Data Coverage
what the system can train on today — normalized games, odds, movement, injuries, priors and live market ticks, per sport
269,064 gamessnapshot 2026-07-25Per-sport training inventory
Soccer
PRIMARY- games
- 224,272
- range
- 2005–2026
- competitions
- 30+ divisions · 16 extra countries
- odds cov.
- 99.8
close — presentmove — presentinj — absentprior — presentticks — present
sft · T3 flagship source — labels are de-vigged soccer closing lines; per-sport JSONL emitted for the soccer adapter
Cricket
QUEUED #2- games
- 10,468
- range
- 2002–2026
- competitions
- 13 comps — Tests, ODI, T20I, IPL + 9 franchise/100-ball leagues
- odds cov.
- —
close — absentmove — absentinj — absentprior — absentticks — present
sft · T1 catalog examples; no T3 (closing-odds labels require an odds source)
Ball-by-ball Cricsheet archives are in hand (raw) — the §9 feature catalogue is derivable once cricket goes online
NBA
DATA-READY- games
- 27,063
- range
- 2005–2025
- competitions
- 1 — NBA (balldontlie schedules/results)
- odds cov.
- 73.0
close — presentmove — absentinj — presentprior — absentticks — present
sft · T1 catalog examples; T3 offset inputs limited (closing lines only, no movement series)
NFL
DATA-READY- games
- 7,261
- range
- 1999–2025
- competitions
- 1 — NFL (nflverse)
- odds cov.
- 72.7
close — presentmove — absentinj — presentprior — absentticks — present
sft · T1 catalog examples; T3 offset inputs limited (closing lines only, no movement series)
Esports · MMA · Tennis +
CATALOG-ONLY- games
- —
- range
- —
- competitions
- Baseball, hockey, golf, motorsport, boxing, NCAA also cataloged
- odds cov.
- —
close — absentmove — absentinj — absentprior — absentticks — present
sft · T1 only — live catalog titles feed market-mapping examples
Cross-sport stores
- injury records
- 38,552
- line-movement snapshots
- 386,067
- · across matches
- 91,715
- ClubElo priors
- 594
- SFT train / val / test
- 34,771 / 4,405 / 4,362
Injuries are NBA+NFL (append-only as_of snapshots); line movement is soccer (T-72h→close, ≤32 books); SFT mix design: T3 reconcile 50% · T1 mapping 25% (M7). Per-sport T1/T3 counts aren't published yet — cards state roles, not invented numbers.
Market families × sports
| Family | What it prices | SOC | CRI | NBA | NFL | TEN | MMA | ESP | OTH |
|---|---|---|---|---|---|---|---|---|---|
| moneyline | Who wins — two-way or three-way (with draw) contest result | SOC: traded on PM/Kalshi today | CRI: traded on PM/Kalshi today | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: traded on PM/Kalshi today | MMA: traded on PM/Kalshi today | ESP: traded on PM/Kalshi today | OTH: traded on PM/Kalshi today |
| spread | Winning margin vs a handicap line — points, goals, maps, rounds | SOC: traded on PM/Kalshi today | CRI: taxonomy-ready, not observed trading | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: taxonomy-ready, not observed trading | MMA: taxonomy-ready, not observed trading | ESP: traded on PM/Kalshi today | OTH: taxonomy-ready, not observed trading |
| total | Over/under on a counted quantity — goals, points, runs, maps | SOC: traded on PM/Kalshi today | CRI: taxonomy-ready, not observed trading | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: taxonomy-ready, not observed trading | MMA: traded on PM/Kalshi today | ESP: traded on PM/Kalshi today | OTH: taxonomy-ready, not observed trading |
| correct_score | The exact final score or exact series scoreline | SOC: traded on PM/Kalshi today | CRI: taxonomy-ready, not observed trading | NBA: taxonomy-ready, not observed trading | NFL: taxonomy-ready, not observed trading | TEN: taxonomy-ready, not observed trading | MMA: not applicable | ESP: traded on PM/Kalshi today | OTH: not applicable |
| team_prop | Team-level side markets — both teams to score, clean sheet, team total | SOC: traded on PM/Kalshi today | CRI: taxonomy-ready, not observed trading | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: not applicable | MMA: not applicable | ESP: taxonomy-ready, not observed trading | OTH: taxonomy-ready, not observed trading |
