ,\n",
       "        0.00102989, 0.00072722, 0.00058441, 0.00075909, 0.00057193,\n",
       "        0.00080466, 0.00062688, 0.00062776, 0.00078428, 0.00073991,\n",
       "        0.0007429 , 0.00076146, 0.00063765, 0.00073938, 0.00081227,\n",
       "        0.00091182, 0.0007496 , 0.00082681, 0.00053123, 0.00081242,\n",
       "        0.00068566, 0.00150734, 0.00177404, 0.0010797 , 0.00169153,\n",
       "        0.00094735, 0.0010565 , 0.001015  , 0.00099179, 0.00089188,\n",
       "        0.00086206, 0.00066467, 0.00054572, 0.00053553, 0.00053361,\n",
       "        0.00063466, 0.00074581, 0.00117031, 0.00093038, 0.00066872,\n",
       "        0.00066004, 0.00093187, 0.00081679, 0.00104672, 0.00059675,\n",
       "        0.00083757, 0.00079583, 0.00092015, 0.00096778, 0.00072159,\n",
       "        0.00179306, 0.00072764, 0.00063007, 0.00075376, 0.0006453 ,\n",
       "        0.00057008, 0.00097508, 0.00065858, 0.00090064, 0.00063577,\n",
       "        0.00063833, 0.00052728, 0.00062021, 0.00065985, 0.00054877,\n",
       "        0.00075933, 0.00102187, 0.00081483, 0.00099401, 0.00072223,\n",
       "        0.00081644, 0.00075031, 0.00064906, 0.00079816, 0.00091364,\n",
       "        0.00073061, 0.00098665, 0.00059859, 0.00067899, 0.00058615,\n",
       "        0.00067889, 0.00056883, 0.00049887, 0.00068052, 0.00062662,\n",
       "        0.00074868, 0.00048124, 0.00064252, 0.00039797, 0.00040187,\n",
       "        0.00039184, 0.00041252, 0.00040656, 0.00049689, 0.00042442,\n",
       "        0.00032334, 0.00036942, 0.00044311, 0.00037231, 0.00042003,\n",
       "        0.00032387, 0.00034798, 0.00032729, 0.00030446, 0.00039133,\n",
       "        0.00036823, 0.00091967, 0.00039128, 0.0003746 , 0.00082574,\n",
       "        0.00032901, 0.00051432, 0.00064652, 0.00045731, 0.00048498,\n",
       "        0.00060848, 0.00042344, 0.00035281, 0.00043011, 0.00045055,\n",
       "        0.00044598, 0.00070067, 0.00139398, 0.00138038, 0.00100843,\n",
       "        0.00125738, 0.00085158, 0.00116471, 0.00141472, 0.0008905 ,\n",
       "        0.00067912, 0.00057207, 0.0007203 , 0.00044953, 0.00046934,\n",
       "        0.00037679, 0.00058167, 0.00049913, 0.00040842, 0.00041488,\n",
       "        0.00057186, 0.00045834, 0.00057272, 0.00057618, 0.00036655,\n",
       "        0.00030882, 0.00034582, 0.00078097, 0.00041069, 0.00042515,\n",
       "        0.00032843, 0.00030747, 0.00035012, 0.00064602, 0.00062946,\n",
       "        0.00048177, 0.00096028, 0.00052801, 0.00043974, 0.00060636,\n",
       "        0.00083528, 0.00068411, 0.00051311, 0.00046994, 0.0005487 ,\n",
       "        0.00045837, 0.00038109, 0.00099197, 0.00042746, 0.00057726,\n",
       "        0.00056188, 0.00042466, 0.00031923, 0.00037639, 0.00043406,\n",
       "        0.00034362, 0.0005575 , 0.00031896, 0.00033363, 0.00030136,\n",
       "        0.00023074, 0.0002381 , 0.00027463, 0.00023042, 0.00023268,\n",
       "        0.00018755, 0.0001625 , 0.00054428, 0.00029583, 0.0002471 ,\n",
       "        0.00040648, 0.00035252, 0.00027938, 0.00031009, 0.00035424,\n",
       "        0.00026358, 0.00029627, 0.00031828, 0.00027798, 0.00030076,\n",
       "        0.00020534, 0.0002288 , 0.00028417, 0.0002168 , 0.00031517,\n",
       "        0.00018127, 0.00015919, 0.00013624, 0.00014604, 0.00030171,\n",
       "        0.00014329, 0.00016968, 0.00029558, 0.00019117, 0.00017497,\n",
       "        0.00019965, 0.00030377, 0.00025561, 0.00019136, 0.00033923,\n",
       "        0.00022602]),\n",
       " ('AR(1)',\n",
       "  'Rolling'): array([0.0016985 , 0.00110951, 0.0005333 , 0.00069307, 0.00087402,\n",
       "        0.00071118, 0.00078805, 0.00132706, 0.00089547, 0.00094915,\n",
       "        0.00167112, 0.00136684, 0.00121968, 0.00082428, 0.00121888,\n",
       "        0.00103363, 0.00070041, 0.00070959, 0.00075965, 0.00078616,\n",
       "        0.00076856, 0.00082784, 0.00067295, 0.00055548, 0.00064014,\n",
       "        0.00069524, 0.00060919, 0.00060835, 0.00067468, 0.00067335,\n",
       "        0.00066769, 0.00085864, 0.00082913, 0.00068833, 0.0005281 ,\n",
       "        0.00053978, 0.00061249, 0.0004939 , 0.00056248, 0.00046788,\n",
       "        0.00048305, 0.00052971, 0.00059847, 0.0004835 , 0.00040421,\n",
       "        0.00039233, 0.0005977 , 0.00051606, 0.00056156, 0.0005213 ,\n",
       "        0.000501  , 0.00042752, 0.00045885, 0.00043753, 0.00041831,\n",
       "        0.00052036, 0.00041944, 0.00046059, 0.00051115, 0.00047613,\n",
       "        0.00052966, 0.00041828, 0.00046931, 0.00052993, 0.00040783,\n",
       "        0.00043142, 0.00043657, 0.00052144, 0.00079698, 0.00094529,\n",
       "        0.00090409, 0.00078171, 0.00079734, 0.00109818, 0.00133907,\n",
       "        0.00108738, 0.00078606, 0.00088915, 0.00081773, 0.0006474 ,\n",
       "        0.00097315, 0.00087737, 0.00073479, 0.00064552, 0.00074321,\n",
       "        0.00110428, 0.00082019, 0.00099983, 0.00190377, 0.00181101,\n",
       "        0.00102119, 0.00072344, 0.00103616, 0.00098709, 0.00118865,\n",
       "        0.00115009, 0.00084588, 0.0011203 , 0.00070502, 0.00088654,\n",
       "        0.00073125, 0.00065568, 0.00074891, 0.00093336, 0.00104173,\n",
       "        0.00099785, 0.00132288, 0.00108953, 0.00087456, 0.00089235,\n",
       "        0.00074459, 0.00071639, 0.00073328, 0.00078187, 0.00079618,\n",
       "        0.00061936, 0.00060872, 0.00073   , 0.00074299, 0.00080764,\n",
       "        0.00069025, 0.00065066, 0.00062383, 0.00062123, 0.00118356,\n",
       "        0.00197663, 0.00280499, 0.00226106, 0.00178213, 0.00322748,\n",
       "        0.00254998, 0.00240464, 0.00213805, 0.00203318, 0.00186311,\n",
       "        0.00286367, 0.00506843, 0.00421312, 0.00432767, 0.00327415,\n",
       "        0.0024244 , 0.00275122, 0.00183106, 0.00193079, 0.00320615,\n",
       "        0.00229031, 0.00207504, 0.00141401, 0.00127057, 0.00129358,\n",
       "        0.00129977, 0.00136206, 0.00140078, 0.00119454, 0.0015382 ,\n",
       "        0.00184812, 0.00101498, 0.00092585, 0.00092247, 0.00111693,\n",
       "        0.00084099, 0.00095859, 0.00077773, 0.00088244, 0.00063672,\n",
       "        0.00150292, 0.00111817, 0.00073278, 0.00073735, 0.00099943,\n",
       "        0.00053087, 0.00056859, 0.00062001, 0.0009019 , 0.0010623 ,\n",
       "        0.00097706, 0.0006946 , 0.00056003, 0.00060881, 0.00073449,\n",
       "        0.00112761, 0.0007438 , 0.00070482, 0.00131715, 0.00090316,\n",
       "        0.00061949, 0.00084879, 0.00129344, 0.001076  , 0.0015524 ,\n",
       "        0.00139307, 0.00119369, 0.00080036, 0.00083437, 0.00068281,\n",
       "        0.00074075, 0.00090341, 0.00078421, 0.00058743, 0.00059718,\n",
       "        0.00075272, 0.00104696, 0.00127642, 0.00084275, 0.00102042,\n",
       "        0.0012583 , 0.00107388, 0.00107019, 0.00365299, 0.00251962,\n",
       "        0.00232739, 0.00109666, 0.00091835, 0.00083183, 0.00128338,\n",
       "        0.00150017, 0.00099699, 0.00075295, 0.00080254, 0.00067045,\n",
       "        0.00060352, 0.00056912, 0.00072951, 0.00076782, 0.00061279,\n",
       "        0.00062312, 0.00056688, 0.00068198, 0.00088623, 0.00067411,\n",
       "        0.0010319 , 0.00202873, 0.00091004, 0.00178887, 0.00218646,\n",
       "        0.00184385, 0.00152163, 0.00181158, 0.00118205, 0.00105351,\n",
       "        0.0014453 , 0.00227825, 0.00180215, 0.00259893, 0.00179794,\n",
       "        0.00151159, 0.0014503 , 0.0024791 , 0.00323242, 0.00324009,\n",
       "        0.00250806, 0.00277724, 0.00312144, 0.0024275 , 0.00160956,\n",
       "        0.00197853, 0.0039121 , 0.0020911 , 0.00200771, 0.00161643,\n",
       "        0.00166111, 0.00171267, 0.00201491, 0.0015985 , 0.00230355,\n",
       "        0.00153734, 0.00185465, 0.00203585, 0.00117018, 0.00098592,\n",
       "        0.00158865, 0.00142141, 0.00151411, 0.00150335, 0.00403914,\n",
       "        0.00202928, 0.00143436, 0.00340298, 0.00275212, 0.00179309,\n",
       "        0.00238013, 0.00295771, 0.00526015, 0.00333894, 0.00475761,\n",
       "        0.00790251, 0.00978996, 0.00387892, 0.00449845, 0.00228181,\n",
       "        0.00299972, 0.00272474, 0.00202067, 0.00567277, 0.00594253,\n",
       "        0.01153318, 0.00900741, 0.00762851, 0.00676832, 0.00752074,\n",
       "        0.00728627, 0.00398585, 0.00379676, 0.00185349, 0.00255006,\n",
       "        0.00398423, 0.00213535, 0.00139561, 0.00125112, 0.0014121 ,\n",
       "        0.00089566, 0.0013377 , 0.00096947, 0.00373589, 0.00288354,\n",
       "        0.0026407 , 0.00233754, 0.0015796 , 0.00090469, 0.00122581,\n",
       "        0.00110233, 0.00088306, 0.0011812 , 0.00084005, 0.00089557,\n",
       "        0.00078443, 0.00102131, 0.00077435, 0.0010417 , 0.00121088,\n",
       "        0.00080624, 0.0011603 , 0.00125222, 0.00081683, 0.00214213,\n",
       "        0.00195381, 0.00191267, 0.00246396, 0.00187476, 0.00146783,\n",
       "        0.0019599 , 0.00287072, 0.0036971 , 0.00199875, 0.00199554,\n",
       "        0.00318594, 0.00220818, 0.00603622, 0.00476974, 0.00316448,\n",
       "        0.00112734, 0.00304776, 0.00188504, 0.00355516, 0.00378231,\n",
       "        0.00405329, 0.00388514, 0.00357079, 0.00522695, 0.00195877,\n",
       "        0.00124515, 0.0015011 , 0.00169526, 0.00125189, 0.00220771,\n",
       "        0.00187504, 0.00126237, 0.00098132, 0.00152122, 0.00183581,\n",
       "        0.0013124 , 0.00092526, 0.00095156, 0.00128421, 0.00143803,\n",
       "        0.00184784, 0.00153222, 0.00137411, 0.00278126, 0.00148972,\n",
       "        0.00108928, 0.00125003, 0.00194122, 0.00127598, 0.00575185,\n",
       "        0.00506173, 0.0045857 , 0.00375434, 0.00515112, 0.00497705,\n",
       "        0.00513741, 0.00325226, 0.00237242, 0.00272791, 0.00192598,\n",
       "        0.00195214, 0.00186681, 0.00157411, 0.00217116, 0.00173438,\n",
       "        0.0025259 , 0.00167883, 0.00089736, 0.00135089, 0.00213061,\n",
       "        0.00157351, 0.00110994, 0.00114488, 0.00111308, 0.00101903,\n",
       "        0.00098597, 0.0008864 , 0.00085799, 0.00075013, 0.0008163 ,\n",
       "        0.00071697, 0.00085013, 0.00177344, 0.00343454, 0.00199461,\n",
       "        0.00141243, 0.0013882 , 0.00116527, 0.00095908, 0.00105485,\n",
