BSD Testbench

test new policy net (copy) (with book) (copy)

new-policy vs old-policy

Duplicate Paused

Wins

Losses

Draws

Score

0.516

Elo Estimate

10.9 ( -75.4, 98.5 )

SPRT

Continue

Configuration

{
  "name": "test new policy net (copy) (with book) (copy)",
  "description": "Fast (current value+policy) vs Experimental (current value+old policy)",
  "system_a": {
    "name": "new-policy",
    "repo_url": "https://github.com/XInTheDark/brawlstarsdraft_0827.git",
    "commit_or_ref": "20251020_base",
    "subdir": null,
    "config_path": "webapp/config.yaml",
    "model": "Large",
    "models_dir": null,
    "download_link": null,
    "fm_device": null,
    "env_overrides": {},
    "request_timeout_seconds": null,
    "port_hint": null
  },
  "system_b": {
    "name": "old-policy",
    "repo_url": "https://github.com/XInTheDark/brawlstarsdraft_0827.git",
    "commit_or_ref": "20251020_old_policy",
    "subdir": null,
    "config_path": "webapp/config.yaml",
    "model": null,
    "models_dir": "models",
    "download_link": null,
    "fm_device": null,
    "env_overrides": {},
    "request_timeout_seconds": null,
    "port_hint": null
  },
  "judges": [
    {
      "name": "website-judge",
      "weight": 1.0,
      "epsilon": null,
      "evaluate_both_sides": true,
      "system": {
        "name": "website-judge",
        "repo_url": "https://github.com/XInTheDark/brawlstarsdraft_0827.git",
        "commit_or_ref": "website",
        "subdir": null,
        "config_path": "webapp/config.yaml",
        "model": "Large",
        "models_dir": null,
        "download_link": null,
        "fm_device": null,
        "env_overrides": {},
        "request_timeout_seconds": null,
        "port_hint": null
      }
    }
  ],
  "judge_policy": "average",
  "time_control": {
    "seconds": 6.0,
    "rollout": "pad",
    "policy_temperature": 0.8,
    "seed_mode": "deterministic",
    "extra_flags": null,
    "normalize_sides": true
  },
  "draft": {
    "bans_per_side": 3,
    "epsilon_draw": 0.05,
    "book_path": "books/ranked.yaml",
    "map_mode_weights": {}
  },
  "sprt": {
    "elo0": 0.0,
    "elo1": 2.0,
    "alpha": 0.05,
    "beta": 0.05
  },
  "fixed_n": null,
  "parallelism_hint": 4,
  "max_trials": null,
  "id": "a06f45c0-8b9c-4ced-99ba-45299cb0fcf8",
  "status": "paused",
  "created_at": "2025-10-20T09:46:07"
}

LLR Progress

Recent Games

Export JSON Export CSV Export NDJSON

Pair	Game	Result	Map	p(A)
851	0	B	bounty / Snake Prairie	0.359
851	1	A	bounty / Snake Prairie	0.918
852	0	A	gemGrab / Gem Fort	0.764
852	1	B	gemGrab / Gem Fort	0.423
850	0	B	hotZone / Open Business	0.376
850	1	A	hotZone / Open Business	0.693
848	0	A	hotZone / Ring of Fire	0.673
848	1	B	hotZone / Ring of Fire	0.360
849	0	A	bounty / Shooting Star	0.667
849	1	B	bounty / Shooting Star	0.486
847	0	A	bounty / Hideout	0.662
847	1	D	bounty / Hideout	0.605
845	0	B	bounty / Excel	0.542
845	1	A	bounty / Excel	0.884
846	0	A	gemGrab / Deathcap Trap	0.601
846	1	B	gemGrab / Deathcap Trap	0.389
844	0	B	brawlBall / Grass Knot	0.423
844	1	A	brawlBall / Grass Knot	0.673
842	0	A	hotZone / Dueling Beetles	0.673
842	1	B	hotZone / Dueling Beetles	0.359
843	0	B	bounty / Layer Cake	0.560
843	1	A	bounty / Layer Cake	0.788
841	0	A	brawlBall / Triple Dribble	0.536
841	1	B	brawlBall / Triple Dribble	0.461
840	0	D	brawlBall / Sneaky Fields	0.524
840	1	D	brawlBall / Sneaky Fields	0.560
839	0	A	brawlBall / Sneaky Fields	0.510
839	1	B	brawlBall / Sneaky Fields	0.460
838	0	B	bounty / Excel	0.493
838	1	A	bounty / Excel	0.901
836	0	A	bounty / Hideout	0.611
836	1	B	bounty / Hideout	0.496
837	0	A	brawlBall / Center Stage	0.630
837	1	B	brawlBall / Center Stage	0.564
835	0	A	brawlBall / Backyard Bowl	0.659
835	1	B	brawlBall / Backyard Bowl	0.469
833	0	A	brawlBall / Pinhole Punt	0.722
833	1	D	brawlBall / Pinhole Punt	0.623
834	0	A	knockout / New Horizons	0.725
834	1	B	knockout / New Horizons	0.652
832	0	D	brawlBall / Sneaky Fields	0.517
832	1	A	brawlBall / Sneaky Fields	0.514
830	0	A	hotZone / Dueling Beetles	0.862
830	1	B	hotZone / Dueling Beetles	0.480
831	0	A	bounty / No Excuses	0.720
831	1	B	bounty / No Excuses	0.238
829	0	B	gemGrab / Hard Rock Mine	0.325
829	1	A	gemGrab / Hard Rock Mine	0.723
827	0	D	bounty / Shooting Star	0.534
827	1	B	bounty / Shooting Star	0.369
828	0	A	brawlBall / Center Stage	0.647
828	1	B	brawlBall / Center Stage	0.575
826	0	A	bounty / Snake Prairie	0.928
826	1	B	bounty / Snake Prairie	0.414
824	0	A	hotZone / Ring of Fire	0.733
824	1	B	hotZone / Ring of Fire	0.416
825	0	B	bounty / Dry Season	0.601
825	1	A	bounty / Dry Season	0.824
821	0	B	brawlBall / Triple Dribble	0.576
821	1	A	brawlBall / Triple Dribble	0.684
820	0	A	hotZone / Parallel Plays	0.808
820	1	B	hotZone / Parallel Plays	0.357
817	0	B	bounty / Dry Season	0.569
817	1	A	bounty / Dry Season	0.858

test new policy net (copy) (with book) (copy)

Configuration

LLR Progress

Recent Games

Draft Details