Refresh to 7,775 spreadsheet runs; route model learns from the bot and corrects for reporting bias

- spreadsheet 5,988 -> 7,775 runs; new route PDF (Black Dragon Gates); register unchanged
- form.yaml: 'Protect / Help Squadmate' -> Protect squadmate
- policy, statistics, graph, metadata rebuilt

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UCzTd31GphpMXDhQqWrCMQ
This commit is contained in:
ekcSchiano
2026-11-12 01:31:11 +00:00
co-authored by Claude Opus 5.5
parent 8f1f0383fe
commit 8713ccff1a
11 changed files with 4403 additions and 2579 deletions
+13 -13
View File
@@ -94,8 +94,8 @@
"description": "Every run from every source, one per row, coded into the same columns (collection/etl/runs.py; the form's rules are core/spec/form.yaml). The `source` column says where each run came from.",
"contentUrl": "https://ekcschiano.github.io/threshing-atlas/data/runs_clean.csv",
"encodingFormat": "text/csv",
"sha256": "522c8501585fa6f067ec5a41dc34f79bc0bf5b262b3d3fb20d6f07723fa3cfaf",
"contentSize": "2300306 B"
"sha256": "0db0627c485ae1a8d21b2fa51a255234fcd809483932298154e52bbfc9cc34e4",
"contentSize": "2910779 B"
},
{
"@type": "cr:FileObject",
@@ -104,8 +104,8 @@
"description": "The community \"Threshing Routes\" form export as published (headers and rows), with email addresses in free-text answers redacted. The source of runs_clean.",
"contentUrl": "https://github.com/ekcSchiano/threshing-atlas/blob/main/data/public/form_responses.json",
"encodingFormat": "application/json",
"sha256": "4c5791f78ce81f5e88f1fb99d933acd06d8223ff9a573e96b6fcb3a6fd9c8d22",
"contentSize": "8293395 B"
"sha256": "51c768a83110799e322d25437da9be07583693dd6fdb592e49db63afef8f5d2c",
"contentSize": "10531911 B"
},
{
"@type": "cr:FileObject",
@@ -124,8 +124,8 @@
"description": "The Unofficial Dragon Register (which dragon each rider bonded), without rider names.",
"contentUrl": "https://ekcschiano.github.io/threshing-atlas/data/register.json",
"encodingFormat": "application/json",
"sha256": "f889c7e18068a934eb6872efda9b43599f4870d2e42bfdb3d447405976d59df8",
"contentSize": "111095 B"
"sha256": "0986dee41088e2649ed235631538371cf2f1d9efd5d663b1420727d57ea58776",
"contentSize": "114247 B"
},
{
"@type": "cr:FileObject",
@@ -134,8 +134,8 @@
"description": "Every number the Atlas page shows, for six filter presets (analysis/stats/site_stats.py; what is measured is analysis/stats/measures.yaml).",
"contentUrl": "https://ekcschiano.github.io/threshing-atlas/data/stats.json",
"encodingFormat": "application/json",
"sha256": "dc79c0762e5c27423f9fa2086f06aa7019b0aec6ae41d6a6dcbe59a1016cffab",
"contentSize": "4269546 B"
"sha256": "5150824f30be790aabad428611b41769d8b07fe40d1c6e3b75324a0173db9f42",
"contentSize": "5000518 B"
},
{
"@type": "cr:FileObject",
@@ -144,8 +144,8 @@
"description": "The route model: the game (core/spec/game.yaml), the count tables behind every roll, and the best decision tree for every dragon colour under two goals, with posterior intervals (planner/build_policy.py).",
"contentUrl": "https://ekcschiano.github.io/threshing-atlas/data/policy.json",
"encodingFormat": "application/json",
"sha256": "26c3adad2236493b4a77b7d46454fe01c84e8f750f0f03b6e7dbe968af32ea4e",
"contentSize": "617892 B"
"sha256": "b518a59da7173a4868873ed09f2a69014f0fea719b61d216faa43b758f12cfa3",
"contentSize": "596891 B"
},
{
"@type": "cr:FileObject",
@@ -154,8 +154,8 @@
"description": "The route graph over every run in runs_clean.csv, counted per source (analysis/network/build-graph.js).",
"contentUrl": "https://ekcschiano.github.io/threshing-atlas/data/graph.json",
"encodingFormat": "application/json",
"sha256": "53e9c7824ebe417c5950c24627ff288ceb74850a3d0f8b4cf97bae97f4c5daa7",
"contentSize": "71944 B"
"sha256": "d4e3a445b9303e95ad4e0e42679544f7ead1881ed9e6a8323f7a1dcf87063fd1",
"contentSize": "73272 B"
}
],
"recordSet": [
@@ -217,7 +217,7 @@
"@type": "cr:Field",
"@id": "runs/bot_mode",
"name": "bot_mode",
"description": "For bot runs, how the bot chose: \"auto_learn\" (its own rules and the route PDF), \"atlas_policy\" (the Atlas's route model), \"explore\" (trying untested options), \"experiment\" (a randomised test), \"manual\" (a person clicked). Matters for analysis: the bot follows a strategy, so its choices are not like a player's. Empty for other sources. Values: auto_learn; atlas_policy; explore; experiment; manual.",
"description": "For bot runs, how the bot chose: \"auto_learn\" (its own rules and the route PDF), \"atlas_policy\" (the Atlas's route model), \"atlas_thompson\" (the route model, sampling routes in proportion to how likely each is best), \"explore\" (trying untested options), \"experiment\" (a randomised test), \"manual\" (a person clicked). Matters for analysis: the bot follows a strategy, so its choices are not like a player's. Empty for other sources. Values: auto_learn; atlas_policy; atlas_thompson; explore; experiment; manual.",
"dataType": "sc:Text",
"source": {
"fileObject": {