       "        0.00121921, 0.0011013 , 0.00090474, 0.00079971, 0.00076817,\n",
       "        0.00083271, 0.00081646, 0.00102501, 0.00105757, 0.00093404,\n",
       "        0.00070566, 0.00069033, 0.00069223, 0.00084568, 0.00070276,\n",
       "        0.00104085, 0.00125603, 0.00092629, 0.00083328, 0.00071798,\n",
       "        0.00087773, 0.00134177, 0.00119698, 0.00088421, 0.00101056,\n",
       "        0.00077195, 0.0007677 , 0.00134839, 0.00097884, 0.0010092 ,\n",
       "        0.00080908, 0.00092984, 0.00085643, 0.00087927, 0.00067059,\n",
       "        0.00069328, 0.00063311, 0.00089594, 0.00095254, 0.00076519,\n",
       "        0.00076372, 0.00085401, 0.00097242, 0.00099883, 0.00109928,\n",
       "        0.00085935, 0.00081451, 0.00069681, 0.00091832, 0.00097381,\n",
       "        0.00113807, 0.00095123, 0.00093698, 0.00083772, 0.00066057,\n",
       "        0.0015445 , 0.00111402, 0.0012015 , 0.001303  , 0.00092764,\n",
       "        0.00058174, 0.0006448 , 0.00070057, 0.00095442, 0.00080063,\n",
       "        0.00087502, 0.00078105, 0.00088037, 0.00068487, 0.00077186,\n",
       "        0.00217324, 0.00150653, 0.00110661, 0.00105635, 0.00068482,\n",
       "        0.00070226, 0.00099987, 0.00067047, 0.00076404, 0.0006854 ,\n",
       "        0.00072277, 0.00059126, 0.00054206, 0.00113055, 0.00100832,\n",
       "        0.00076351, 0.00068557, 0.00093988, 0.00065289, 0.00102891,\n",
       "        0.00069767, 0.00065471, 0.00102124, 0.0008089 , 0.00090722,\n",
       "        0.00079132, 0.00078373, 0.00079375, 0.00077509, 0.00100814,\n",
       "        0.00088477, 0.00098865, 0.00062167, 0.00081567, 0.00079372,\n",
       "        0.00145095, 0.00227255, 0.00123746, 0.00171383, 0.00101062,\n",
       "        0.00085672, 0.000951  , 0.00092028, 0.00084723, 0.00101655,\n",
       "        0.00077261, 0.00058494, 0.0005509 , 0.00059913, 0.0006125 ,\n",
       "        0.00092302, 0.00107021, 0.00104179, 0.00081748, 0.00080397,\n",
       "        0.00094997, 0.00094199, 0.00090203, 0.00064526, 0.00117655,\n",
       "        0.00093215, 0.00116495, 0.00118064, 0.00074509, 0.00133081,\n",
       "        0.00077851, 0.00062361, 0.0007426 , 0.00071255, 0.00063538,\n",
       "        0.00122797, 0.00081453, 0.00093421, 0.00081527, 0.00064082,\n",
       "        0.00057904, 0.00067065, 0.00076488, 0.0006602 , 0.00102419,\n",
       "        0.00160433, 0.00083145, 0.00107384, 0.00074333, 0.00081995,\n",
       "        0.00082285, 0.00072913, 0.00074003, 0.0013844 , 0.0007828 ,\n",
       "        0.00109289, 0.00068533, 0.00073221, 0.00062336, 0.00071841,\n",
       "        0.0006785 , 0.00054654, 0.00066974, 0.00071115, 0.0006984 ,\n",
       "        0.00054164, 0.00067947, 0.00050222, 0.0004972 , 0.00047269,\n",
       "        0.00053274, 0.00051976, 0.00051612, 0.00049165, 0.00039512,\n",
       "        0.00044032, 0.00049163, 0.00052226, 0.00049396, 0.00043782,\n",
       "        0.00046611, 0.00044552, 0.00038926, 0.00043873, 0.00048577,\n",
       "        0.00069241, 0.00054222, 0.00055934, 0.00062905, 0.00041481,\n",
       "        0.00062432, 0.00078143, 0.00054192, 0.00055386, 0.00063487,\n",
       "        0.00048866, 0.00047569, 0.00055954, 0.00054273, 0.00050615,\n",
       "        0.00081764, 0.00114878, 0.00185992, 0.00101545, 0.00129839,\n",
       "        0.00077497, 0.00091173, 0.00101168, 0.00075139, 0.00059834,\n",
       "        0.00054408, 0.00064845, 0.00050736, 0.00044664, 0.00042474,\n",
       "        0.00064406, 0.00074287, 0.00046265, 0.00051287, 0.00060561,\n",
       "        0.00052211, 0.00067437, 0.00055057, 0.0004097 , 0.00034692,\n",
       "        0.00038237, 0.00095336, 0.00045843, 0.00043805, 0.0004049 ,\n",
       "        0.00038435, 0.0004528 , 0.00062677, 0.00066014, 0.00056886,\n",
       "        0.00109787, 0.00059767, 0.0005387 , 0.00057488, 0.00075753,\n",
       "        0.00106329, 0.00058865, 0.00050302, 0.00057248, 0.00051949,\n",
       "        0.00044462, 0.00079066, 0.00056198, 0.00077745, 0.00066855,\n",
       "        0.00049748, 0.0003778 , 0.00038655, 0.00045688, 0.00038437,\n",
       "        0.00071907, 0.00041714, 0.00042201, 0.00037424, 0.00032996,\n",
       "        0.00033677, 0.00047564, 0.0003419 , 0.00036347, 0.0003132 ,\n",
       "        0.00026509, 0.00046071, 0.00043111, 0.000367  , 0.00046025,\n",
       "        0.00051088, 0.00041263, 0.00040636, 0.00040124, 0.00034024,\n",
       "        0.00041263, 0.00041849, 0.00041696, 0.00038461, 0.00030614,\n",
       "        0.0003015 , 0.0003838 , 0.00031437, 0.00041043, 0.00027088,\n",
       "        0.00024398, 0.00021265, 0.00022157, 0.00038467, 0.00028998,\n",
       "        0.00030581, 0.00035701, 0.00029108, 0.00026386, 0.00026692,\n",
       "        0.0003869 , 0.00030096, 0.00027982, 0.00047653, 0.00030927]),\n",
       " ('AR(30)',\n",
       "  'Rolling'): array([ 4.21668655e-04,  6.87616159e-04,  7.83666254e-04,  6.96650281e-04,\n",
       "         5.67356160e-04,  5.60460905e-04,  5.48456235e-04,  4.85047515e-04,\n",
       "         4.15448865e-04,  5.06379410e-04,  3.50844267e-04,  3.62742503e-04,\n",
       "         4.35187212e-04,  4.75310178e-04,  3.33461206e-04,  3.28090320e-04,\n",
       "         3.12029594e-04,  4.11194088e-04,  3.04029558e-04,  4.05334021e-04,\n",
       "         3.71226520e-04,  3.31446558e-04,  3.46918233e-04,  3.88489038e-04,\n",
       "         3.10301796e-04,  2.43170455e-04,  4.20819593e-04,  2.68158653e-04,\n",
       "         3.31434287e-04,  3.62215592e-04,  3.22296973e-04,  4.06440228e-04,\n",
       "         3.01788768e-04,  3.60695310e-04,  3.60540041e-04,  2.31653937e-04,\n",
       "         3.24283604e-04,  3.43540893e-04,  3.64160197e-04,  6.19440575e-04,\n",
       "         7.77612653e-04,  7.69062885e-04,  7.55240091e-04,  7.78386108e-04,\n",
       "         1.02744487e-03,  1.24873877e-03,  1.06665657e-03,  8.57414219e-04,\n",
       "         9.48586529e-04,  7.19703956e-04,  6.55029858e-04,  1.08382769e-03,\n",
       "         7.13311080e-04,  6.49589491e-04,  6.84684763e-04,  6.56679855e-04,\n",
       "         9.52480972e-04,  6.90453814e-04,  1.04141067e-03,  1.83519860e-03,\n",
       "         1.51473520e-03,  1.04388258e-03,  1.05978914e-03,  1.06779101e-03,\n",
       "         8.89446442e-04,  1.44616664e-03,  1.21496397e-03,  7.01689295e-04,\n",
       "         1.16788072e-03,  7.38822513e-04,  1.08133595e-03,  8.45738154e-04,\n",
       "         7.23119802e-04,  6.39125853e-04,  8.21897770e-04,  8.70233991e-04,\n",
       "         9.79566462e-04,  1.36906367e-03,  9.10938185e-04,  1.12610351e-03,\n",
       "         1.08091228e-03,  6.49403560e-04,  6.25722138e-04,  9.95310608e-04,\n",
       "         9.60975202e-04,  7.72846246e-04,  6.41142903e-04,  5.82050739e-04,\n",
       "         7.93501651e-04,  6.73510811e-04,  8.43088057e-04,  6.72799204e-04,\n",
       "         6.27576528e-04,  5.59064867e-04,  5.84786271e-04,  1.17914611e-03,\n",
       "         1.79737506e-03,  2.71166723e-03,  2.22350983e-03,  2.18292838e-03,\n",
       "         3.61301815e-03,  2.50124844e-03,  3.00498217e-03,  2.67962099e-03,\n",
       "         2.15876322e-03,  1.85633991e-03,  3.33281024e-03,  5.00198229e-03,\n",
       "         4.30624522e-03,  5.06920148e-03,  3.65683256e-03,  2.76402854e-03,\n",
       "         3.12434344e-03,  2.15567876e-03,  2.36279138e-03,  3.56481938e-03,\n",
       "         2.26014837e-03,  2.50241920e-03,  1.80509567e-03,  1.14391995e-03,\n",
       "         1.51114810e-03,  1.70295644e-03,  1.15783184e-03,  1.25573528e-03,\n",
       "         1.32507406e-03,  1.72475450e-03,  1.91485927e-03,  1.11052395e-03,\n",
       "         9.50314522e-04,  8.20569619e-04,  1.03013907e-03,  1.40623076e-03,\n",
       "         1.53887122e-03,  8.72045114e-04,  9.26912191e-04,  6.60158816e-04,\n",
       "         1.59943236e-03,  9.12741038e-04,  7.80225099e-04,  1.05470820e-03,\n",
       "         1.03566047e-03,  3.80833424e-04,  7.28574623e-04,  6.21471174e-04,\n",
       "         7.36283246e-04,  9.90742066e-04,  1.01627756e-03,  6.90785827e-04,\n",
       "         5.61298468e-04,  6.12780732e-04,  8.22366137e-04,  9.97941958e-04,\n",
       "         5.93234203e-04,  7.93148818e-04,  1.27580486e-03,  6.61196065e-04,\n",
       "         8.00649024e-04,  9.14307453e-04,  1.10516940e-03,  8.82391183e-04,\n",
       "         1.81490728e-03,  1.29950474e-03,  1.16048036e-03,  8.95602093e-04,\n",
       "         1.09055851e-03,  6.48653229e-04,  7.82515925e-04,  9.09665387e-04,\n",
       "         8.33370774e-04,  6.45953689e-04,  6.06785506e-04,  6.11328959e-04,\n",
       "         8.13360041e-04,  1.03813850e-03,  7.83299958e-04,  1.01911012e-03,\n",
       "         1.03928214e-03,  9.23175479e-04,  1.34247516e-03,  3.46772389e-03,\n",
       "         1.93359225e-03,  2.76803954e-03,  1.56651017e-03,  1.24051069e-03,\n",
       "         1.01648790e-03,  1.89842450e-03,  1.70838907e-03,  1.09991233e-03,\n",
       "         8.55618354e-04,  6.88202289e-04,  5.78075960e-04,  6.44845766e-04,\n",
       "         4.53609219e-04,  5.73581394e-04,  3.34447238e-04,  4.87325187e-04,\n",
       "         5.24802161e-04,  6.54814118e-04,  5.08986051e-04,  1.01611031e-03,\n",
       "         5.24695107e-04,  1.00311431e-03,  1.56066661e-03,  1.31410383e-03,\n",
       "         2.12612899e-03,  2.08222358e-03,  1.60515129e-03,  1.73167812e-03,\n",
       "         2.02922085e-03,  1.21574783e-03,  1.46497037e-03,  1.73071011e-03,\n",
       "         2.13282706e-03,  1.81604941e-03,  2.73796419e-03,  1.58430402e-03,\n",
       "         1.66994046e-03,  1.51302893e-03,  2.55640747e-03,  3.00420838e-03,\n",
       "         3.26686009e-03,  2.70047935e-03,  3.02826894e-03,  3.07542583e-03,\n",
       "         2.72973966e-03,  2.29267328e-03,  2.85838093e-03,  3.54300297e-03,\n",
       "         2.09587150e-03,  2.94690089e-03,  1.89404180e-03,  1.53611710e-03,\n",
       "         1.91036560e-03,  2.16834881e-03,  1.53702973e-03,  2.04051507e-03,\n",
       "         1.63270422e-03,  1.82897632e-03,  2.18536059e-03,  1.24596071e-03,\n",
       "         1.17965302e-03,  1.58706612e-03,  1.34304176e-03,  1.72534842e-03,\n",
       "         1.84492090e-03,  3.62507501e-03,  1.42719856e-03,  2.49846807e-03,\n",
       "         3.30812669e-03,  2.08257360e-03,  2.13947518e-03,  3.64871533e-03,\n",