| player_prop | One participant's stat line — goals, yards, wickets, kills | SOC: traded on PM/Kalshi today | CRI: taxonomy-ready, not observed trading | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: taxonomy-ready, not observed trading | MMA: taxonomy-ready, not observed trading | ESP: taxonomy-ready, not observed trading | OTH: taxonomy-ready, not observed trading |
| outright | Season/tournament futures — champion, awards, season win totals | SOC: traded on PM/Kalshi today | CRI: traded on PM/Kalshi today | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: traded on PM/Kalshi today | MMA: taxonomy-ready, not observed trading | ESP: traded on PM/Kalshi today | OTH: traded on PM/Kalshi today |
| series | Multi-game series winner and best-of series score | SOC: taxonomy-ready, not observed trading | CRI: taxonomy-ready, not observed trading | NBA: traded on PM/Kalshi today | NFL: not applicable | TEN: not applicable | MMA: not applicable | ESP: traded on PM/Kalshi today | OTH: taxonomy-ready, not observed trading |
| sequencenew | Next/first event — next goal, first TD, wicket in next over | SOC: taxonomy-ready, not observed trading | CRI: taxonomy-ready, not observed trading | NBA: taxonomy-ready, not observed trading | NFL: taxonomy-ready, not observed trading | TEN: taxonomy-ready, not observed trading | MMA: not applicable | ESP: taxonomy-ready, not observed trading | OTH: taxonomy-ready, not observed trading |
| methodnew | How the result happens — KO vs decision, method of dismissal | SOC: taxonomy-ready, not observed trading | CRI: taxonomy-ready, not observed trading | NBA: not applicable | NFL: not applicable | TEN: taxonomy-ready, not observed trading | MMA: traded on PM/Kalshi today | ESP: not applicable | OTH: taxonomy-ready, not observed trading |
| placementnew | Race/ranking finishes — podium, top-N, head-to-head position | SOC: not applicable | CRI: taxonomy-ready, not observed trading | NBA: not applicable | NFL: not applicable | TEN: not applicable | MMA: not applicable | ESP: taxonomy-ready, not observed trading | OTH: traded on PM/Kalshi today |
| combonew | Multivariate contracts — “Team A wins AND total over X” (Kalshi multivariate events) | SOC: taxonomy-ready, not observed trading | CRI: taxonomy-ready, not observed trading | NBA: taxonomy-ready, not observed trading | NFL: taxonomy-ready, not observed trading | TEN: taxonomy-ready, not observed trading | MMA: taxonomy-ready, not observed trading | ESP: taxonomy-ready, not observed trading | OTH: taxonomy-ready, not observed trading |
| micro | Sub-minute in-play micro-events — next ball, next point | SOC: gated — paid feeds + fast pricing loop required | CRI: gated — paid feeds + fast pricing loop required | NBA: gated — paid feeds + fast pricing loop required | NFL: gated — paid feeds + fast pricing loop required | TEN: gated — paid feeds + fast pricing loop required | MMA: gated — paid feeds + fast pricing loop required | ESP: gated — paid feeds + fast pricing loop required | OTH: gated — paid feeds + fast pricing loop required |
| special | Catch-all yes/no event props — overtime, red card, record broken | SOC: traded on PM/Kalshi today | CRI: taxonomy-ready, not observed trading | NBA: traded on PM/Kalshi today | NFL: traded on PM/Kalshi today | TEN: taxonomy-ready, not observed trading | MMA: traded on PM/Kalshi today | ESP: traded on PM/Kalshi today | OTH: taxonomy-ready, not observed trading |
● traded on PM/Kalshi today · ○ taxonomy-ready, not observed trading · ⨯ gated — paid feeds + fast pricing loop required · — not applicable
14 market families (blueprint §6) — new marks the four adopted pre-M8 so the T1 adapter learns the full label space once: sequence, method, placement, combo. Cells reflect blueprint §3/§7 platform observations at snapshot date plus our gating rules; micro stays gated by locked decision #8 (paid feeds + a fast classical pricing loop — a 975B MoE over an API is never in a sub-second path).
