SHEGOTGAME
Methodology · Our Seven Models

The GM's Read™

Every pick starts with The GM's Read™ — our proprietary blend of five statistical scouts — then gets fine-tuned by six situational models that read the game like a coach.

The head coach

She Got Game runs on seven proprietary models. The GM's Read is our head coach — it weighs recent form, matchup, availability, usage, and line value into a single confidence score. Then our six situational models (SSI™, Jet Lag Index™, Gas Tank Index™, Mental Load™, Mind Lock-in Score™, BURST™) act as scouts whispering in the GM's ear, nudging the read up or down based on things no box score can see.

The GM's Read: 91 confidence · HIGH
Boosted by: BURST (+2), Jet Lag Index (−1)

The five scouts inside The GM's Read

The base engine blends five statistical scouts into one number:

  • Recent Form — how the player has actually produced lately.
  • Matchup — opponent defense and pace vs this stat.
  • Availability & Context — status, minutes, and role tonight.
  • Usage — how central the player is to the offense.
  • Line Value — where the book's number sits vs the model.

The six situational models

Six scouts whisper in the GM's ear, nudging the read up or down:

SSI™
Speed/Shooting Index
A situational modifier that reads shot quality and pace-of-play signals the box score misses.
Jet Lag Index™
Travel fatigue
Circadian misalignment and cross-country travel legs that quietly sap a player’s edge.
Gas Tank Index™
Rest / energy
Rest days, back-to-backs, and cumulative minutes load — how full the tank is at tip-off.
Mental Load™
Cognitive fatigue
The cognitive tax of a heavy schedule and high-leverage stretches that dulls sharp play.
Mind Lock-in Score™
Focus / engagement
A focus and engagement read — who’s dialed in tonight and who’s coasting.
BURST™
Speed Scoring Index
An explosive-scoring modifier that flags players primed for a high-ceiling scoring night.

Which markets are we betting right now?

The backtester filters picks through per-market gates. A market can be paused entirely if the historical edge isn't reliable, and each enabled market has a minimum edge (in points) and a Kelly staking cap so we don't oversize into noise. Right now: 8 enabled, 2 paused.

MarketStateMin edgeKelly capκ (dispersion)
AssistsastEnabled≥ 4.0 pt3.0%
Pts+AstpaEnabled≥ 3.0 pt4.0%
Pts+RebprEnabled≥ 3.0 pt5.0%
Pts+Reb+AstpraEnabled≥ 3.0 pt5.0%
PointsptsEnabled≥ 5.0 pt3.0%
Reb+AstraEnabled≥ 3.0 pt5.0%
ReboundsrebEnabled≥ 5.0 pt5.0%
3-Pointers MadethreesEnabled≥ 3.0 pt4.0%
Double-DoubleddPaused
Triple-DoubletdPaused

Gates update daily from live backtest results. A paused market means our current models don't clear the historical noise floor there — we'd rather show you nothing than a coin flip. The κ column is the empirical spread of L10 residuals per market: higher values = harder to project.

Think of it like a front office

The GM makes the call — but only after five scouts and six situational analysts have their say.