       "         3.32999835e-03,  4.39970450e-03,  3.56131749e-03,  5.91905133e-03,\n",
       "         7.29140736e-03,  8.92265257e-03,  5.04018099e-03,  6.70521753e-03,\n",
       "         3.39932957e-03,  2.73299811e-03,  4.17020801e-03,  3.85412997e-03,\n",
       "         6.96552780e-03,  6.33827606e-03,  1.11049117e-02,  8.53622256e-03,\n",
       "         7.54441683e-03,  7.22569287e-03,  6.12428828e-03,  5.68540589e-03,\n",
       "         5.09038003e-03,  6.03457564e-03,  3.41930903e-03,  3.51773637e-03,\n",
       "         5.31786685e-03,  4.39843403e-03,  1.08457482e-03,  3.09224512e-03,\n",
       "         1.16739686e-03,  9.51570862e-04,  3.93505031e-04,  1.66354181e-03,\n",
       "         5.00263280e-03,  2.19502257e-03,  2.48592725e-03,  6.34613549e-04,\n",
       "         1.81467775e-03, -3.26492384e-04,  1.09745580e-03,  5.91122600e-04,\n",
       "         7.13008941e-04,  1.97926024e-03,  1.86833131e-03,  1.50824132e-03,\n",
       "         6.71432081e-04,  1.38318612e-03,  7.31083755e-04,  9.83935536e-04,\n",
       "         1.44575250e-03,  8.57110737e-04,  1.07831041e-03,  1.34332898e-03,\n",
       "         1.11274239e-03,  1.55651899e-03,  1.80283270e-03,  1.55894672e-03,\n",
       "         2.34851958e-03,  1.91575318e-03,  1.48116688e-03,  1.71779543e-03,\n",
       "         2.96364051e-03,  3.59503664e-03,  2.18686297e-03,  2.58378799e-03,\n",
       "         3.07513584e-03,  2.14911623e-03,  5.36087203e-03,  5.14606361e-03,\n",
       "         3.72398390e-03,  1.44129325e-03,  3.25191132e-03,  1.82834686e-03,\n",
       "         3.00333659e-03,  4.63710276e-03,  4.81338959e-03,  3.42144368e-03,\n",
       "         3.80813599e-03,  4.76218974e-03,  2.29573161e-03,  1.67702817e-03,\n",
       "         2.08353803e-03,  1.54512285e-03,  8.64040213e-04,  2.64050222e-03,\n",
       "         2.73943403e-03,  1.41517268e-03,  1.01202583e-03,  1.31970294e-03,\n",
       "         1.11321446e-03,  7.18352658e-04,  1.42678761e-03,  7.66574283e-04,\n",
       "         1.38678637e-03,  1.03696663e-03,  2.19113708e-03,  1.27135971e-03,\n",
       "         1.17675561e-03,  2.41001430e-03,  1.07067331e-03,  9.49450379e-04,\n",
       "         1.27046141e-03,  2.31382074e-03,  1.43199520e-03,  5.38288654e-03,\n",
       "         4.98374783e-03,  4.90956695e-03,  3.86233812e-03,  5.32715695e-03,\n",
       "         4.39856540e-03,  4.86758977e-03,  4.04258341e-03,  3.18594944e-03,\n",
       "         2.97247016e-03,  2.93141819e-03,  1.93725127e-03,  2.34902202e-03,\n",
       "         2.29472311e-03,  2.45706421e-03,  1.62408373e-03,  2.26205545e-03,\n",
       "         1.55118281e-03,  1.03210393e-03,  1.69356762e-03,  1.86262868e-03,\n",
       "         9.25884391e-04,  1.08854265e-03,  6.68936890e-04,  8.51906950e-04,\n",
       "         5.66913147e-04,  9.57133432e-04,  3.64613918e-04,  8.59091830e-04,\n",
       "         9.13712185e-04,  9.20167569e-04,  7.92173982e-04,  8.65728515e-04,\n",
       "         1.89067442e-03,  3.33935205e-03,  1.91846471e-03,  1.47424903e-03,\n",
       "         1.42943306e-03,  1.06688374e-03,  6.40671288e-04,  1.04869488e-03,\n",
       "         1.47897800e-03,  1.23472469e-03,  9.59357901e-04,  1.12834300e-03,\n",
       "         5.83853007e-04,  7.96899681e-04,  9.44846049e-04,  9.19358714e-04,\n",
       "         8.73320331e-04,  7.95513374e-04,  7.48462369e-04,  7.13378870e-04,\n",
       "         6.93436638e-04,  7.07625092e-04,  4.13538187e-04,  7.59976976e-04,\n",
       "         7.83783093e-04,  8.41918426e-04,  9.06052036e-04,  7.26825352e-04,\n",
       "         8.77738678e-04,  1.34007219e-03,  1.09829281e-03,  8.65772130e-04,\n",
       "         9.91632815e-04,  8.61391592e-04,  7.07195156e-04,  1.16626566e-03,\n",
       "         1.03566542e-03,  9.39984538e-04,  8.05900963e-04,  8.95174199e-04,\n",
       "         6.90223967e-04,  8.18979429e-04,  7.83866784e-04,  6.76467397e-04,\n",
       "         5.53173433e-04,  8.02806081e-04,  7.19534821e-04,  6.92611958e-04,\n",
       "         7.69579945e-04,  8.55075899e-04,  7.98707737e-04,  8.35797507e-04,\n",
       "         9.60916165e-04,  7.92641651e-04,  8.15495416e-04,  6.75996536e-04,\n",
       "         7.91126825e-04,  9.14602144e-04,  1.00460450e-03,  9.51452452e-04,\n",
       "         9.29284822e-04,  7.76560374e-04,  6.10957478e-04,  1.36368115e-03,\n",
       "         1.10462913e-03,  1.16194829e-03,  1.30743490e-03,  9.54703501e-04,\n",
       "         5.04405223e-04,  5.47470415e-04,  7.07180336e-04,  8.34084316e-04,\n",
       "         7.01058807e-04,  9.62813437e-04,  6.29963971e-04,  7.66000445e-04,\n",
       "         6.93866895e-04,  6.19134873e-04,  1.89241235e-03,  1.39968499e-03,\n",
       "         1.00405599e-03,  1.05045514e-03,  6.91813025e-04,  5.58969997e-04,\n",
       "         7.87823268e-04,  8.37125646e-04,  6.12757752e-04,  6.01435029e-04,\n",
       "         8.77541603e-04,  2.76751248e-04,  5.30974368e-04,  1.28334985e-03,\n",
       "         8.20274782e-04,  7.11435404e-04,  6.00611950e-04,  7.17294406e-04,\n",
       "         5.39481920e-04,  8.77174202e-04,  7.75984345e-04,  5.22095309e-04,\n",
       "         8.50871255e-04,  6.15020507e-04,  5.84585162e-04,  8.41181914e-04,\n",
       "         8.36326276e-04,  7.87424798e-04,  7.83264956e-04,  8.74004573e-04,\n",
       "         7.55298691e-04,  9.03434773e-04,  7.27408356e-04,  6.76445615e-04,\n",
       "         7.47149297e-04,  1.18995535e-03,  1.93968920e-03,  1.25094302e-03,\n",
       "         1.62898345e-03,  1.23554140e-03,  6.52435504e-04,  8.65737527e-04,\n",
       "         8.27467232e-04,  8.49077563e-04,  1.01941852e-03,  8.64526136e-04,\n",
       "         6.61060007e-04,  3.62562854e-04,  7.43880266e-04,  4.74355428e-04,\n",
       "         7.13592908e-04,  9.98859749e-04,  7.42153308e-04,  7.63330058e-04,\n",
       "         7.81856121e-04,  9.07825115e-04,  7.92335326e-04,  7.25032823e-04,\n",
       "         5.22710368e-04,  7.28880576e-04,  8.84411567e-04,  1.04097380e-03,\n",
       "         1.16925109e-03,  9.87560282e-04,  1.28884580e-03,  7.24655607e-04,\n",
       "         5.71139070e-04,  6.06541088e-04,  6.73755216e-04,  7.21184191e-04,\n",
       "         1.05510553e-03,  8.99520718e-04,  8.59683293e-04,  6.17171445e-04,\n",
       "         6.73785204e-04,  2.88272046e-04,  5.25741027e-04,  7.52610302e-04,\n",
       "         4.90867115e-04,  9.25757070e-04,  1.50855905e-03,  7.75887297e-04,\n",
       "         9.25143090e-04,  8.20673248e-04,  6.31476567e-04,  5.18327880e-04,\n",
       "         7.00037699e-04,  7.52625370e-04,  1.16703150e-03,  1.05870450e-03,\n",
       "         1.07813592e-03,  5.47169181e-04,  8.77993996e-04,  3.41518024e-04,\n",
       "         5.05476073e-04,  6.88684934e-04,  4.01647537e-04,  6.83698156e-04,\n",
       "         8.24693600e-04,  5.13247596e-04,  5.22570409e-04,  5.05012322e-04,\n",
       "         3.58412615e-04,  1.88327845e-04,  3.37505233e-04,  4.76098870e-04,\n",
       "         3.63244193e-04,  5.80702407e-04,  4.75907338e-04,  2.36819540e-04,\n",
       "         3.50844624e-04,  2.44667587e-04,  3.45400065e-04,  3.94141654e-04,\n",
       "         3.73275310e-04,  4.03335860e-04,  3.95056605e-04,  2.57183036e-04,\n",
       "         3.40879464e-04,  3.49387172e-04,  5.49863471e-04,  4.18282038e-04,\n",
       "         4.26068892e-04,  5.43631792e-04,  3.22299340e-04,  4.72732499e-04,\n",
       "         6.65103594e-04,  4.66078382e-04,  4.38515958e-04,  5.78244758e-04,\n",
       "         4.40288855e-04,  3.01651061e-04,  4.89737265e-04,  5.16006804e-04,\n",
       "         3.29158675e-04,  7.16016811e-04,  1.00170211e-03,  1.50881588e-03,\n",
       "         1.08919814e-03,  1.25105819e-03,  8.40304696e-04,  7.18703709e-04,\n",
       "         7.70745772e-04,  6.33013867e-04,  5.99186740e-04,  7.13911579e-04,\n",
       "         6.45705976e-04,  5.65477517e-04,  3.12676796e-04,  4.76862138e-04,\n",
       "         5.47353685e-04,  6.30474068e-04,  4.49351931e-04,  3.09118087e-04,\n",
       "         6.18430660e-04,  5.23374113e-04,  5.65230481e-04,  4.14918439e-04,\n",
       "         3.10663596e-04,  2.09682679e-04,  9.49992870e-05,  7.39766319e-04,\n",
       "         5.24648813e-04,  4.87788247e-04,  4.51980236e-04,  4.30411898e-04,\n",
       "         4.16004269e-04,  4.63035053e-04,  5.47681100e-04,  5.87634170e-04,\n",
       "         9.40136041e-04,  7.18397576e-04,  4.27366099e-04,  5.01245411e-04,\n",
       "         6.32791665e-04,  9.31982923e-04,  6.41514551e-04,  3.96506110e-04,\n",
       "         4.60656414e-04,  5.09825764e-04,  5.94527224e-04,  4.07765955e-04,\n",
       "         5.25268191e-04,  7.32700990e-04,  6.20375715e-04,  4.03355389e-04,\n",
       "         2.68515319e-04,  3.67264434e-04,  4.59540383e-04,  4.25147305e-04,\n",
       "         4.74138291e-04,  2.95938751e-04,  2.63775187e-04,  3.54000037e-04,\n",
       "         2.36939148e-04,  2.87755800e-04,  3.08876852e-04,  2.69099062e-04,\n",
       "         2.18386502e-04,  1.07553656e-04,  3.03205682e-04,  2.97737596e-04,\n",
       "         4.27008871e-04,  3.89700059e-04,  2.27686359e-04,  3.27708390e-04,\n",
       "         3.46299652e-04,  2.14274222e-04,  3.26521220e-04,  3.92288176e-04,\n",
       "         3.60319520e-04,  3.93825262e-04,  2.88050254e-04,  2.39374417e-04,\n",
       "         1.93995578e-04,  2.70949749e-04,  2.64030591e-04,  2.53825927e-04,\n",
       "         3.27359810e-04,  1.67744709e-04,  1.63856230e-04,  1.43355122e-04,\n",
       "         8.89713499e-05,  2.45923088e-04,  2.36076725e-04,  1.44168771e-04,\n",
       "         2.49739615e-04,  1.96424014e-04,  1.42093131e-04,  1.68590277e-04,\n",
       "         2.68917505e-04,  2.47095178e-04,  2.08408861e-04,  3.39765524e-04,\n",
       "         2.38084958e-04]),\n",
       " ('HAR',\n",
       "  'Rolling'): array([0.0005806 , 0.00057587, 0.00074059, 0.00071929, 0.00060944,\n",
       "        0.00047355, 0.00047682, 0.00053228, 0.00042274, 0.00046552,\n",
       "        0.00037244, 0.00036841, 0.00040196, 0.00045736, 0.00036124,\n",
       "        0.00029381, 0.00027561, 0.00044627, 0.00038259, 0.00041513,\n",
       "        0.00038129, 0.00036893, 0.00031192, 0.00033976, 0.0003176 ,\n",
       "        0.00029597, 0.00038301, 0.00030549, 0.00033364, 0.00037734,\n",
