Route model: first attempts are their own shadow roll; experiment plan and automated bot cycle

- game.yaml: `givens` (first or later attempt) and `distinct` levels that are not pooled; the shadow roll
  depends on the attempt first. Form submissions show the shadow on 9% of first attempts and 0.3% of later ones.
- Reporting bias: a roll whose bias cannot be measured can borrow another roll's (`borrow: {shadow: meadow}`),
  so form submissions inform the shadow again. An assumption, stated in planner/policy.yaml.
- policy.json: targets.<colour>.firstAttempt (this attempt, eventually). Black: about 1.4% on a first attempt,
  0.07% on a later one.
- solver.js: advisor options `attempt`, `fix`, sampling "uniform"; assign() and armFor() for the plan.
- planner/experiments.yaml + build_experiments.py -> data/public/experiments.json (progress and status per
  experiment, rebuilt from the bot's runs).
- python -m collection cycle: import bot runs, refit, re-plan, deliver the kit; .gitea/workflows/bot-cycle.yml.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014jqCC2LZ4x1WHQ1V57nTsv
This commit is contained in:
ekcSchiano
2026-11-12 03:31:33 +00:00
co-authored by Claude Fable 5.2
parent ee07ee143e
commit ab3664f369
25 changed files with 1938 additions and 573 deletions
+9
View File
@@ -74,6 +74,15 @@ def main():
a, e, l = t["thisAttempt"], t["eventually"], lb["targets"][X]["thisAttempt"]
att = e["expectedAttempts"]["mean"]
L.append(f"| {X} | {pct(a['pX']['mean'])} ({pct(a['pX']['lo'])}–{pct(a['pX']['hi'])}) | {pct(l['pX']['mean'])} ({pct(l['pX']['lo'])}–{pct(l['pX']['hi'])}) | {pct(t['crossFit']['heldOut'])} | {pct(e['ratio']['mean'])} | {'–' if att is None else js_round(att)} |")
L += ["", "## First attempts", ""]
L += ["The shadow at the tree is close to a first-attempt event, so an account's first attempt is solved separately (`givens` in `core/spec/game.yaml`). Only the colours that can come from the shadow differ much.", ""]
L += ["| Colour | First attempt | Low-bias sample | Later attempt |", "|---|---|---|---|"]
for X, t in s["targets"].items():
f, fl, a = t["firstAttempt"]["thisAttempt"]["pX"], lb["targets"][X]["firstAttempt"]["thisAttempt"]["pX"], t["thisAttempt"]["pX"]
L.append(f"| {X} | {pct(f['mean'])} ({pct(f['lo'])}–{pct(f['hi'])}) | {pct(fl['mean'])} ({pct(fl['lo'])}–{pct(fl['hi'])}) | {pct(a['mean'])} ({pct(a['lo'])}–{pct(a['hi'])}) |")
sh = (s["model"].get("reportingBias") or {}).get("form, chasing Blue/Black", {}).get("shadow", {})
if sh.get("used", "").startswith("assumed"):
L += ["", f"The shadow's reporting bias cannot be measured yet (complete logs have never seen one), so it is assumed equal to the meadow Blue's (`borrow` in `planner/policy.yaml`): form reports of a shadow count 1/{sh['rho']:.1f} each. The bot's first-attempt runs will replace this assumption."]
L += ["", "## How much do earlier choices matter?", ""]
L += ["Each chance event can depend on the choices before it. The model learns how much from the data (empirical-Bayes concentration; larger means the history matters less):", ""]
L += ["| Chance event | Depends on (most important first) | Concentration |", "|---|---|---|"]