Skip to content
Four sports · every number cited

The analytics,
and the receipt
for every number.

A bottomless, honestly-measured look at basketball, baseball, soccer, and tennis. No edge claims — only calibrated findings you can check.

Scout answers only from precomputed, receipt-cited data. It never invents a number and says NO_DATA when it has none.

The Glass Court
1,549
entity cards · 4 sports
74
analytics modules
287
mechanism verdicts
130 confirmed126 null31 not testable
96/96
reproducibility checks green

In plain terms: a mechanism verdict is one tested cause-and-effect claim; a null means we tested it and found nothing, and we keep those on purpose; a Brier score rates how calibrated a prediction is, where lower is better.

Four ways in

The Forecaster

The calibrated prediction engine, walked forward and paired against the devigged close — calibration, never a dollar edge.

17 markets scored out-of-sample · 3 sports
The Loop

The self-improving AI, keeping score on itself: propose, gate, verdict, then graveyard or verified.

259 forward-tested families · 5 verdicts flipped
Explore

Browse every analytic. Importance-weighted so a growing catalog stays scannable, nulls at equal prominence.

74 analytics modules
Ask Scout

Ask a plain question. Scout answers only from precomputed, receipt-cited data, and says NO_DATA when it has none.

precomputed · receipt-cited · no LLM at runtime

Six novel stats

See all six →
LHLnovel packaging
Line Half-Life

soccer_intl completes half its pre-game line motion by 2.7h before tip; nba not until 4.8h. tennis, wnba move most of their line >6h before tip.

LCFincremental novel
Live-Clock Fraction

mlb stays contested later (LCF 0.833) than soccer_intl (LCF 0.748). (nba comeback-rate cross-check 0.0146)

LBIincremental
Load-Bearing Index

Most one-star-fragile: DEN (Nikola Jokić, delta_winprob 0.5822). Two estimators name the same #1 player for 1/30 teams (directional cross-check, different seasons).

MFPincremental
Market Foresight Premium

mlb: MFP rises from 0.06 to 0.24 over the game (mean 0.25); soccer_intl: MFP rises from 0.64 to 1.71 over the game (mean 0.92)

OHGnovel self critical cross
Overreaction Harvest Gap

soccer_intl: overshoot 0.286 but model beats market only 22% of the time at max disagreement -> OHG 0.081 (a failure to harvest, not an edge); mlb: overshoot 0.072 but model beats market only 38% of the time at max disagreement -> OHG 0.009 (a failure to harvest, not an edge)

SFTincremental
Schedule Fatigue Tax

Most-taxed schedule: DEN 2025-26 (-0.38 pts/100 ORtg season-averaged); least: MEM 2025-26 (-0.27). League schedule-inequality range 0.11 pts/100 across 90 team-seasons.

Explore the analytics

All 74 modules →
confirmednull (a finding)not testabledescriptivepending
NBA · FatigueBack-to-Back Rest PenaltyMargin cost of a zero-rest game, leak-free over two seasons of team-game frames.
-1.73pts
NBA · In-gameWin-Prob vs Market, End Q1Paired Brier delta -0.0084 against the close — negative means the market is still sharper.
Market sharper
MLB · SkillBrier Skill Scores
Null
Cross-sportCalibration Scoreboard
17rows
Tennis · SurfaceSurface TransferHow a player's form carries across clay, grass, and hard courts.
278players
The honest railnulls (351) outnumber confirms (168) 2.1x -- we publish our nulls — The market is efficient; we aim to match the devigged close within noise. No dollar edge is claimed, anywhere. Every number links to the artifact that produced it, and the six retracted figures live only inside their retraction. We even deflate our own sample sizes down to the honest count of independent games.
Ask Scout about this
Does playing on the second night of a back-to-back hurt production?Does the model actually beat the betting market?Does a tennis prior trained on one tour transfer to the other?Ask anything →