Walked every tab on desktop and phone, clicked every link and button:
- Colours: new map inks and page colours, checked with simulated
protanopia/deuteranopia/tritanopia (worst pair ~12 CIEDE2000, was 1.4 for
Red/Green on the map and 9.7 for Orange/Brown on the page). Map label text
is always dark ink; dragon colour goes on borders. docs/COLOURS.md and
research/python/check_colours.py.
- Valley map: dragons get a paper halo and a visibility floor (faded ones never
drop below 0.3; bonding dragons start at 0.5); "no shadow" roll labelled
correctly.
- Wording: methods updated (strata, the newer form's questions, missing
answers, a new section on the solver, The Odds cards and genuine attempts);
findings use readable factor names and no longer claim a date adjustment;
'Load a newer route export above'; code links include solver.js and
explorer.js; The Odds shows 93% instead of '1 in 1'.
- Deep Dive route map treats all six dragons alike ('Bond a dragon' row,
default 'Bond any dragon'; rare sightings separate).
- Phone: Deep Dive no longer overflows (sections and selects can shrink).
- scripts/qa/crawl.js and mapcheck.js for future release checks.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
'Become a rider' and 'riders sent to Malek' now leave out runs where the rider
died on purpose (burned at an encounter they did not genuinely attempt, the
solver's genuineAttempt rule) and runs with no recorded outcome.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Remove the 'Your odds, honestly' heading and intro. The two summary cards now
lead the page: 'become a rider' (runs that ended bonded to a dragon) and the
unchanged 'riders sent to Malek'. The reach-the-hour figure moves into the
first card's note.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Bond rates for every dragon, the two Greens and the first-sighting table now
sit in Deep Dive next to the rare-dragon analysis; The Dragons tab is the
register only. The first-sighting table also drops rows whose company was
not recorded.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
In-page links now keep their sub-path (#valley/blue) when switching tabs, and
a link to a dragon opens it fresh on its most probable route (earlier detours
cleared). Clicking the Valley tab itself keeps the current selection.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
Original silhouettes built from the same parts as the dragons: your squadmate
with a raised sword facing the two bullies (sword and dagger), the wounded
cadet on the ground, and a thorny bush at the hour's walk. Each has a short
hover note and fades when a chosen route does not pass it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
- engine.js / analyze.py: one MISSING set ("Unknown", "Unsure",
"Unsure/unknown", "Multiple", "(not recorded)", "(not asked)", ...).
Factor tables compare only runs that answered each question and report how
many were left out; models are complete-case for every choice variable;
route rankings group only fully known routes; the route flow skips
missing steps.
- "I've lost count" now counts as a later attempt.
- The chase goal keeps its "not asked" stratum as an adjustment only; it is
no longer displayed as a finding.
- Page: no unknown levels in charts, tables or filters; factor explorer notes
how many runs were left out; Methods explains the rule.
JS/Python parity still exact.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
- New tab order: The Valley, The Odds, The Dragons, Deep Dive, Data & Methods
(Route Flow and Blue & Black fold into Deep Dive; methods into Data &
Methods; old #links redirect).
- explorer.js replaces the trails map and the route planner, with progressive
disclosure: the painted valley shows landmarks and the dragons where riders
meet them (faint = rarely met), no routes and no deaths; hover for a note;
choose a dragon to draw every route to every place it can be bonded, with
width = chance of still being on the route, and solid / dashed / dotted and
opacity for how a route compares with the best; click a choice point to
branch off and re-plan; "Why this route?" holds intervals and risks; a
toggle shows the community's actual traffic; #valley/blue/meadow-Blue links.
- dragons.js: original dragon silhouettes in six poses, built from simple
parts.
- Solver: route to one specific encounter (only: [...]); ENCOUNTERS per
colour, including the Greens rescue for Orange and Brown.
- The Odds rewritten for all six dragons, ranked easiest to hardest.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
- New "Route Planner" tab: pick a dragon colour, a goal (only this dragon /
best odds this run) and a data sample. The solver runs in the browser from
policy.json and walks the valley: every decision shows each option's
chance of bonding the colour (with the best play afterwards) and of
reaching an encounter alive; every roll shows its probabilities; deaths
show where they happen; "stays best in N% of re-runs" and "no clear
winner" flag noisy recommendations. Click any option on the map or in the
list to explore another branch.
- solver.js: values now carry P(reaching an encounter alive); advisor.step()
gives the planner the next decision, roll or outcome at any history.
- valley.js shares its painting with the planner.
- Fix: route-map target chips no longer capture clicks on other chip rows.
- Pages build gets a doctype and charset.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
- packages/core/src/solver.js: decision tree with perfect recall over the
community form's choices; chance events (first dragon, deaths, shadow,
meadow dragon, encounter outcome) depend on earlier choices through
hierarchical Dirichlet back-off with empirical-Bayes concentration per
event; backward induction for P(colour this attempt) and Dinkelbach for
P(colour) / (P(colour) + P(other colour)); 200 posterior draws for
credible intervals and "still best" rates; cross-fitted held-out check.
- Only genuine attempts inform encounter outcomes (goal / Malek intent).
- data/public/policy.json (all runs and a low-bias sample) and the readable
docs/POLICY.md; published on Pages with solver.js for the bot.
- Bot advisor: re-solves from any history, maps advice onto screen text,
reports margins so ties can be randomised. docs/BOT_INTEGRATION.md.
- Tests on a synthetic world with known answers; daily source update now
rebuilds the policy.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
The Threshing Routes form was redesigned (5,651 responses). Both implementations
(engine.js and clean.py) now read either version:
- one direction question + one colour question, with the older per-direction
columns used only to fill gaps and cross-check
- cadet encounter as an action path (scream, first and second combat move)
plus a separate outcome
- new fields: sighting_action, combat_second, rare_reported, malek_intent,
left_squadmates, tail; "Why are you doing this run?" mapped onto the goal
("Blue or Black" for rare chasers)
- the rare-dragon question feeds saw_blue_meadow / saw_shadow and the branch
The old export still codes identically. JS/Python parity: all match.
Data: form 2,573 -> 5,651 rows, Dragon Register 216 -> 273, new route PDF
fingerprint. Supersedes the automated pull request #2.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
- apps/atlas: the web page (valley map, route flow, Dragon Register)
- packages/core: coding rules and statistics (engine.js) and the shared route
graph (graph.js): scene graph, decision-variable graph, bot-to-form crosswalk,
Beta credible intervals
- packages/ingest: DragonKind bot importer to anonymised trajectories (no prose,
no account identifiers); data/bot holds Ken's sample
- research/python: independent reference implementation + update_sources.py
- tests: JS/Python parity checks and graph/importer unit tests
- CI, GitHub Pages deploy, daily check of the community sheets and route PDF
- Email addresses in free-text form answers are redacted; the raw spreadsheet
is not committed
- MIT code licence, CC BY 4.0 data licence
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