Settlers / Research

75 pages · Search titles and descriptions

↑ ↓ to navigate · Enter to open · Esc to closeLocal search
Play the game

Confirmed through the protocol arena

Win rate by competitor

95% interval
expectimax-v2-plan · slot 058.8% · n=80
eta · slot 113.8% · n=80
fast · slot 212.5% · n=80
eta · slot 315.0% · n=80
0%50%100%
run record. 80 completed of 80 planned games. 95% Wilson intervals; descriptive evidence only.

The candidate is the corrected expectimax-v2 at depth 3 under a 1.5 s budget with eight common random scenarios and four sampled worlds, frozen in the policy registry as expectimax-v2-plan before the cohort. The lineup and rule are the ones the pre-fix candidate had failed a day earlier with 31 wins against 25. Twelve matches ran at once against one server; the wall-clock budget makes contention search less, never longer, so any effect of sharing the machine works against the candidate.

The mean game lasted 69 turns. The candidate averaged 8.40 points to the builders' 6.5 to 7.1, and the two identical ETA policies finished within one win of each other this time.

One game, viewed from above

The first scheduled game, selected by schedule rather than outcome.

eta slot 1’s turn

Turn 0 · Place a settlement · 10 points to win

Ore4Ore, rolls 4, 3 of 36 dice outcomesWool11Wool, rolls 11, 2 of 36 dice outcomesLumber12Lumber, rolls 12, 1 of 36 dice outcomesGrain8Grain, rolls 8, 5 of 36 dice outcomesOre3Ore, rolls 3, 2 of 36 dice outcomesBrick6Brick, rolls 6, 5 of 36 dice outcomesBrick9Brick, rolls 9, 4 of 36 dice outcomesWool5Wool, rolls 5, 4 of 36 dice outcomesLumber10Lumber, rolls 10, 3 of 36 dice outcomesBrick11Brick, rolls 11, 2 of 36 dice outcomesGrain5Grain, rolls 5, 4 of 36 dice outcomesWool10Wool, rolls 10, 3 of 36 dice outcomesGrain2Grain, rolls 2, 1 of 36 dice outcomesWool9Wool, rolls 9, 4 of 36 dice outcomesOre4Ore, rolls 4, 3 of 36 dice outcomesDesertDesert, blocked by the robberLumber6Lumber, rolls 6, 5 of 36 dice outcomesLumber3Lumber, rolls 3, 2 of 36 dice outcomesGrain8Grain, rolls 8, 5 of 36 dice outcomes3:1Any resource 3:1 port2:1Ore 2:1 port3:1Any resource 3:1 port2:1Lumber 2:1 port2:1Wool 2:1 port3:1Any resource 3:1 port2:1Brick 2:1 port3:1Any resource 3:1 port2:1Grain 2:1 port
Settlement City Robber Roll frequency
Bank19Brick19Lumber19Wool19Grain19Ore25 development cards left

expectimax-v2-plan slot 0

0 VP
0 resource cards0 development cards0 settlements0 cities0 roads0 knights played
Longest route: 0

eta slot 1

Playing

0 VP
0 resource cards0 development cards0 settlements0 cities0 roads0 knights played
Longest route: 0

fast slot 2

0 VP
0 resource cards0 development cards0 settlements0 cities0 roads0 knights played
Longest route: 0

eta slot 3

0 VP
0 resource cards0 development cards0 settlements0 cities0 roads0 knights played
Longest route: 0

Actions & conversation

1 events · since the previous position

  1. Table

    The game started.

Game 1 in the scheduled cohort.

109 of 109 recorded positions. Board changes between positions are grouped.

Replay sources

Game aa659137-6e03-4c83-b126-5c63e561edbc · version 14

Frames: 4b9783ed0a6f0eb46818295534984f46b92ee8ece278548c02926e5587a15204

Public events: ea05520a0f84c7ec9e0a0591ce5ff4532b5d38f6119b1e4312202bd1d38e0fde

Methods and reproduction

The registered protocol under study search/expectimax-v2, run with just run 49427921-feaf-4291-a5bf-79500acdcae7 --parallel 12. The exact test counts wins by the candidate slot and the compared ETA slot only: P(X ≥ 47 | n = 58, p = 0.5) = 1.02 × 10⁻⁶. The candidate's exact configuration is the doubled-brace JSON in policies/registry.toml; the run manifest records the server remote binary hash and both repositories' Git state (server commit 59433b7 carries the planning fix). Public frame source SHA-256: 4b9783ed0a6f0eb46818295534984f46b92ee8ece278548c02926e5587a15204. Attribution: Claude Fable 5.1 (claude-fable-5-1) through Claude Code.

Detailed result and cohort context

80-game protocol cohort completed 80 of 80 games with no timeout moves. expectimax-v2-plan won 47 (58.8%); the ETA builder in the compared slot won 11 (13.8%), the other ETA slot 12 and the fast builder 10. The registered rule asked for a 10-point gap and a one-sided exact test at p ≤ 0.05: the gap is 45 points and p = 1.0 × 10⁻⁶. The candidate is the new protocol-arena baseline.