- 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
A form row recording a Green bonded at the shadow Blue let the "eventually" objective
find a route whose only possible bond was that Green (ratio 1, near-zero chance).
game.yaml now lists the possible bonds per encounter (own colour; the two Greens also
Orange and Brown); anything else is a data-entry error and is not counted.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
- game.yaml: chance events can have `requires` (a gate) and `otherwise`; the shadow
requires "Investigate the scream" and "Help your squadmate" (the route PDF's Black
Dragon Gates; spreadsheet: 78 shadows in 4199 gated tree runs, 2 in 646 others)
- model and advisor: a closed gate fixes the roll; runs off the gate are not counted
- form.yaml: "Protect / Help Squadmate" (new answer) -> Protect squadmate
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Step 4b of the Python-first core.
- core/solver.js is now a 250-line advisor (was 900 lines): it reads the game (core/spec/game.yaml,
published inside policy.json as model.game) and the count tables, and re-solves from any point in
a run with the same interpreter as the Python model. Same API for the site and Ken's bot
(advisor, recommendOption, solverHistory). The bot's on-screen text for each option moves to
game.yaml (bot_text); the deadly rolls carry their risk labels there too.
- tests/core/advisor.test.js: the advisor gives exactly the Python model's values and choices for
every colour and goal (to 1e-12, every decision on every published route), walks a run in the
game's order, and maps advice onto the bot's screen.
- Removed core/engine.js and everything that only compared it with Python (tests/parity/run_engine,
compare_runs, compare_results, compare_routes, compare_solver, compare_stats, analysis/stats/
analyze.py). In their place: tests/core/test_stats.py checks the statistics against known answers,
test_solver.py checks the route model on a synthetic world (tests/fixtures/synthetic_policy.py).
- analysis/network/build-graph.js reads the Python-coded runs; graph.json is byte-identical.
- docs/BOT_INTEGRATION.md: what changed for the bot (history keys, new choices, re-sync all three files).
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
1. Letting the Red pass. Only climbing down quickly puts you above the Red, so a tree run that
passed on the Red and then met the Greens let it pass. Those 16 runs were coded as slow
descents; they are now "Descend quickly: let the Red pass" and count for the Greens'
encounter, so the route model's "let it pass" option rests on its own runs. A tree run with
both the Red and the Greens attempted (impossible in one run) is now flagged.
2. Leaving your squadmates. The form only offered "> Left them" from 2026-10-27 (0 of 504 squad
runs before chose it, 40 of 144 after). Earlier and undated squad runs are now unknown rather
than "stayed together" (form.yaml: left_squadmates_asked_from).
3. Bot text. The wounded cadet's other two choices map onto the form; a test checks every choice
the bot has logged maps onto a form variable (encounter answers excepted, not modelled yet).
Each rule has a test in tests/core/test_coding.py. Data rebuilt (runs, graph, policy, report);
the headline odds move by at most 0.02 points. The Methods tab describes both rules.
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Step 2 of the Python-first, declarative core.
- core/spec/game.yaml: the game as data: every decision and its options, every roll and the
earlier steps its odds may depend on, every bonding encounter, and the flow of a run from the
entrance to the dragons (set / keep / drop / go / end / encounter). Also how the form's coded
answers map onto the game.
- core/atlas/model.py: what each coded run reveals about each step, count tables, empirical-Bayes
shrinkage, and the posterior draws (same random numbers as before, reproducible from the seed).
- core/atlas/solver.py: one interpreter of game.yaml used three ways (solve, evaluate a fixed
policy, export the published tree) instead of the tree written out five times. Vectorised
over the 200 posterior draws: the build takes about 2 minutes instead of 4, with no memory flag.
- planner/build_policy.py and policy_report.py replace the JS build and report.
Checked against the JavaScript model: every number in policy.json agrees to within 4e-12
(last-digit differences in log/cos). Three "still best" shares differ by one draw in 200, at
near-ties under the "eventually" objective. POLICY.md is otherwise identical.
tests/parity/compare_solver.py keeps checking JS against Python in CI until the site's JS advisor
replaces solver.js; tests/core/test_game.py checks game.yaml is complete and consistent.
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
Step 1 of moving the repo to Python with declarative rules.
- core/spec/form.yaml: every question (how it is found in an export), every answer and what it
is coded as, missing answers, email redaction, burn order, first-dragon pools.
- core/atlas/coding.py reads it and replaces core/python/clean.py (python -m atlas.coding).
- tests/core/test_coding.py: 16 rule tests, one hand-written row per rule, checking the rules
are right rather than only that JS and Python agree.
- analysis/stats/analyze.py takes the colours and missing answers from the spec instead of its
own copy; the daily update job codes rows through the package.
- pyproject.toml for the package; pyyaml and pytest added to requirements.
The coded runs are byte-identical to before (5,988 runs); all parity checks still match.
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>