       "        0.00034735, 0.00038957, 0.00030639, 0.00034592, 0.00039606,\n",
       "        0.00029605, 0.00030751, 0.00030441, 0.00037314, 0.00062587,\n",
       "        0.00077871, 0.00077473, 0.00070503, 0.00073131, 0.00100605,\n",
       "        0.00125634, 0.00106108, 0.00080642, 0.00090905, 0.00084783,\n",
       "        0.00069201, 0.00096978, 0.00086675, 0.00073056, 0.00064426,\n",
       "        0.00071896, 0.00103643, 0.00080128, 0.00095532, 0.00177311,\n",
       "        0.00174203, 0.00108964, 0.00083698, 0.00110618, 0.00107424,\n",
       "        0.00125456, 0.00119736, 0.00090262, 0.00114058, 0.00078694,\n",
       "        0.00093267, 0.00079109, 0.00070415, 0.00076509, 0.00092497,\n",
       "        0.00101083, 0.00098821, 0.00128303, 0.00110533, 0.00093588,\n",
       "        0.00095798, 0.00082389, 0.00078394, 0.0007858 , 0.00080342,\n",
       "        0.00080188, 0.00063888, 0.00061498, 0.00071482, 0.00072086,\n",
       "        0.00077287, 0.00066895, 0.00062906, 0.00060492, 0.00060186,\n",
       "        0.00109736, 0.00183978, 0.00266381, 0.00224482, 0.00188363,\n",
       "        0.00326069, 0.00275879, 0.00271563, 0.00247819, 0.00236139,\n",
       "        0.00221614, 0.0031313 , 0.00500371, 0.0044284 , 0.00461097,\n",
       "        0.00379524, 0.00310755, 0.00337499, 0.00247417, 0.00246625,\n",
       "        0.0034503 , 0.0025932 , 0.00235015, 0.00174255, 0.00152401,\n",
       "        0.0015177 , 0.0014741 , 0.00142413, 0.00141032, 0.00119082,\n",
       "        0.0015399 , 0.00178995, 0.00109861, 0.00100354, 0.0009728 ,\n",
       "        0.00111684, 0.00087512, 0.00093528, 0.00073644, 0.00080711,\n",
       "        0.00061444, 0.00134197, 0.00103811, 0.00073359, 0.00073234,\n",
       "        0.00095089, 0.00056685, 0.00059674, 0.00059064, 0.0008015 ,\n",
       "        0.00094959, 0.0008988 , 0.00066402, 0.00056181, 0.00060347,\n",
       "        0.00070785, 0.00102332, 0.00071225, 0.00066921, 0.00117373,\n",
       "        0.0008793 , 0.00067024, 0.0008507 , 0.00119304, 0.00105094,\n",
       "        0.0014604 , 0.00134375, 0.00121151, 0.00092706, 0.00095067,\n",
       "        0.00079715, 0.00082134, 0.00090395, 0.00077633, 0.00059099,\n",
       "        0.00058664, 0.00069833, 0.00094165, 0.00114631, 0.00081585,\n",
       "        0.00096379, 0.00118328, 0.00107332, 0.00109067, 0.00319473,\n",
       "        0.00241809, 0.00236208, 0.00143428, 0.00127263, 0.00118938,\n",
       "        0.00154807, 0.00157728, 0.00110146, 0.0008193 , 0.00083879,\n",
       "        0.00072475, 0.00066139, 0.00059282, 0.0006677 , 0.00068295,\n",
       "        0.00055656, 0.00055367, 0.00050551, 0.00059826, 0.00077274,\n",
       "        0.00060703, 0.00089537, 0.0017359 , 0.00090611, 0.00164827,\n",
       "        0.00204403, 0.00183821, 0.00164258, 0.00190922, 0.00137997,\n",
       "        0.00129291, 0.00157572, 0.00221187, 0.00184689, 0.00251543,\n",
       "        0.00190265, 0.00170308, 0.00167981, 0.00252559, 0.00315415,\n",
       "        0.00324406, 0.00267831, 0.00293917, 0.00329331, 0.00281798,\n",
       "        0.00211784, 0.00233998, 0.00380807, 0.00242674, 0.00230936,\n",
       "        0.00191018, 0.00189072, 0.00193259, 0.00215754, 0.00169325,\n",
       "        0.00222802, 0.00163113, 0.00188713, 0.00203512, 0.00136244,\n",
       "        0.00115481, 0.00159359, 0.00141212, 0.00148387, 0.00147418,\n",
       "        0.00348735, 0.00205448, 0.00166143, 0.00321801, 0.00284155,\n",
       "        0.00217674, 0.0026606 , 0.00301196, 0.00487092, 0.00366756,\n",
       "        0.00477139, 0.00746347, 0.0093834 , 0.0051687 , 0.00557616,\n",
       "        0.00359398, 0.0040886 , 0.00364352, 0.00283439, 0.00551484,\n",
       "        0.00594533, 0.01083624, 0.00909527, 0.00814807, 0.0076627 ,\n",
       "        0.00845254, 0.00820326, 0.00529838, 0.00476583, 0.00283028,\n",
       "        0.00339357, 0.00453901, 0.00273073, 0.00184053, 0.00162615,\n",
       "        0.00164037, 0.00113249, 0.0014665 , 0.00112271, 0.00373949,\n",
       "        0.00292809, 0.00274475, 0.00247407, 0.00178263, 0.0011394 ,\n",
       "        0.00142609, 0.00124614, 0.00100407, 0.00122792, 0.00088428,\n",
       "        0.00091617, 0.00080423, 0.00100907, 0.00077524, 0.0010164 ,\n",
       "        0.00116385, 0.00079844, 0.00111898, 0.00120894, 0.00081657,\n",
       "        0.00202638, 0.00187487, 0.00185334, 0.00237855, 0.00186964,\n",
       "        0.00151161, 0.00197443, 0.00280222, 0.00357547, 0.00206541,\n",
       "        0.00205445, 0.00313743, 0.00229621, 0.00576911, 0.00469748,\n",
       "        0.00327257, 0.00146489, 0.00317721, 0.00213407, 0.00362817,\n",
       "        0.00376751, 0.00398556, 0.00385892, 0.00365559, 0.00514376,\n",
       "        0.00231882, 0.00163224, 0.00178619, 0.00188965, 0.00143317,\n",
       "        0.00223069, 0.00184883, 0.00129152, 0.00103837, 0.00151237,\n",
       "        0.00178677, 0.00133038, 0.00095692, 0.00095438, 0.00124537,\n",
       "        0.00139267, 0.00176001, 0.00147446, 0.00134227, 0.00261714,\n",
       "        0.00150801, 0.00115396, 0.00128884, 0.00189069, 0.00130824,\n",
       "        0.00533481, 0.00477574, 0.00445937, 0.00381688, 0.00514637,\n",
       "        0.00509149, 0.00533434, 0.00360676, 0.00275109, 0.00300467,\n",
       "        0.00224544, 0.00218145, 0.00202454, 0.00167502, 0.0021733 ,\n",
       "        0.00177237, 0.00246287, 0.00170828, 0.00098998, 0.00137834,\n",
       "        0.00207774, 0.0015692 , 0.00114288, 0.00113589, 0.00109021,\n",
       "        0.00100995, 0.00096893, 0.000846  , 0.00080065, 0.00069399,\n",
       "        0.00074593, 0.0006497 , 0.00076626, 0.00161099, 0.00315559,\n",
       "        0.00190603, 0.0014117 , 0.00140683, 0.00122168, 0.00104235,\n",
       "        0.00110808, 0.00119573, 0.00106812, 0.00088062, 0.00077299,\n",
       "        0.00073558, 0.00079154, 0.00077182, 0.00095154, 0.0009784 ,\n",
       "        0.00087021, 0.00066612, 0.00065074, 0.00064869, 0.00078565,\n",
       "        0.00064921, 0.00095192, 0.00115181, 0.00086487, 0.00078695,\n",
       "        0.00068568, 0.00082892, 0.00125788, 0.00113532, 0.00085051,\n",
       "        0.00096571, 0.00075174, 0.00075134, 0.00127626, 0.00093413,\n",
       "        0.00095822, 0.00078093, 0.00088701, 0.00082418, 0.00084608,\n",
       "        0.00064347, 0.0006556 , 0.00059247, 0.00082902, 0.00088013,\n",
       "        0.00071138, 0.00070746, 0.00079243, 0.00090476, 0.0009371 ,\n",
       "        0.00103078, 0.00081487, 0.00077559, 0.00067052, 0.00086875,\n",
       "        0.00091859, 0.00106738, 0.0008986 , 0.00088834, 0.00080007,\n",
       "        0.00064414, 0.00144217, 0.00106518, 0.00114425, 0.00124248,\n",
       "        0.00090875, 0.0005956 , 0.00065047, 0.00067764, 0.00089728,\n",
       "        0.00074663, 0.00079679, 0.00070499, 0.00080483, 0.00063064,\n",
       "        0.0007083 , 0.00197897, 0.00140728, 0.00106403, 0.00102656,\n",
       "        0.00069282, 0.00070315, 0.00096704, 0.00063723, 0.00070108,\n",
       "        0.00062254, 0.00064708, 0.00052941, 0.00048332, 0.00100439,\n",
       "        0.00090792, 0.00069244, 0.00062411, 0.0008523 , 0.00060201,\n",
       "        0.00094444, 0.00064259, 0.00059626, 0.0009238 , 0.00074065,\n",
       "        0.00082778, 0.00072915, 0.00071499, 0.00072583, 0.00071265,\n",
       "        0.00091678, 0.00080916, 0.00090102, 0.00057479, 0.00074449,\n",
       "        0.00072619, 0.00132223, 0.00207268, 0.00117511, 0.0016108 ,\n",
       "        0.00100371, 0.00086957, 0.00095578, 0.00091564, 0.00081671,\n",
       "        0.00095913, 0.00072248, 0.00054766, 0.00051001, 0.00054376,\n",
       "        0.00054765, 0.00082172, 0.0009537 , 0.00093456, 0.00074332,\n",
       "        0.0007379 , 0.00087473, 0.000875  , 0.00083977, 0.00060412,\n",
       "        0.00107225, 0.00086192, 0.00107515, 0.00109529, 0.00070911,\n",
       "        0.00123408, 0.00075205, 0.00060085, 0.00070017, 0.00066186,\n",
       "        0.00058118, 0.00111156, 0.00073604, 0.00084513, 0.00074525,\n",
       "        0.00058908, 0.00053114, 0.00061254, 0.00068159, 0.00058552,\n",
       "        0.00090886, 0.00143709, 0.00076659, 0.00099078, 0.00070473,\n",
       "        0.00077277, 0.00077906, 0.0006904 , 0.00068032, 0.00125604,\n",
       "        0.00072447, 0.00100348, 0.00064423, 0.000682  , 0.0005842 ,\n",
       "        0.00066563, 0.00061349, 0.00049266, 0.00058978, 0.00062633,\n",
       "        0.00061461, 0.00047399, 0.00059301, 0.00043399, 0.00042724,\n",
       "        0.00040001, 0.0004491 , 0.00043366, 0.00042981, 0.00040477,\n",
       "        0.00031782, 0.00035348, 0.00039801, 0.00042242, 0.00039733,\n",
       "        0.00034712, 0.0003711 , 0.00035419, 0.00030181, 0.00034028,\n",
       "        0.00038139, 0.00056567, 0.0004364 , 0.00045302, 0.00051901,\n",
       "        0.00033408, 0.00052122, 0.0006659 , 0.00045435, 0.00046535,\n",
       "        0.00053793, 0.00040783, 0.0003979 , 0.00047085, 0.00045134,\n",
       "        0.00042027, 0.00069938, 0.00100579, 0.00165878, 0.00092991,\n",
       "        0.00119498, 0.00074254, 0.00087122, 0.00096449, 0.00073074,\n",
       "        0.00057339, 0.00051367, 0.00059485, 0.00046713, 0.00040506,\n",
       "        0.00037552, 0.00056948, 0.00066212, 0.00041204, 0.00045572,\n",
       "        0.00054073, 0.00046112, 0.00060589, 0.00050151, 0.00037051,\n",
       "        0.00031279, 0.00034358, 0.00085297, 0.00041806, 0.00040469,\n",
       "        0.00037144, 0.00035418, 0.00041913, 0.0005688 , 0.00059047,\n",
       "        0.00051729, 0.00097065, 0.00055646, 0.00050988, 0.00054019,\n",
       "        0.00069118, 0.00094476, 0.00057374, 0.00047477, 0.00052554,\n",
       "        0.00048185, 0.0004177 , 0.00067407, 0.00048664, 0.00066987,\n",
       "        0.00059837, 0.00047107, 0.00037577, 0.00037434, 0.00040716,\n",
       "        0.00034513, 0.00058423, 0.00035056, 0.00034914, 0.00031248,\n",
       "        0.00027701, 0.00027545, 0.00038175, 0.00026525, 0.00027847,\n",
       "        0.00023617, 0.0001935 , 0.00034308, 0.0003261 , 0.0002744 ,\n",
       "        0.00034833, 0.00039192, 0.00032491, 0.0003277 , 0.00032132,\n",
       "        0.00027286, 0.00032874, 0.00032982, 0.00032488, 0.00030024,\n",
       "        0.0002373 , 0.00023018, 0.00029448, 0.00023605, 0.00031325,\n",
