Engine internals

Shapley attribution

In plain English

A fair way to split the credit/blame for the score among the 13 factors — each factor gets exactly its marginal contribution.

How it works

Shapley 1953 game-theoretic value: across all permutations of factor inclusion, what's factor X's average marginal contribution to the final composite? This gives a cleaner attribution than simple "factor i has weight w_i × score" because it accounts for factor interactions. The Score Composition waterfall on every ticker page shows Shapley contributions.

Where you see this in Framler
Score Composition waterfall on /stocks/[ticker]. The +X.X numbers are Shapley contributions.
Primary citation
Shapley 1953, Value of an n-person game

Related — Engine internals

Framler score (composite)Verdict (BUY/MIXED/SELL)Forward return (expected)Market regime (BOCPD)Conformal prediction intervalTail-dependence (copula)

← All glossary terms · Full methodology →

Shapley attribution — Framler glossary | Framler