WPA-POWERED BASKETBALL ANALYTICS

Who contributes to winning — at every level.

Alleygorithm is a validated win-probability system for player impact and group contribution. It measures who drives winning, which combinations have actually worked, and how much confidence to place in the sample.

Player impact

Find who drives winning on their own minutes.

Pair contribution

See which duos have actually driven winning together.

Lineup strength

Move from names to groups with the same contribution language.

Players
Pairs scored
Possessions
0.22
Prediction r

Read The Product

The model is only useful if the interface teaches you how to read it.

Alleygorithm is designed around three questions: who drives winning, which groups have actually contributed, and how much confidence you should place in the answer.

Player signal

Start with win contribution

This is the public-facing impact score: a possession-level winning signal, normalized per 100 possessions and stabilized for public use.

Group signal

Use group contribution first

Pairs and lineups are most useful as contribution views: which combinations have actually driven winning together. Surplus is an advanced read, not the default story.

Decision signal

Read possessions with confidence

Small-sample groups can look extreme. Possessions, shrinkage, and confidence badges help separate story-worthy signal from early noise.

Who wins together

Win contribution per 100 possessions, EB-shrunk

Individual impact

Shapley WPA per 100 possessions

All players

Best lineups

5-man groups by win contribution

Explore all lineups

How it works

01
Win probability model

LSTM trained on play-by-play data predicts home win probability at every event. WPA = change per possession.

02
Shapley attribution

Ridge regression assigns WPA to players (coefficients = Shapley values). Cross-season r=0.22.

03
Every grouping level

Score players, pairs, trios, quartets, lineups. Empirical Bayes shrinkage. Confidence tiers.