       "        0.00020394, 0.00017805, 0.00015142, 0.00015647, 0.00028721,\n",
       "        0.00021207, 0.00022357, 0.00026738, 0.0002201 , 0.00020113,\n",
       "        0.000201  , 0.00029829, 0.0002278 , 0.00020989, 0.00037004,\n",
       "        0.00024128]),\n",
       " ('HAR+I',\n",
       "  'Rolling'): array([0.00051718, 0.00053353, 0.00064974, 0.00090883, 0.00057868,\n",
       "        0.00044662, 0.00051085, 0.00055666, 0.00040561, 0.00050365,\n",
       "        0.00032714, 0.00033229, 0.00052776, 0.00046011, 0.00031238,\n",
       "        0.00027791, 0.00027482, 0.00035547, 0.00029915, 0.00031971,\n",
       "        0.00042274, 0.00027284, 0.00028296, 0.00024003, 0.00034251,\n",
       "        0.00026791, 0.00042251, 0.00027108, 0.00030207, 0.00050212,\n",
       "        0.00035281, 0.000386  , 0.00029561, 0.00032118, 0.00037086,\n",
       "        0.00026377, 0.00025493, 0.00030747, 0.00040332, 0.00062642,\n",
       "        0.00058506, 0.00057473, 0.00068264, 0.00064982, 0.00080971,\n",
       "        0.00171999, 0.00082778, 0.00082362, 0.00085717, 0.00073218,\n",
       "        0.00068586, 0.00081046, 0.00091186, 0.00069219, 0.00076691,\n",
       "        0.00066883, 0.00124933, 0.0007452 , 0.00078965, 0.00219017,\n",
       "        0.00147849, 0.00116731, 0.00081565, 0.00098413, 0.00121915,\n",
       "        0.00080141, 0.00122181, 0.00081173, 0.00087527, 0.00086687,\n",
       "        0.00108259, 0.00086609, 0.00070941, 0.00062094, 0.00114971,\n",
       "        0.00082048, 0.00102805, 0.001118  , 0.00096656, 0.0008675 ,\n",
       "        0.00095542, 0.00082303, 0.00068591, 0.00084102, 0.00092263,\n",
       "        0.0006154 , 0.00059792, 0.0005401 , 0.00074272, 0.00058706,\n",
       "        0.0008736 , 0.00059465, 0.00063434, 0.0006285 , 0.00058653,\n",
       "        0.00163823, 0.00147387, 0.00280955, 0.00163764, 0.00238018,\n",
       "        0.00255113, 0.00215481, 0.00297099, 0.00216816, 0.00245441,\n",
       "        0.00206378, 0.00247886, 0.00438652, 0.00457832, 0.00387079,\n",
       "        0.00315651, 0.00268763, 0.00321692, 0.00212512, 0.00297693,\n",
       "        0.0033576 , 0.00205354, 0.0024703 , 0.00179749, 0.00145174,\n",
       "        0.00163786, 0.00133016, 0.00132756, 0.00133999, 0.0011215 ,\n",
       "        0.00189497, 0.00143187, 0.001082  , 0.00097398, 0.00083498,\n",
       "        0.00115519, 0.00078338, 0.00087535, 0.00065985, 0.0006078 ,\n",
       "        0.00061985, 0.00108175, 0.00085478, 0.00066122, 0.00076875,\n",
       "        0.00064945, 0.00045269, 0.00055091, 0.00052883, 0.0007617 ,\n",
       "        0.00101658, 0.00085569, 0.00060717, 0.00057261, 0.00048719,\n",
       "        0.00071411, 0.00112645, 0.00058463, 0.00078127, 0.00085337,\n",
       "        0.00072596, 0.0006567 , 0.00069808, 0.00074145, 0.00114273,\n",
       "        0.00152591, 0.00113718, 0.00147127, 0.00093308, 0.00095565,\n",
       "        0.00067775, 0.0009379 , 0.00089218, 0.00068113, 0.00057964,\n",
       "        0.00058285, 0.00074393, 0.0010609 , 0.00085156, 0.00073991,\n",
       "        0.0009769 , 0.00115728, 0.00091497, 0.00093729, 0.00199698,\n",
       "        0.00208307, 0.00192166, 0.00143106, 0.00133656, 0.0011285 ,\n",
       "        0.00166438, 0.00159603, 0.00093234, 0.00070176, 0.0008699 ,\n",
       "        0.00060429, 0.0006779 , 0.0006105 , 0.0006726 , 0.00062497,\n",
       "        0.00058762, 0.00052769, 0.00053398, 0.00055513, 0.0007126 ,\n",
       "        0.00049624, 0.00109819, 0.00143894, 0.00079856, 0.00155067,\n",
       "        0.00247008, 0.00144122, 0.00143157, 0.00199296, 0.00122608,\n",
       "        0.00136306, 0.00183963, 0.00199382, 0.00167162, 0.00182266,\n",
       "        0.00156944, 0.001488  , 0.00166895, 0.00352491, 0.00305927,\n",
       "        0.00252388, 0.00302103, 0.0030038 , 0.00369223, 0.00293694,\n",
       "        0.00193361, 0.00272803, 0.00381219, 0.00231844, 0.00243567,\n",
       "        0.00177789, 0.00206773, 0.00177291, 0.00259164, 0.00139164,\n",
       "        0.00262998, 0.0014588 , 0.00218071, 0.00162756, 0.00128262,\n",
       "        0.00116185, 0.00172935, 0.00133009, 0.00182846, 0.00147262,\n",
       "        0.00368453, 0.00154941, 0.00155752, 0.00376386, 0.00239775,\n",
       "        0.00204708, 0.00283671, 0.00416691, 0.0038754 , 0.00321349,\n",
       "        0.00516724, 0.00857816, 0.00795254, 0.0053844 , 0.00549337,\n",
       "        0.00370026, 0.00404045, 0.00273702, 0.0027582 , 0.00860151,\n",
       "        0.00436727, 0.00996264, 0.00917622, 0.00715278, 0.00737124,\n",
       "        0.00682472, 0.0072133 , 0.00490084, 0.0041396 , 0.00261361,\n",
       "        0.00326417, 0.00386568, 0.00275642, 0.00153219, 0.00175466,\n",
       "        0.00143307, 0.00099167, 0.0012499 , 0.00099063, 0.00358693,\n",
       "        0.00227365, 0.002394  , 0.00244825, 0.00152266, 0.00103212,\n",
       "        0.00125954, 0.00106895, 0.00119044, 0.0010265 , 0.00075752,\n",
       "        0.00082229, 0.00069784, 0.00124065, 0.00068208, 0.00121773,\n",
       "        0.00107178, 0.00071468, 0.00126814, 0.00105893, 0.00071396,\n",
       "        0.00212722, 0.00148567, 0.00207718, 0.00218523, 0.00232094,\n",
       "        0.00149624, 0.00185531, 0.00279093, 0.00294179, 0.00191221,\n",
       "        0.00201513, 0.00265331, 0.00280574, 0.00486329, 0.00355329,\n",
       "        0.00313267, 0.00141889, 0.00250817, 0.00187999, 0.00282415,\n",
       "        0.0032191 , 0.0048549 , 0.00348933, 0.00445068, 0.00393769,\n",
       "        0.00195672, 0.00199732, 0.00173938, 0.0016749 , 0.00137057,\n",
       "        0.00186284, 0.00163519, 0.00103149, 0.00089996, 0.00150902,\n",
       "        0.0014466 , 0.00120255, 0.00104546, 0.00093806, 0.00121971,\n",
       "        0.00113527, 0.00226763, 0.00135983, 0.00160537, 0.00231591,\n",
       "        0.00141372, 0.00125246, 0.00153928, 0.00162864, 0.0011994 ,\n",
       "        0.00485156, 0.00352539, 0.0057384 , 0.00319151, 0.0049831 ,\n",
       "        0.00476841, 0.00438832, 0.0030634 , 0.00248781, 0.00255697,\n",
       "        0.00225727, 0.00204999, 0.00165722, 0.00147461, 0.00183713,\n",
       "        0.00145093, 0.00178633, 0.00150469, 0.00092911, 0.00182017,\n",
       "        0.00169505, 0.00147225, 0.00106824, 0.00112301, 0.00090335,\n",
       "        0.00130244, 0.00085678, 0.0008458 , 0.00085948, 0.00063062,\n",
       "        0.00071679, 0.00062407, 0.0007282 , 0.00179039, 0.00269528,\n",
       "        0.00171324, 0.0011192 , 0.00125649, 0.00104629, 0.00112839,\n",
       "        0.00098011, 0.0010741 , 0.00109224, 0.0008082 , 0.00071905,\n",
       "        0.00073692, 0.00084315, 0.00073039, 0.00084506, 0.0008729 ,\n",
       "        0.00085973, 0.00061831, 0.00059745, 0.00066274, 0.00077469,\n",
       "        0.00057199, 0.00077087, 0.00094686, 0.0007504 , 0.00066345,\n",
       "        0.00066255, 0.00074879, 0.00098896, 0.00109693, 0.00077168,\n",
       "        0.00095461, 0.00073618, 0.00068717, 0.00136792, 0.00100472,\n",
       "        0.00098557, 0.00075344, 0.00109462, 0.00076744, 0.00079997,\n",
       "        0.0006104 , 0.00062194, 0.00058333, 0.00088463, 0.00076473,\n",
       "        0.0006547 , 0.00069391, 0.00080252, 0.0008093 , 0.00103077,\n",
       "        0.00092481, 0.0007999 , 0.00068935, 0.0006817 , 0.00084498,\n",
       "        0.00102471, 0.00102722, 0.00084531, 0.00086799, 0.00073032,\n",
       "        0.00060574, 0.00125061, 0.00104933, 0.00093956, 0.001012  ,\n",
       "        0.00077489, 0.0005564 , 0.00063232, 0.00074922, 0.00080441,\n",
       "        0.00069864, 0.0007199 , 0.00059155, 0.00073834, 0.00067493,\n",
       "        0.00063896, 0.00187779, 0.00109663, 0.00101739, 0.0009188 ,\n",
       "        0.00064423, 0.00073609, 0.00082055, 0.0006212 , 0.00061987,\n",
       "        0.00057495, 0.00060841, 0.00050124, 0.00050629, 0.00085384,\n",
       "        0.00102797, 0.00071352, 0.00055871, 0.00068652, 0.00054502,\n",
       "        0.00073993, 0.0006057 , 0.00059997, 0.00069758, 0.00072476,\n",
       "        0.00067857, 0.00072881, 0.00061458, 0.00071331, 0.00076928,\n",
       "        0.00084179, 0.00069446, 0.00074567, 0.00049999, 0.00080432,\n",
       "        0.00065615, 0.00140275, 0.00172602, 0.00102053, 0.00164366,\n",
       "        0.00086021, 0.00099959, 0.0009163 , 0.00093456, 0.00082243,\n",
       "        0.0007837 , 0.00061226, 0.00051475, 0.00050481, 0.00051061,\n",
       "        0.00061595, 0.00068922, 0.00114096, 0.00087515, 0.00065127,\n",
       "        0.00064133, 0.00088861, 0.00075712, 0.00102927, 0.00056934,\n",
       "        0.0007823 , 0.00076057, 0.00089091, 0.00094405, 0.00068958,\n",
       "        0.00176161, 0.00069253, 0.00059508, 0.00070058, 0.00059105,\n",
       "        0.00055811, 0.00094211, 0.00061395, 0.00084418, 0.00060741,\n",
       "        0.00061773, 0.00050238, 0.00059574, 0.00059864, 0.00052853,\n",
       "        0.00074028, 0.0009633 , 0.00078023, 0.0009821 , 0.00068363,\n",
       "        0.00077909, 0.00071345, 0.00061755, 0.00078   , 0.00093484,\n",
       "        0.00069634, 0.00093172, 0.0005747 , 0.00063009, 0.00055721,\n",
       "        0.00063078, 0.00054243, 0.00047159, 0.00066253, 0.00060318,\n",
       "        0.00074741, 0.00046637, 0.00061969, 0.00038901, 0.00038893,\n",
       "        0.00038155, 0.00040887, 0.00039413, 0.00049844, 0.00042422,\n",
       "        0.00032279, 0.00037369, 0.00044221, 0.0003761 , 0.00041556,\n",
       "        0.00032485, 0.00035288, 0.00032571, 0.00030681, 0.00039442,\n",
       "        0.00037865, 0.0009376 , 0.00040099, 0.00036416, 0.00083225,\n",
       "        0.00032337, 0.00049762, 0.00058858, 0.00043625, 0.00046963,\n",
       "        0.00058989, 0.00041431, 0.00034533, 0.00042321, 0.0004408 ,\n",
       "        0.00043244, 0.00064718, 0.00143148, 0.00125827, 0.00095872,\n",
       "        0.00115207, 0.00078684, 0.0011178 , 0.00137942, 0.00085113,\n",
       "        0.00064062, 0.00056538, 0.00071045, 0.00042828, 0.00046154,\n",
       "        0.00037963, 0.00056446, 0.00051235, 0.00041167, 0.00042865,\n",
       "        0.00055216, 0.00046838, 0.00054671, 0.00061984, 0.00035415,\n",
       "        0.00031697, 0.00035078, 0.00071915, 0.00041642, 0.00043311,\n",
