From fea6792f254d081375afb083a7013634487deafe Mon Sep 17 00:00:00 2001 From: Federico Pasqua Date: Sun, 27 Sep 2026 18:37:10 +0200 Subject: [PATCH] Record the final 6.8 run with the fixed bots; Phase 6 complete Co-Authored-By: Claude Opus 5.5 (1M context) --- agents/plans/2026-09-25-dojo-board-and-sim.md | 47 +++++++++++++++++-- 1 file changed, 42 insertions(+), 5 deletions(-) diff --git a/agents/plans/2026-09-25-dojo-board-and-sim.md b/agents/plans/2026-09-25-dojo-board-and-sim.md index 37cc998..f8debc7 100644 --- a/agents/plans/2026-09-25-dojo-board-and-sim.md +++ b/agents/plans/2026-09-25-dojo-board-and-sim.md @@ -179,7 +179,7 @@ Deliverables: - One change at a time. When a change is adopted, re-run the 6.1 baseline so later steps start from the current rules, and log the decision (D-number) here and in `RULES.md`. - Results go in this file as a table per step (command line, scenario, key measures). -**Next action:** redo the 6.8 run with the fixed bots (D51), then 6.8.5 and Phase 7. 6.2 is suspended (D43), 6.9 skipped (D44), 6.4.1 parked (D49). Everything up to D52 is committed; commit only when the developer asks. +**Next action:** Phase 6 is complete; Phase 7 (7.1 freeze the rules) when the developer says so. 6.9 skipped (D44), 6.4.1 parked (D49). Everything up to D52 is committed; the 6.8 final run is not; commit only when the developer asks. **Targets** (confirmed by the developer as a starting point, D32): @@ -401,8 +401,8 @@ Deliverables: Exam pass rate unchanged (40–44%). Reading: −1 saves 3–5 minutes on the mean and 6–12 on the p90; 0 adds only 1–2 more. The price is the comeback: the first Brown keeps retrying instead of dropping back (fewer different challengers), lead changes fall by ~1 per game, the halfway leader wins 3–8 points more, the worst player is out a little longer. No setting brings casual under p90 110. - [ ] 6.9 Catch-up candidate A — **skipped (D44)**. Was: `RulesConfig` switch, Allenamento +2 for a player 2+ belts behind the leader (bot aware of it); compare on the contention, lead and length measures and on the style win rates. -- [ ] 6.8 Consolidation: the final configuration at 3 / 4 / 6 / 8 players (start 0 as for balancing, plus start 3 at 6 and 8 as in the rules, D48) with all lineups except `random` (D45), 2,000 games each; every target checked; `RULES.md` updated to the next version with the decisions; open items moved to Phase 7. - - [x] 6.8.1 Run (rules v0.9 = v0.8 + D47–D50, board draft 2; 2,000 games per scenario, 3 min 27 s): `uv run dojo sim --games 2000 --seed 0 --players 3 4 6 8 --start 0 --lineup balanced,farmer,aggressive balanced casual aggressive`, and the same with `--players 6 8 --start 3`. All games finished; Exam pass rate 41–44% everywhere; win rate by bot style (mixed) within 4 pp at every count. +- [x] 6.8 Consolidation: the final configuration at 3 / 4 / 6 / 8 players (start 0 as for balancing, plus start 3 at 6 and 8 as in the rules, D48) with all lineups except `random` (D45), 2,000 games each; every target checked; `RULES.md` updated to the next version with the decisions; open items moved to Phase 7. + - [x] 6.8.1 First run, **with the biased bots (before