4ku 5.1

4ku 5.1 is essentially the same as 5.0 but contains a bug fix. Use this instead of 5.0.

4ku 5.0

4ku 5.0 contains a bug that can negatively impact strength for multithreaded searches. Do not use this release, use 5.1 instead.

10+0.1s:
Score of 4ku-5.0 vs 4ku-4.0: 6075 - 5287 - 8642  [0.520] 20004
Elo difference: 13.7 +/- 3.4, LOS: 100.0 %, DrawRatio: 43.2 %

40+0.4s:
Score of 4ku-5.0 vs 4ku-4.0: 5760 - 4707 - 9533  [0.526] 20000
Elo difference: 18.3 +/- 3.5, LOS: 100.0 %, DrawRatio: 47.7 %

4ku 4.0

10+0.1s:
Score of 4ku-4.0 vs 4ku-3.1: 1023 - 236 - 741  [0.697] 2000
Elo difference: 144.5 +/- 12.4, LOS: 100.0 %, DrawRatio: 37.0 %

100+1s:
Score of 4ku-4.0 vs 4ku-3.1: 1653 - 252 - 1095  [0.734] 3000
Elo difference: 175.9 +/- 10.3, LOS: 100.0 %, DrawRatio: 36.5 %

4ku 3.1

This version of 4ku is playing in the TCEC S25 Swiss event.

10+0.1s:
Score of 4ku 3.1 vs 4ku 3.0: 833 - 774 - 993  [0.511] 2600
Elo difference: 7.9 +/- 10.5, LOS: 92.9 %, DrawRatio: 38.2 %

100+1s:
Score of 4ku 3.1 vs 4ku 3.0: 2322 - 1611 - 3067  [0.551] 7000
Elo difference: 35.4 +/- 6.1, LOS: 100.0 %, DrawRatio: 43.8 %

4ku 3.0

This version of 4ku-mini is due to play in the upcoming TCEC 4k2 tournament.

8.0+0.08s:
Score of 4ku 3.0 vs 4ku 2.0: 11598 - 3434 - 4968  [0.704] 20000
Elo difference: 150.6 +/- 4.4, LOS: 100.0 %, DrawRatio: 24.8 %

40.0+0.4s:
Score of 4ku 3.0 vs 4ku 2.0: 11796 - 2846 - 5358  [0.724] 20000
Elo difference: 167.3 +/- 4.4, LOS: 100.0 %, DrawRatio: 26.8 %

4ku 2.0

This version of 4ku-mini played in the TCEC's S24 Swiss 4 tournament. Various strength and size improvements since 1.1

10+0.1s:
Score of 4ku 2.0 vs 4ku 1.1: 1659 - 373 - 968  [0.714] 3000
Elo difference: 159.2 +/- 10.7, LOS: 100.0 %, DrawRatio: 32.3 %

New Contributors

• @Diazepawn made their first contribution in #121

Full Changelog: https://github.com/kz04px/4ku/commits/v2.0

4ku 1.1

4ku 1.1 is identical to 1.0 in every way besides improved UCI parsing.

4ku 1.0

This version of 4ku-mini played in the TCEC's S23 4k tournament.

4ku 1.0 has an issue where "movestogo" appearing before "wtime" or "btime" will cause the engine to incorrectly parse the input. This version should never be tested or even used. Use the fixed 1.1 instead.

New Contributors

• @GediminasMasaitis made their first contribution in #1
• @lambda0x00 made their first contribution in #5
• @Alex2262 made their first contribution in #13
• @gcp made their first contribution in #49
• @uwuplant made their first contribution in #64
• @PGG106 made their first contribution in #67
• @CGLemon made their first contribution in #92
• @dsmsgms made their first contribution in #93
• @unaiic made their first contribution in #96

Full Changelog: https://github.com/kz04px/4ku/commits/v1.0