       "        0.00033975, 0.00032323, 0.00038732, 0.00062602, 0.00065349,\n",
       "        0.00048758, 0.00093442, 0.00051787, 0.0004577 , 0.00061149,\n",
       "        0.00078653, 0.00068196, 0.00049778, 0.00045808, 0.00055025,\n",
       "        0.00046314, 0.00039798, 0.00094162, 0.00042939, 0.00058225,\n",
       "        0.00054068, 0.00042776, 0.00034301, 0.00039312, 0.0004528 ,\n",
       "        0.00036463, 0.00057037, 0.00030927, 0.00033507, 0.00031204,\n",
       "        0.00026122, 0.00026358, 0.00031501, 0.00024592, 0.00024122,\n",
       "        0.00020809, 0.00019178, 0.00055025, 0.00032045, 0.0002552 ,\n",
       "        0.00039744, 0.00032894, 0.00029334, 0.00031753, 0.00036936,\n",
       "        0.00028256, 0.00030426, 0.00031191, 0.00028329, 0.00031472,\n",
       "        0.00022395, 0.0002473 , 0.00029392, 0.00022692, 0.00030232,\n",
       "        0.00018498, 0.00017325, 0.00015307, 0.00017223, 0.00030613,\n",
       "        0.00015745, 0.00017992, 0.00031122, 0.00021703, 0.00018504,\n",
       "        0.0001911 , 0.00031374, 0.00027349, 0.00019551, 0.00031681,\n",
       "        0.00022145])}"
      ]
     },
     "execution_count": 430,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "predictions_dict['bnbbtc']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "353310f7-40fc-4c19-bc4e-4cc0b7c09b06",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "12b70364-cebf-4c1d-97f7-4e7b67fa1e0f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done with AR(1).\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\wang-\\OneDrive\\Term_2\\Dissertation\\experiments\\scratches\\utils\\q_likelihood_oos_details.py:13: RuntimeWarning: invalid value encountered in log\n",
      "  q_likelihood = ratio - np.log(ratio) - 1\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with bnbbtc, which is index 0.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with adabtc, which is index 1.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with ethbtc, which is index 2.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with batbtc, which is index 3.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with btcusdc, which is index 4.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with btcusdt, which is index 5.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with neobtc, which is index 6.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with btctusd, which is index 7.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with etcbtc, which is index 8.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with ltcbtc, which is index 9.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with xrpbtc, which is index 10.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with trxbtc, which is index 11.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with eosbtc, which is index 12.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with xmrbtc, which is index 13.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with zecbtc, which is index 14.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with qtumbtc, which is index 15.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with dcrbtc, which is index 16.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with xlmbtc, which is index 17.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with iotabtc, which is index 18.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with manabtc, which is index 19.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with dashbtc, which is index 20.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with zrxbtc, which is index 21.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Calendar.\n",
      "Done with AR(1).\n",
      "Done with AR(30).\n",
      "Done with HAR.\n",
      "Done with HAR+I.\n",
      "Done with Rolling.\n",
      "Done with repbtc, which is index 22.\n"
     ]
    }
   ],
   "source": [
    "# store LOSSES\n",
    "summary_details_dict = dict()\n",
    "for i, symbol in enumerate(symbols_list):\n",
    "    oos_results = make_empty_oos_results()\n",
    "    summary_details = {key: None for key in product(model_names, data_type_names, summary_names)}\n",
    "    for data_type in ['Calendar', 'Rolling']:\n",
    "        for model_name, function in model_name_to_function.items():\n",
    "            res, y, X = function(\n",
    "                input_folder_name=input_folder_name, \n",
    "                symbol=symbol, \n",
    "                data_type=data_type\n",
    "            )\n",
    "            window_length = 365\n",
    "            beta_daily, y_daily, X_daily = daily_rolling_OLS(y, X, window_length, data_type)\n",
    "            \n",
    "            q_likelihoods = q_likelihood_oos_details(beta_daily, y_daily, X_daily, window_length)\n",
    "            squared_errors = squared_errors_oos(beta_daily, y_daily, X_daily, window_length)\n",
    "            \n",
    "            summary_details[model_name, data_type, 'QLIKE'] = q_likelihoods\n",
    "            summary_details[model_name, data_type, 'MSE'] = squared_errors\n",
    "            \n",
    "            summaries = [\n",
    "                r_squared_oos(beta_daily, y_daily, X_daily, window_length),\n",
    "                squared_errors.mean(), # we should fix these to be nanmeans\n",
    "                q_likelihoods.mean()\n",
    "            ]\n",
    "            summary_names = ['$R^2$', 'MSE', 'QLIKE']\n",
    "            for summary_name, summary in zip(summary_names, summaries):\n",
    "                oos_results[model_name, data_type].loc[summary_name] = summary\n",
    "            print(f\"Done with {model_name}.\")\n",
    "        print(f\"Done with {data_type}.\")\n",
    "    oos_results_dict[symbol] = oos_results\n",
    "    summary_details_dict[symbol] = summary_details # we have this\n",
    "    print(f\"Done with {symbol}, which is index {i}.\")\n",
    "oos_results_mean = pd.concat(oos_results_dict).groupby(level=1).mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "fc15ca07-0eae-45b6-ad0b-e91073b29a73",
   "metadata": {},
   "outputs": [],
   "source": [
    "pd.set_option('display.precision', 9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "efd2610e-824c-4819-b078-ad16efbfe4ed",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(1)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(30)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR+I</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Data</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>$R^2$</th>\n",
       "      <td>0.555393165</td>\n",
       "      <td>0.562172869</td>\n",
       "      <td>0.517090336</td>\n",
       "      <td>0.597589017</td>\n",
       "      <td>0.573172815</td>\n",
       "      <td>0.585285049</td>\n",
       "      <td>0.613970806</td>\n",
       "      <td>0.632705558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MSE</th>\n",
       "      <td>0.000002288</td>\n",
       "      <td>0.000002243</td>\n",
       "      <td>0.000002567</td>\n",
       "      <td>0.000002098</td>\n",
       "      <td>0.000002274</td>\n",
       "      <td>0.000002199</td>\n",
       "      <td>0.000001954</td>\n",
       "      <td>0.000001869</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QLIKE</th>\n",
       "      <td>0.156234390</td>\n",
       "      <td>0.155706945</td>\n",
       "      <td>0.162725042</td>\n",
       "      <td>0.141415095</td>\n",
       "      <td>0.143821868</td>\n",
       "      <td>0.141886788</td>\n",
       "      <td>0.135665942</td>\n",
       "      <td>0.133970752</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Model        AR(1)                    AR(30)                       HAR  \\\n",
       "Data      Calendar      Rolling     Calendar      Rolling     Calendar   \n",
       "$R^2$  0.555393165  0.562172869  0.517090336  0.597589017  0.573172815   \n",
       "MSE    0.000002288  0.000002243  0.000002567  0.000002098  0.000002274   \n",
       "QLIKE  0.156234390  0.155706945  0.162725042  0.141415095  0.143821868   \n",
       "\n",
       "Model                     HAR+I               \n",
       "Data       Rolling     Calendar      Rolling  \n",
       "$R^2$  0.585285049  0.613970806  0.632705558  \n",
       "MSE    0.000002199  0.000001954  0.000001869  \n",
       "QLIKE  0.141886788  0.135665942  0.133970752  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oos_results_mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "69ed9c86-7a03-44eb-b4e6-5e27eb934423",
   "metadata": {},
   "outputs": [],
   "source": [
    "def make_empty_hypothesis_results():\n",
    "    model_names = ['AR(1)', 'AR(30)', 'HAR', 'HAR+I']\n",
    "    # data_type_names = ['Calendar', 'Rolling']\n",
    "\n",
    "    summary_names = ['MSE', 'QLIKE']\n",
    "\n",
    "    column_iterable = [model_names]\n",
    "    column_iterable_name = [\"Model\"]\n",
    "    column_index = pd.MultiIndex.from_product(column_iterable, names=column_iterable_name)\n",
    "\n",
    "    row_index = summary_names\n",
    "\n",
    "    return pd.DataFrame(np.full([2, 4], np.nan), index=row_index, columns=column_index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "b53ad2ac-a6a0-49c2-8e33-8f1980931e18",
   "metadata": {},
   "outputs": [],
   "source": [
    "hypothesis_results = make_empty_hypothesis_results()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "041d2330-bcac-4e72-8caf-09f98cb02cf7",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th>AR(1)</th>\n",
       "      <th>AR(30)</th>\n",
       "      <th>HAR</th>\n",
       "      <th>HAR+I</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MSE</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QLIKE</th>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Model AR(1) AR(30) HAR HAR+I\n",
       "MSE     NaN    NaN NaN   NaN\n",
       "QLIKE   NaN    NaN NaN   NaN"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hypothesis_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "57196b10-d4a9-4871-b57d-080b82c7c957",
   "metadata": {},
   "outputs": [],
   "source": [
    "summary_names = ['MSE', 'QLIKE']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ea350946-cd1b-473f-a75e-b1a471a7083a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c50263fb-eb69-4615-b036-408df303a131",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_name_to_calendar"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "id": "190aafe4-d857-4870-9730-4542aa672669",