Trainable data points per sport
Soccer9/23 in handblueprint §10
Team strength & priors
- Elo ratings (computed pregame, leakage-tested)
- ClubElo priors (as_of-stamped)
- opponent-adjusted recent form
- home-advantage term
- referee tendenciesroadmap
- fixture importanceroadmap
xG family
- xG / xGAroadmap
- expected assistsroadmap
- shot location & qualityroadmap
- pressures / PPDAroadmap
- field tiltroadmap
- set-piece outputroadmap
Lineups & availability
- expected / confirmed lineupsroadmap
- formationroadmap
- injuries & suspensionsroadmap
- player minutes & fatigueroadmap
Schedule & congestion
- rest days (computed pregame)
- travelroadmap
- congestion windowsroadmap
Market state
- closing 1X2 odds (99.8% of games)
- T-72h→close line movement, ≤32 books
- 21-book de-vigged live consensus
- PM/Kalshi price ticks incl. websocket stream
Catalogue: docs/research/market-taxonomy-blueprint.md §10 (in-repo)
Cricket10/29 in handblueprint §9
Context
- format: Test / ODI / T20 / 100-ball
- competition & stage (13 comps)
- toss winner + decision (raw)
- venue
- playing XI (raw)
- pitch behaviourroadmap
- weather / dew / DLS stateroadmap
Batting
- per-batter form & rolling averagesroadmap
- strike rate / boundary % / dot-ball %roadmap
- phase splits: powerplay · middle · deathroadmap
- pace-vs-spin & matchup splitsroadmap
- entry position & partnership contributionsroadmap
Bowling
- economy / wickets / strike rateroadmap
- phase usage & death-over performanceroadmap
- batter matchupsroadmap
- workload & recent oversroadmap
Fielding
- catches & dropsroadmap
- run-outs & direct hitsroadmap
- keeper statsroadmap
- runs savedroadmap
Ball-by-ball stream
- per-delivery events (Cricsheet raw archives)
- innings totals
- partnership state (raw)
- required-rate state (raw)
Derived
- Elo streams (men-only, computed)
- WASP-lite resource %roadmap
- wicket / boundary hazardroadmap
- collapse probabilityroadmap
- venue-adjusted par scoreroadmap
Catalogue: docs/research/market-taxonomy-blueprint.md §9 (in-repo)
NBA5/14 in handblueprint §12
Pre-game
- schedules & results 2005–2025
- moneylines (73% of games)
- injuries & minute restrictions (as_of snapshots)
- rest & back-to-backs (computed)
- pace / off-def ratingsroadmap
- expected lineupsroadmap
Box score & live
- shot attempts & locationsroadmap
- eFG% / rebounds / assists / turnoversroadmap
- lineup combinations & usageroadmap
- plus-minus / on-offroadmap
Derived
- margin-Elo + Gaussian margin model (computed)
- possessions remainingroadmap
- player minute projectionroadmap
- foul-out / fatigue riskroadmap
Catalogue: docs/research/market-taxonomy-blueprint.md §12 (in-repo)
NFL5/16 in handblueprint §11
Pre-game
- schedules & results 1999–2025
- moneylines (72.7% of games)
- injury reports (as_of snapshots)
- rest days (computed)
- QB statusroadmap
- weather & windroadmap
- coaching tendenciesroadmap
Play-by-play
- down & distance stateroadmap
- EPA / win probabilityroadmap
- drive resultsroadmap
- formation & personnelroadmap
Efficiency splits
- margin-Elo (computed)
- pressure & sack rateroadmap
- explosive-play rateroadmap
- red-zone / third-down efficiencyroadmap
- pace & neutral pass rateroadmap
Catalogue: docs/research/market-taxonomy-blueprint.md §11 (in-repo)
Esports1/14 in handblueprint §27
Universal
- game title / version / patchroadmap
- roster version & substitutesroadmap
- best-of formatroadmap
- map pool & veto orderroadmap
- side selectionroadmap
MOBA
- draft picks & bansroadmap
- gold / XP curvesroadmap
- objectivesroadmap
- item timingsroadmap
Tactical shooters
- map & side economyroadmap
- opening duelsroadmap
- clutchesroadmap
- utility usageroadmap
Market state
- PM/Kalshi catalog + price ticks (12 title families observed)
Catalogue: docs/research/market-taxonomy-blueprint.md §27 (in-repo)
MMA · Tennis1/11 in handblueprint §23 · §15
MMA
- reach / stance / age / weight-cutroadmap
- layoff & recent activityroadmap
- strikes landed & absorbedroadmap
- takedowns & control timeroadmap
- method & round historyroadmap
Tennis
- surface-adjusted ratingroadmap
- serve / return splitsroadmap
- head-to-head & style matchuproadmap
- fatigue & match-duration historyroadmap
- point-by-point stateroadmap
Market state
- PM/Kalshi catalog + price ticks
Catalogue: docs/research/market-taxonomy-blueprint.md §23 · §15 (in-repo)