D51; superseded by 6.8.6)** (rules v0.9 = v0.8 + D47–D50, board draft 2; 2,000 games per scenario, 3 min 27 s): `uv run dojo sim --games 2000 --seed 0 --players 3 4 6 8 --start 0 --lineup balanced,farmer,aggressive balanced casual aggressive`, and the same with `--players 6 8 --start 3`. All games finished; Exam pass rate 41–44% everywhere; win rate by bot style (mixed) within 4 pp at every count. | Players | Start | Lineup | Minutes mean / p90 | Exam passed | Halfway leader wins | Out of contention; worst player | Rivals within 1 belt | Win rate by seat | |---|---|---|---|---|---|---|---|---| @@ -431,13 +431,50 @@ Deliverables: | 8 | 3 | casual | 144 / 203 | 42% | 24% | 11%; 40% | 2.33 | 11 / 12 / 11 / 13 / 13 / 13 / 14 / 13 | | 8 | 3 | aggressive | 135 / 202 | 41% | 27% | 17%; 49% | 1.05 | 10 / 12 / 10 / 14 / 14 / 14 / 14 / 13 | - - [x] 6.8.2 Targets at 4 players (start 0): inside for mixed and balanced on every measure. Accepted misses (D50): casual 91 / 123 and aggressive 90 / 128 minutes; aggressive seats 21–29% and rivals 0.51 (edge). 3 players: short (53–71 min), halfway leader wins 45–52% (edge of 35–55%), rivals 0.36–0.60 with heuristic bots. 6–8 players: long (D48 accepted), worst player out 27–49% of turns, halfway leader wins 23–32%. + - [x] 6.8.2 Targets at 4 players (start 0; first run, see 6.8.6 for the final figures): inside for mixed and balanced on every measure. Accepted misses (D50): casual 91 / 123 and aggressive 90 / 128 minutes; aggressive seats 21–29% and rivals 0.51 (edge). 3 players: short (53–71 min), halfway leader wins 45–52% (edge of 35–55%), rivals 0.36–0.60 with heuristic bots. 6–8 players: long (D48 accepted), worst player out 27–49% of turns, halfway leader wins 23–32%. - [x] 6.8.3 `RULES.md` v0.9: D47–D50 in the text; "Rule variants under test" replaced by "Tested alternatives" (the simulator switches). - [x] 6.8.4 Seat order (finding; studied in 6.2, closed by D51 and D52): the first seat is behind at every count, and more so with more players — 4 players 21–24% (fair 25%), 6 players 11–16% (fair 17%), 8 players 9–12% (fair 12.5%); later seats rise to 20% (6) and 15% (8). At 6 players seats 4 and 5 are the two players alone in a corner (`seat_corners` gives 0, 1, 0, 1, 2, 3) and win most (17–22%), so sharing a Safe Room costs too. 6.2 (suspended, D43) would be the step to study it. - - [ ] 6.8.5 Open items moved to Phase 7. + - [x] 6.8.5 Open items moved to Phase 7 (7.0). + - [x] 6.8.6 Final run with the fixed bots (D51, D52; same commands as 6.8.1). All games finished; Exam pass rate 40–44% everywhere; bot-style spread (mixed) 2–4 pp. + + | Players | Start | Lineup | Minutes mean / p90 | Exam passed | Halfway leader wins | Out of contention; worst player | Rivals within 1 belt | Win rate by seat | Bot spread | + |---|---|---|---|---|---|---|---|---|---| + | 3 | 0 | mixed | 53 / 72 | 41% | 50% | 2%; 6% | 0.71 | 31 / 34 / 36 | 4 pp | + | 3 | 0 | balanced | 51 / 70 | 40% | 48% | 1%; 4% | 0.74 | 31 / 32 / 37 | — | + | 3 | 0 | casual | 70 / 97 | 44% | 45% | 2%; 5% | 1.06 | 34 / 32 / 34 | — | + | 3 | 0 | aggressive | 61 / 86 | 41% | 52% | 4%; 11% | 0.45 | 33 / 33 / 35 | — | + | 4 | 