   "metadata": {},
   "outputs": [],
   "source": [
    "def summary_difference_averaged(summary_details_dict, model_name, summary_name):\n",
    "    list_of_summary_calendar = []\n",
    "    list_of_summary_rolling = []\n",
    "    for symbol in symbols_list:\n",
    "        summary_details = summary_details_dict[symbol]\n",
    "        summary_calendar = summary_details[model_name, 'Calendar', summary_name].tolist()\n",
    "        summary_rolling = summary_details[model_name, 'Rolling', summary_name].tolist()\n",
    "        list_of_summary_calendar.append(summary_calendar)\n",
    "        list_of_summary_rolling.append(summary_rolling)\n",
    "    list_of_summary_calendar_np = np.array(list_of_summary_calendar)\n",
    "    list_of_summary_rolling_np = np.array(list_of_summary_rolling)\n",
    "    difference_averaged = (list_of_summary_calendar_np - list_of_summary_rolling_np).mean(axis=0)\n",
    "    return difference_averaged"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "id": "8ea3ea19-0527-40ed-bcef-52a75c3ffb53",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_name = 'AR(30)'\n",
    "model_name = 'HAR+I'\n",
    "summary_name = 'QLIKE'\n",
    "a = summary_difference_averaged(summary_details_dict, model_name, summary_name)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "id": "1c9519e7-8d1d-4a62-b0c9-3dfe73e98da3",
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_newey_west_error(a, maxlags):\n",
    "    a = a[~ np.isnan(a)]\n",
    "    ones = np.ones_like(a)\n",
    "    res = sm.OLS(a, ones).fit()\n",
    "    error = res.get_robustcov_results(cov_type='HAC', maxlags=maxlags).bse[0]\n",
    "    return error"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "id": "b869c2a2-c151-48e8-95f2-b6e1c071a793",
   "metadata": {},
   "outputs": [],
   "source": [
    "model_name = 'AR(1)'\n",
    "summary_name = 'QLIKE'\n",
    "a = summary_difference_averaged(summary_details_dict, model_name, summary_name)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "id": "270eaf76-24f1-4b88-b885-5edf634bc396",
   "metadata": {},
   "outputs": [],
   "source": [
    "auto = [autocovariance(a, len(a), lag, a.mean()) for lag in range(0, 30)]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 218,
   "id": "5ada5588-f479-4dd1-82f6-770ae6d768cc",
   "metadata": {},
   "outputs": [],
   "source": [
    "b = np.arange(0, 30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "id": "f25adb60-ddd4-43db-b75c-25bf85492018",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x21cac75eec8>]"
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(b, auto)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 302,
   "id": "26a2685e-6c67-4d13-bf68-9671b1bd9f56",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "c:\\users\\wang-\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\pandas\\core\\indexing.py:1732: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  self._setitem_single_block(indexer, value, name)\n",
      "c:\\users\\wang-\\appdata\\local\\programs\\python\\python37\\lib\\site-packages\\pandas\\core\\indexing.py:723: SettingWithCopyWarning: \n",
      "A value is trying to be set on a copy of a slice from a DataFrame\n",
      "\n",
      "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
      "  iloc._setitem_with_indexer(indexer, value, self.name)\n"
     ]
    }
   ],
   "source": [
    "hypothesis_results = make_empty_hypothesis_results()\n",
    "hypothesis_p = make_empty_hypothesis_results()\n",
    "summary_names = ['MSE', 'QLIKE']\n",
    "for model_name in model_names:\n",
    "    for summary_name in summary_names:\n",
    "        difference_averaged = summary_difference_averaged(summary_details_dict, model_name, summary_name)\n",
    "        mean = np.nanmean(difference_averaged)\n",
    "        variance_d = np.nanvar(difference_averaged) / len(difference_averaged)\n",
    "        # variance_d = diebold_mariano_variance_estimate(difference_averaged, 1)\n",
    "        variance_d = get_newey_west_error(difference_averaged, 5) ** 2\n",
    "        dm_stat = normalised_stat(mean, variance_d)\n",
    "        p = normal_p_value(dm_stat)\n",
    "        hypothesis_results[model_name].loc[summary_name] = dm_stat\n",
    "        hypothesis_p[model_name].loc[summary_name] = p\n",
    "        # print(f\"Done with {model_name} with {summary_name}.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 303,
   "id": "0f7513b2-200f-4b50-8aa6-b22a50c883b2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th>AR(1)</th>\n",
       "      <th>AR(30)</th>\n",
       "      <th>HAR</th>\n",
       "      <th>HAR+I</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MSE</th>\n",
       "      <td>4.020825865</td>\n",
       "      <td>7.501772076</td>\n",
       "      <td>4.989002519</td>\n",
       "      <td>2.722700675</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QLIKE</th>\n",
       "      <td>2.733009938</td>\n",
       "      <td>4.293141713</td>\n",
       "      <td>9.485333488</td>\n",
       "      <td>5.600149225</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Model        AR(1)       AR(30)          HAR        HAR+I\n",
       "MSE    4.020825865  7.501772076  4.989002519  2.722700675\n",
       "QLIKE  2.733009938  4.293141713  9.485333488  5.600149225"
      ]
     },
     "execution_count": 303,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hypothesis_results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 314,
   "id": "b7a7415e-f553-4df5-8149-b20c5fa508c1",
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'results/hypothesis_results.tex'\n",
    "\n",
    "precision_digits = 3\n",
    "format_type='g'\n",
    "alternate = '#'\n",
    "my_formatter = f\"{{:{alternate}.{precision_digits}{format_type}}}\".format\n",
    "\n",
    "\n",
    "hypothesis_results.to_latex(\n",
    "    file_name,\n",
    "    escape=False, \n",
    "    float_format=my_formatter, \n",
    "    multirow=True, \n",
    "    na_rep=\"\", \n",
    "    multicolumn_format=\"c\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "854c0c24-70c6-442b-8b83-8458262582f5",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "id": "95ee18e6-a52a-48db-a68f-4570e4f472b1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(1)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(30)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR+I</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Data</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>$R^2$</th>\n",
       "      <td>0.555393165</td>\n",
       "      <td>0.562172869</td>\n",
       "      <td>0.517090336</td>\n",
       "      <td>0.597589017</td>\n",
       "      <td>0.573172815</td>\n",
       "      <td>0.585285049</td>\n",
       "      <td>0.613970806</td>\n",
       "      <td>0.632705558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MSE</th>\n",
       "      <td>0.000002288</td>\n",
       "      <td>0.000002243</td>\n",
       "      <td>0.000002567</td>\n",
       "      <td>0.000002098</td>\n",
       "      <td>0.000002274</td>\n",
       "      <td>0.000002199</td>\n",
       "      <td>0.000001954</td>\n",
       "      <td>0.000001869</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QLIKE</th>\n",
       "      <td>0.156234390</td>\n",
       "      <td>0.155706945</td>\n",
       "      <td>0.162725042</td>\n",
       "      <td>0.141415095</td>\n",
       "      <td>0.143821868</td>\n",
       "      <td>0.141886788</td>\n",
       "      <td>0.135665942</td>\n",
       "      <td>0.133970752</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Model        AR(1)                    AR(30)                       HAR  \\\n",
       "Data      Calendar      Rolling     Calendar      Rolling     Calendar   \n",
       "$R^2$  0.555393165  0.562172869  0.517090336  0.597589017  0.573172815   \n",
       "MSE    0.000002288  0.000002243  0.000002567  0.000002098  0.000002274   \n",
       "QLIKE  0.156234390  0.155706945  0.162725042  0.141415095  0.143821868   \n",
       "\n",
       "Model                     HAR+I               \n",
       "Data       Rolling     Calendar      Rolling  \n",
       "$R^2$  0.585285049  0.613970806  0.632705558  \n",
       "MSE    0.000002199  0.000001954  0.000001869  \n",
       "QLIKE  0.141886788  0.135665942  0.133970752  "
      ]
     },
     "execution_count": 224,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oos_results_mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 299,
   "id": "37d4810c-1c1b-4343-9db5-d99a585a5272",
   "metadata": {},
   "outputs": [],
   "source": [
    "precision_digits = 3\n",
    "format_type='g'\n",
    "my_formatter = f\"{{:.{precision_digits}{format_type}}}\".format"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 298,
   "id": "8ac6d352-af37-47cf-90ba-7a7ece575300",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(1)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(30)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR+I</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Data</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>$R^2$</th>\n",
       "      <td>0.555393165</td>\n",
       "      <td>0.562172869</td>\n",
       "      <td>0.517090336</td>\n",
       "      <td>0.597589017</td>\n",
       "      <td>0.573172815</td>\n",
       "      <td>0.585285049</td>\n",
       "      <td>0.613970806</td>\n",
       "      <td>0.632705558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MSE</th>\n",
       "      <td>0.000002288</td>\n",
       "      <td>0.000002243</td>\n",
       "      <td>0.000002567</td>\n",
       "      <td>0.000002098</td>\n",
       "      <td>0.000002274</td>\n",
       "      <td>0.000002199</td>\n",
       "      <td>0.000001954</td>\n",
       "      <td>0.000001869</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QLIKE</th>\n",
       "      <td>0.156234390</td>\n",
       "      <td>0.155706945</td>\n",
       "      <td>0.162725042</td>\n",
       "      <td>0.141415095</td>\n",
       "      <td>0.143821868</td>\n",
       "      <td>0.141886788</td>\n",
       "      <td>0.135665942</td>\n",
       "      <td>0.133970752</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Model        AR(1)                    AR(30)                       HAR  \\\n",
