shawn’s tests

Active 0 Workers / 0 Training / 0 Test threads / 0 MNPS

No active tests

Finished 15 shown

Author	Engine	Identifier	Code	Time control	Results
shawn	WakWak	more_stuff #8255 Failed Error Duck960	diff ↗	8.0+0.08	LLR: -2.32 (-2.25, 2.89) [0.00, 4.20] Games: 42420 W: 16814 L: 16760 D: 8846 Ptnml(0-2): 2868, 3432, 8560, 3478, 2872
butterfly duckfutationss
shawn	WakWak	whatif2 #8234 Failed Error Duck960	diff ↗	8.0+0.08	LLR: -2.49 (-2.25, 2.89) [0.00, 4.20] Games: 85252 W: 34009 L: 33749 D: 17494 Ptnml(0-2): 5766, 6869, 17181, 6959, 5851
what if????
shawn	WakWak	deny_duck #8250 Failed Duck960	diff ↗	8.0+0.08	LLR: -3.02 (-2.25, 2.89) [0.00, 4.20] Games: 734 W: 191 L: 419 D: 124 Ptnml(0-2): 111, 75, 131, 31, 19
doth this work
shawn	WakWak	dfk2 #8235 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.27 (-2.25, 2.89) [0.00, 4.20] Games: 8398 W: 3316 L: 3432 D: 1650 Ptnml(0-2): 605, 690, 1686, 652, 566
what
shawn	WakWak	qs_selective #8236 Passed Duck960	diff ↗	8.0+0.08	LLR: 3.10 (-2.25, 2.89) [0.00, 4.20] Games: 13798 W: 5604 L: 5319 D: 2875 Ptnml(0-2): 869, 1101, 2801, 1132, 996
qs selective
shawn	WakWak	dfk #8231 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.45 (-2.25, 2.89) [0.00, 4.20] Games: 39896 W: 15936 L: 15904 D: 8056 Ptnml(0-2): 2648, 3231, 8224, 3131, 2714
asdfas
shawn	WakWak	whatif #8230 Passed Duck960	diff ↗	8.0+0.08	LLR: 3.18 (-2.25, 2.89) [0.00, 4.20] Games: 1814 W: 847 L: 616 D: 351 Ptnml(0-2): 85, 100, 391, 161, 170
what if?
shawn	WakWak	weak_duckfutations #8229 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.36 (-2.25, 2.89) [0.00, 4.20] Games: 9404 W: 3679 L: 3795 D: 1930 Ptnml(0-2): 649, 800, 1900, 724, 629
Weak duckfuatations
shawn	WakWak	lmr_more #8228 Finished Duck960	diff ↗	8.0+0.08	LLR: -1.40 (-2.25, 2.89) [0.00, 4.20] Games: 10776 W: 4288 L: 4331 D: 2157 Ptnml(0-2): 736, 920, 2148, 819, 765
dmr
shawn	WakWak	that #8224 Finished Duck960	diff ↗	8.0+0.08	LLR: -1.10 (-2.25, 2.89) [0.00, 4.20] Games: 6078 W: 2503 L: 2550 D: 1025 Ptnml(0-2): 489, 412, 1251, 431, 456
ldp saver
shawn	WakWak	roast_duck #8226 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.88 (-2.25, 2.89) [0.00, 4.20] Games: 1894 W: 681 L: 883 D: 330 Ptnml(0-2): 196, 153, 368, 117, 113
roast duck
shawn	WakWak	this #8222 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.65 (-2.25, 2.89) [0.00, 4.20] Games: 1034 W: 351 L: 534 D: 149 Ptnml(0-2): 120, 77, 226, 54, 40
PVDCP
shawn	WakWak	abcd #8219 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.32 (-2.25, 2.89) [0.00, 4.20] Games: 6500 W: 2605 L: 2738 D: 1157 Ptnml(0-2): 526, 478, 1324, 447, 475
net: promote net-502E9197 (second decay tail on the 23C62E38 lineage) Same architecture and data as 8C99AF21. The 8C99AF21 run's raw SB-5000 checkpoint is resumed with its optimiser state and trained 1,000 more superbatches at the same low learning-rate tail (1e-6 cosine to 1e-7), SWA over the last 250. Measured against 8C99AF21, same binary both sides, only the net differing: STC 10+0.1 -1.4 +-2.1 27,772 games H0 (#3587) LTC 40+0.4 +2.4 +-1.8 36,000 games H1 (#3586) Cross-engine round robin on a modern host (11 engines, 6,000 games each at 10+0.1, ~89% draws): the three nets 23C62E38 / 8C99AF21 / 502E9197 finish within CI of each other (10 / 8 / 9), i.e. no regression against outside opposition, which is what that gate is for. Deployment and rating-list play are at the longer time control, where the head-to-head gain is. Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_0176jDMArbavpkTdKKeLBb6Q
shawn	WakWak	asdf #8220 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.29 (-2.25, 2.89) [0.00, 4.20] Games: 4084 W: 1602 L: 1743 D: 739 Ptnml(0-2): 331, 300, 862, 277, 272
6x effect 66x elo
shawn	WakWak	asdf #8218 Failed Duck960	diff ↗	8.0+0.08	LLR: -2.28 (-2.25, 2.89) [0.00, 6.77] Games: 11526 W: 4733 L: 4737 D: 2056 Ptnml(0-2): 878, 803, 2396, 817, 869
asdlkfjaskldfjalksdjfklasjdfklajsdklfklasj