0 | mixed | 67 / 91 | 42% | 43% | 3%; 11% | 0.95 | 23 / 25 / 26 / 26 | 3 pp | + | 4 | 0 | balanced | 65 / 86 | 42% | 38% | 2%; 8% | 0.96 | 24 / 25 / 26 / 25 | — | + | 4 | 0 | casual | 90 / 122 | 43% | 38% | 3%; 10% | 1.48 | 24 / 26 / 26 / 24 | — | + | 4 | 0 | aggressive | 82 / 114 | 41% | 40% | 6%; 17% | 0.65 | 22 / 26 / 26 / 27 | — | + | 6 | 0 | mixed | 101 / 135 | 42% | 33% | 8%; 27% | 1.27 | 13 / 17 / 16 / 15 / 18 / 20 | 3 pp | + | 6 | 0 | balanced | 96 / 128 | 42% | 30% | 7%; 25% | 1.23 | 14 / 16 / 15 / 17 / 18 / 20 | — | + | 6 | 0 | casual | 130 / 176 | 44% | 29% | 7%; 25% | 1.91 | 15 / 16 / 16 / 16 / 19 / 17 | — | + | 6 | 0 | aggressive | 121 / 167 | 41% | 31% | 11%; 32% | 0.96 | 15 / 15 / 16 / 16 / 18 / 20 | — | + | 8 | 0 | mixed | 130 / 171 | 43% | 25% | 10%; 34% | 1.56 | 10 / 12 / 12 / 13 / 12 / 12 / 14 / 15 | 3 pp | + | 8 | 0 | balanced | 123 / 164 | 43% | 25% | 10%; 36% | 1.41 | 11 / 11 / 12 / 13 / 14 / 13 / 13 / 13 | — | + | 8 | 0 | casual | 168 / 224 | 44% | 23% | 10%; 36% | 2.25 | 11 / 13 / 12 / 13 / 12 / 14 / 12 / 14 | — | + | 8 | 0 | aggressive | 154 / 212 | 43% | 27% | 14%; 42% | 1.18 | 11 / 11 / 11 / 14 / 13 / 11 / 15 / 14 | — | + | 6 | 3 | mixed | 80 / 111 | 43% | 33% | 6%; 25% | 1.35 | 15 / 14 / 17 / 17 / 18 / 19 | 3 pp | + | 6 | 3 | balanced | 77 / 106 | 42% | 30% | 6%; 24% | 1.29 | 14 / 15 / 16 / 18 / 19 / 18 | — | + | 6 | 3 | casual | 108 / 150 | 44% | 30% | 7%; 26% | 1.97 | 16 / 15 / 16 / 16 / 18 / 19 | — | + | 6 | 3 | aggressive | 95 / 136 | 43% | 34% | 10%; 34% | 1.01 | 14 / 16 / 15 / 19 / 18 / 18 | — | + | 8 | 3 | mixed | 102 / 144 | 43% | 28% | 8%; 34% | 1.68 | 10 / 12 / 10 / 14 / 13 / 13 / 14 / 14 | 2 pp | + | 8 | 3 | balanced | 97 / 138 | 42% | 28% | 10%; 37% | 1.49 | 11 / 10 / 12 / 12 / 15 / 14 / 14 / 12 | — | + | 8 | 3 | casual | 140 / 198 | 43% | 25% | 10%; 38% | 2.39 | 12 / 11 / 12 / 14 / 11 / 13 / 15 / 12 | — | + | 8 | 3 | aggressive | 122 / 178 | 41% | 28% | 14%; 45% | 1.27 | 10 / 12 / 12 / 14 / 12 / 15 / 13 / 12 | — | + + - Targets at 4 players: mixed (67 / 91) and balanced (65 / 86) inside on every measure; seats 22–27% in every lineup. Accepted misses (D50): casual 90 / 122 (mean at the edge, p90 over), aggressive 82 / 114 (p90 over). Rivals within one belt 0.65–1.48. + - 3 players: 51–70 min; halfway leader wins 45–52% (edge); rivals 0.45–1.06. 6–8 players: long (D48), worst player out 24–45% of turns, halfway leader wins 23–34%; from Yellow, 6 players take 77–108 min. ### Phase 7 — Finalisation (later) +- [ ] 7.0 Open items carried over from Phase 6 (none blocks 7.1): + - Q2 board asymmetry (deferred by the developer). + - 6.4.1 chance square (parked, D49). + - Heuristic bots draw a Nebbia card on only 2.3% of turns (6.6): card effects are measured mostly through the casual bot. + - Remaining small bot tie-breaks by absolute board order (Tunnel destination; the engine keeps the first path per outcome): corner residual 24.0–25.4% after D51. + - At 6 players the two seats alone in a Safe Room win 17–20% vs 13–17% for the four sharing (6.2.3); accepted with D52. - [ ] 7.1 Freeze rules. - [ ] 7.2 Italian edition. - [ ] 7.3 Physical tracking method and components; printable board.