       "Data      Calendar      Rolling     Calendar      Rolling     Calendar   \n",
       "$R^2$  0.555393165  0.562172869  0.517090336  0.597589017  0.573172815   \n",
       "MSE    0.000002288  0.000002243  0.000002567  0.000002098  0.000002274   \n",
       "QLIKE  0.156234390  0.155706945  0.162725042  0.141415095  0.143821868   \n",
       "\n",
       "Model                     HAR+I               \n",
       "Data       Rolling     Calendar      Rolling  \n",
       "$R^2$  0.585285049  0.613970806  0.632705558  \n",
       "MSE    0.000002199  0.000001954  0.000001869  \n",
       "QLIKE  0.141886788  0.135665942  0.133970752  "
      ]
     },
     "execution_count": 298,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oos_results_mean"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 315,
   "id": "1ee46876-b40f-4f31-bf0c-e1f125fba228",
   "metadata": {},
   "outputs": [],
   "source": [
    "file_name = 'results/out_of_sample_summary.tex'\n",
    "\n",
    "precision_digits = 3\n",
    "format_type='g'\n",
    "alternate = '#'\n",
    "my_formatter = f\"{{:{alternate}.{precision_digits}{format_type}}}\".format\n",
    "\n",
    "oos_results_mean.to_latex(\n",
    "    file_name,\n",
    "    escape=False, \n",
    "    float_format=my_formatter, \n",
    "    multirow=True, \n",
    "    na_rep=\"\", \n",
    "    multicolumn_format=\"c\",\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a476707a-f99c-4295-9141-0a75f988bc70",
   "metadata": {},
   "outputs": [],
   "source": [
    "# code to fix nans"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 345,
   "id": "c3b1a6eb-8724-4f00-86f1-7c306da21b40",
   "metadata": {},
   "outputs": [],
   "source": [
    "for i, symbol in enumerate(symbols_list):\n",
    "    oos_results = oos_results_dict[symbol]\n",
    "    for data_type in ['Calendar', 'Rolling']:\n",
    "        for model_name, function in model_name_to_function.items():\n",
    "            # check if QLIKE and MSE are accurate\n",
    "            for summary_name in ['MSE', 'QLIKE']:\n",
    "                details = summary_details_dict[symbol][(model_name, data_type, summary_name)]\n",
    "                details_average = np.nanmean(details)\n",
    "                recorded_value = oos_results_dict[symbol][model_name, data_type].loc[summary_name]\n",
    "                try:\n",
    "                    assert details_average == recorded_value\n",
    "                except AssertionError:\n",
    "                    oos_results_dict[symbol][model_name, data_type].loc[summary_name] = details_average"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 349,
   "id": "8d53a1ec-3666-4af3-84ae-78d011a0384a",
   "metadata": {},
   "outputs": [],
   "source": [
    "# now make heatmap"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 350,
   "id": "49a591b2-8c2e-4907-943a-d580a4d43c51",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead tr th {\n",
       "        text-align: left;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr>\n",
       "      <th>Model</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(1)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">AR(30)</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR</th>\n",
       "      <th colspan=\"2\" halign=\"left\">HAR+I</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Data</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "      <th>Calendar</th>\n",
       "      <th>Rolling</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>$R^2$</th>\n",
       "      <td>0.684850869</td>\n",
       "      <td>0.693576768</td>\n",
       "      <td>0.622131743</td>\n",
       "      <td>0.704514844</td>\n",
       "      <td>0.692784615</td>\n",
       "      <td>0.706449237</td>\n",
       "      <td>0.702643035</td>\n",
       "      <td>0.721859714</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MSE</th>\n",
       "      <td>0.000000728</td>\n",
       "      <td>0.000000707</td>\n",
       "      <td>0.000000901</td>\n",
       "      <td>0.000000705</td>\n",
       "      <td>0.000000733</td>\n",
       "      <td>0.000000700</td>\n",
       "      <td>0.000000709</td>\n",
       "      <td>0.000000663</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>QLIKE</th>\n",
       "      <td>0.129213476</td>\n",
       "      <td>0.128466050</td>\n",
       "      <td>0.132168628</td>\n",
       "      <td>0.118811147</td>\n",
       "      <td>0.110365146</td>\n",
       "      <td>0.108656074</td>\n",
       "      <td>0.109987856</td>\n",
       "      <td>0.107097533</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Model        AR(1)                    AR(30)                       HAR  \\\n",
       "Data      Calendar      Rolling     Calendar      Rolling     Calendar   \n",
       "$R^2$  0.684850869  0.693576768  0.622131743  0.704514844  0.692784615   \n",
       "MSE    0.000000728  0.000000707  0.000000901  0.000000705  0.000000733   \n",
       "QLIKE  0.129213476  0.128466050  0.132168628  0.118811147  0.110365146   \n",
       "\n",
       "Model                     HAR+I               \n",
       "Data       Rolling     Calendar      Rolling  \n",
       "$R^2$  0.706449237  0.702643035  0.721859714  \n",
       "MSE    0.000000700  0.000000709  0.000000663  \n",
       "QLIKE  0.108656074  0.109987856  0.107097533  "
      ]
     },
     "execution_count": 350,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "oos_results_dict['bnbbtc']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 351,
   "id": "97530277-0b9f-454e-aaba-1dc5c80d90b2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# one for each of R^2, MSE, and QLIKE"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0cde6292-e38a-4202-9f7e-fa1f8e759948",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 353,
   "id": "5233d88b-eb98-43bb-891c-12711d43735e",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.concat(oos_results_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 365,
   "id": "27b9ddf6-8eb7-4531-a614-2e83969d1afc",
   "metadata": {},
   "outputs": [],
   "source": [
    "df.index.names = ['Symbol', 'Summary']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 398,
   "id": "b3e8794d-c0a7-4b9a-9751-6d8920e8772c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary_to_plot = '$R^2$'\n",
    "df_to_plot = (\n",
    "    df[df.index.get_level_values(1).isin([summary_to_plot])]\n",
    "    .droplevel('Summary', axis=0)\n",
    ")\n",
    "ax = sns.heatmap(df_to_plot, cmap=\"Blues\", yticklabels=True)\n",
    "filename = 'results/r_squared_heatmap.png'\n",
    "dpi = 300\n",
    "bbox_inches = 'tight'\n",
    "\n",
    "ax.figure.savefig(\n",
    "    filename,\n",
    "    dpi=dpi,\n",
    "    bbox_inches=bbox_inches,\n",
    "    pad_inches=pad_inches\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 399,
   "id": "f02bd5c9-a0be-473f-8dcf-31c781945a8f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary_to_plot = '$R^2$'\n",
    "df_to_plot = (\n",
    "    df[df.index.get_level_values(1).isin([summary_to_plot])]\n",
    "    .droplevel('Summary', axis=0)\n",
    "    .drop(index=['dcrbtc', 'repbtc'])\n",
    ")\n",
    "\n",
    "ax = sns.heatmap(df_to_plot, cmap=\"Blues\", yticklabels=True)\n",
    "filename = 'results/r_squared_heatmap_without_outliers.png'\n",
    "dpi = 300\n",
    "bbox_inches = 'tight'\n",
    "\n",
    "ax.figure.savefig(\n",
    "    filename,\n",
    "    dpi=dpi,\n",
    "    bbox_inches=bbox_inches,\n",
    "    pad_inches=pad_inches\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 400,
   "id": "db89e235-d20f-4d69-a8d9-0149eb153733",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary_to_plot = 'MSE'\n",
    "df_to_plot = (\n",
    "    df[df.index.get_level_values(1).isin([summary_to_plot])]\n",
    "    .droplevel('Summary', axis=0)\n",
    ")\n",
    "\n",
    "ax = sns.heatmap(df_to_plot, cmap=\"Blues\", yticklabels=True)\n",
    "filename = 'results/mse_heatmap.png'\n",
    "dpi = 300\n",
    "bbox_inches = 'tight'\n",
    "\n",
    "ax.figure.savefig(\n",
    "    filename,\n",
    "    dpi=dpi,\n",
    "    bbox_inches=bbox_inches,\n",
    "    pad_inches=pad_inches\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 404,
   "id": "1654c47d-a2d2-4302-ad06-a94cdc646201",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary_to_plot = 'MSE'\n",
    "df_to_plot = (\n",
    "    df[df.index.get_level_values(1).isin([summary_to_plot])]\n",
    "    .droplevel('Summary', axis=0)\n",
    "    .drop(index=['manabtc'])\n",
    ")\n",
    "\n",
    "ax = sns.heatmap(df_to_plot, cmap=\"Blues\", yticklabels=True)\n",
    "filename = 'results/mse_heatmap_without_outliers.png'\n",
    "dpi = 300\n",
    "bbox_inches = 'tight'\n",
    "\n",
    "ax.figure.savefig(\n",
    "    filename,\n",
    "    dpi=dpi,\n",
    "    bbox_inches=bbox_inches,\n",
    "    pad_inches=pad_inches\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 407,
   "id": "6f05a959-5e7e-4867-be50-135a8f808968",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary_to_plot = 'QLIKE'\n",
    "df_to_plot = (\n",
    "    df[df.index.get_level_values(1).isin([summary_to_plot])]\n",
    "    .droplevel('Summary', axis=0)\n",
    ")\n",
    "ax = sns.heatmap(df_to_plot, cmap=\"Blues\", yticklabels=True)\n",
    "filename = 'results/qlike_heatmap.png'\n",
    "dpi = 300\n",
    "bbox_inches = 'tight'\n",
    "\n",
    "ax.figure.savefig(\n",
    "    filename,\n",
    "    dpi=dpi,\n",
    "    bbox_inches=bbox_inches,\n",
    "    pad_inches=pad_inches\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 408,
   "id": "659911b8-8f7f-432f-8a46-dad18160f57d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "summary_to_plot = 'QLIKE'\n",
    "df_to_plot = (\n",
    "    df[df.index.get_level_values(1).isin([summary_to_plot])]\n",
    "    .droplevel('Summary', axis=0)\n",
    "    .drop(columns='AR(30)')\n",
    ")\n",
    "\n",
    "ax = sns.heatmap(df_to_plot, cmap=\"Blues\", yticklabels=True)\n",
    "filename = 'results/qlike_heatmap_without_outliers.png'\n",
    "dpi = 300\n",
    "bbox_inches = 'tight'\n",
    "\n",
    "ax.figure.savefig(\n",
    "    filename,\n",
    "    dpi=dpi,\n",
    "    bbox_inches=bbox_inches,\n",
    "    pad_inches=pad_inches\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "884bd5a2-5d27-4c72-8016-e0fc42071bd7",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}