docs: Clarify staterror modifier specification

[Moelf]

clarify and fix typo for Staterror modifier.

Description

The \delta parameter is the absolute error in each bin due to MC Stat at +/- 1 sigma. Thus, the \sigma of the Gaussian constraint term should have \delta / nominal value not the other way around.

And by play around with the logpdf() output, I'm convinced that the input in JSON is \delta rather than \delta^2 (https://scikit-hep.org/pyhf/intro.html#id25)

Checklist Before Requesting Reviewer

• Tests are passing
• "WIP" removed from the title of the pull request
• Selected an Assignee for the PR to be responsible for the log summary

Before Merging

For the PR Assignees:

• Summarize commit messages into a comprehensive review of the PR

* Correct the description of the calculation of the scale of the constraint
to indicate it is calculated from the relative statistical uncertainty from the
channel's event rate.
* Add additional channel subscripts to make the notation be more explicit that
staterror is per-channel.
* Remove typo of additional 'are'.
* Add Jerry Ling to contributor list.

[codecov]

Codecov Report

Merging #1856 (a8eb612) into master (de4ce07) will not change coverage.
The diff coverage is n/a.

❗ Current head a8eb612 differs from pull request most recent head ef98feb. Consider uploading reports for the commit ef98feb to get more accurate results

@@           Coverage Diff           @@
##           master    #1856   +/-   ##
=======================================
  Coverage   98.16%   98.16%           
=======================================
  Files          68       68           
  Lines        4315     4315           
  Branches      725      725           
=======================================
  Hits         4236     4236           
  Misses         46       46           
  Partials       33       33           

Flag	Coverage Δ
contrib	26.44% <0.00%> (ø)
doctest	60.78% <0.00%> (ø)
unittests-3.10	96.03% <0.00%> (ø)
unittests-3.7	96.02% <0.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update b4abe41...ef98feb. Read the comment docs.

[matthewfeickert]

Thanks for the PR @Moelf. We'll try to get to it before the end of the weekend.

As you didn't make a GitHub Issue for this can you give a full summary in the PR body and then also finish off the rest of the CONTRIBUTING guidelines (hint: adding yourself to contributors. 😉)?

[Moelf]

I think the \delta^2 in the Table can be understood as:

1. we take the quadrature sum of \delta_b across all the samples with the same MC error
2. we pass the sqrt(sum) into the staterror constructor.

So I think the table is not wrong.

[matthewfeickert]

Note to myself for when I review this tomorrow: Relevant RTD build section is here: https://pyhf--1856.org.readthedocs.build/en/1856/likelihood.html#mc-statistical-uncertainty-staterror

[matthewfeickert]

@lukasheinrich @kratsg while I think we should maybe discuss more the notational consistency across the docs I don't think that this needs to hold up Jerry's PR. So if we get approvals from both of you then I this LGTM for now.

[matthewfeickert]

@kratsg @lukasheinrich if this can get reviewed today by one (ideally both) of you then I think we can merge this.

[matthewfeickert]

Thanks for your PR @Moelf!