Stackruns

Run T-0072 passed

· Evidence: SPRT Sequential probability ratio test. Games are played until the accumulated evidence reaches one of two bounds; then the test stops.

Hypothesis

Together with the int16 evaluation core, the third network generation replaces the hand evaluation and gains significantly in playing strength compared with the bound predecessor.

translated from German. Original: Die dritte Netzgeneration ersetzt zusammen mit dem int16-Bewertungskern die Handbewertung und gewinnt gegenüber dem gebundenen Vorgänger deutlich an Spielstärke.

Stages

Elo against the predecessorElo intervals: short stage +222.52 ± 36.23; long stage +260.44 ± 39.40.0+100+200+300short stage+222.52 ± 36.23short stage · passed+222.52 ± 36.23 Elo · 414 gameslong stage+260.44 ± 39.40long stage · passed+260.44 ± 39.40 Elo · 378 games
Point estimate with the reported 95-percent half-width, per stage. The dashed line is zero. Estimates from a stopped sequential test are biased by the stopping rule; the interval is nominal.
Field short stagelong stage
Verdictpassedpassed
Elo Strength difference to the opponent, estimated from the games, with a 95-percent half-width where the oracle reported one. Never an absolute rating.+222.52 ± 36.23+260.44 ± 39.40
nElo Normalised Elo: the Elo difference divided by the spread of the game results, so that the SPRT bounds mean the same at different time controls.+268.49 ± 33.47+331.64 ± 35.02
Games414378
Wins / draws / losses284 / 80 / 50279 / 60 / 39
pentanomial Games are played in pairs with swapped colours; the five counts are the pairs scoring 0, ½, 1, 1½ and 2 points.7 / 6 / 39 / 56 / 993 / 4 / 35 / 44 / 103
LLR Log-likelihood ratio, the running evidence of an SPRT. It starts at 0; a stage passes at the upper bound (about +2.94) and fails at the lower one (about −2.94).+2.95+2.96
Bounds−2.94 … +2.94−2.94 … +2.94
SPRT Sequential probability ratio test. Games are played until the accumulated evidence reaches one of two bounds; then the test stops.[0.0, 5.0][0.0, 5.0]
Error rates α / β0.05 / 0.050.05 / 0.05
Modelnormalizednormalized
time control Base time plus increment per move in seconds, e.g. 8+0.08: eight seconds per game plus 0.08 seconds per move.8+0.0840+0.4
BookAB
Opponentvs predecessorvs predecessor
Duration3 min15 min

Provenance

Date
2026-09-08
Started
2026-09-08 06:42:42 UTC
Duration
18 min
Author
Agent B
Epoch
4
Gatekeeper
passed
Bench nodes
3,406
Commit
b4ad0521baa1d6da3e24074be7dc8aaeb915ee58
SHA-256 A cryptographic hash. The same bytes always give the same hash, so a hash identifies exactly one version of a file or binary.
07cc4ef7b1ec1d39a95867aa1a5dd5b270fcf3a87337d7743146442ae0f23338
SHA-256 A cryptographic hash. The same bytes always give the same hash, so a hash identifies exactly one version of a file or binary.
365999ff8d66e92dc9a8f5421c3f77328de9327cc4eedec7648009ce4e61cc72
SHA-256 A cryptographic hash. The same bytes always give the same hash, so a hash identifies exactly one version of a file or binary.
5c547b3a8043c403ca4a4e59f21e52a24da77ca00b1e09d9581d6740b44f8532
SHA-256 A cryptographic hash. The same bytes always give the same hash, so a hash identifies exactly one version of a file or binary.
d483f6d2bfb8c0e96b442ff1371fd7c2f156ca1d0aa6c6e001088cb8c6c25ee9