Records
OLSS-II with a two-strategy portfolio against full KLUSS at the 200 ms budget
Registration 092a047f-bad7-4d56-98f0-5e28094c6ec9 (study subgame-solving/opponent-portfolio), run 2026-09-19 on AMD Ryzen AI Max (Strix Halo), 32 threads, 125 GB unified memory, Radeon 8060S via ROCm, Linux 6.18. Server commit e3107a0, research commit 9588223 (registration files as committed; this record added afterwards). Other agents shared the machine during the run; per-seed results are deterministic regardless of load.
Command:
cd server && cargo run --release -p armies-arena --bin matchup -- --red 'olss2 n2 200ms=kluss:{"mode":"olss2","samples":64,"expansions":10000,"cfr_every":64,"portfolio":[{"policy":"greedy"},{"policy":"self","parameters":{"temperature":0.3}}]}' --blue 'kluss 200ms=kluss:{"samples":64,"shadows":4,"expansions":5000,"cfr_every":64}' --games 200 --seed 0 --rules training --threads 8 --record-dir ../research/runs/092a047f-bad7-4d56-98f0-5e28094c6ec9/records --out ../research/runs/092a047f-bad7-4d56-98f0-5e28094c6ec9/matchup.json --tag 092a047f-bad7-4d56-98f0-5e28094c6ec9
cd research && python3 scripts/paired_contrast.py runs/092a047f-bad7-4d56-98f0-5e28094c6ec9/matchup.json| Player | Wins | Draws | Losses | Effective | 95% Wilson | Mean plies | ms/move |
|---|---|---|---|---|---|---|---|
| olss2 N=2 (greedy, self 0.3; 10,000 expansions) | 124 | 0 | 76 | 0.620 | 0.551 to 0.684 | 197 | 489.547 |
| kluss (5,000 expansions) | 76 | 0 | 124 | 0.380 | 0.316 to 0.449 | 197 | 388.057 |
| ending: flag_captured | 190 | ||||||
| ending: no_moves | 10 |
Per-seed paired contrast for the olss2 player (score as Red plus score as Blue minus one): mean 0.240, 95% normal interval 0.085 to 0.395 over 100 seeds (46 positive, 32 split, 22 negative). The olss2 player won 70 games as Red and 54 as Blue. Wall time 2,220 s on 8 threads. Time per move is as played in the arena under load; the calibration bench (single thread, middlegame positions, other jobs sharing the machine) read 190 ms for the olss2 side at 10,000 expansions and 170 ms for kluss at 5,000.
Decision rule
The hypothesis (the order reverses at 200 ms, kluss ahead) is not supported: the contrast interval 0.085 to 0.395 lies entirely above zero, so olss2 with the two-strategy portfolio is still ahead at this budget.
Provenance
Archived JSON: runs/092a047f-bad7-4d56-98f0-5e28094c6ec9/matchup.json (SHA-256 9221624a107d52edbe7f259b35eeab1f7faf60b6f9f2b7f25d10ddb7ffaaf4f5), with per-game moves in the arena record files under the same directory. Games planned 200, attempted 200, completed 200, interrupted 0. No deviation from the registration.