Skip to content
AI supply chain term

Information coefficient (rank IC)

The rank correlation between one night’s conviction scores and the price moves that followed - did higher-scored names actually move more?

What it means

The information coefficient asks the sharpest question you can put to a scoring engine: across all names on a given recorded night, did the ranking predict the ranking of what happened next? THE ENTITY computes it as a Spearman rank correlation between the stored scores of a past night and each name’s forward return over the seven days that followed (grading at T-7 is by design: the night being graded is always a week old, so every grade is fully out-of-sample). An IC of zero means the ordering carried no information; sustained positive IC means the engine’s ranking genuinely led prices. The Scorecard shows the rolling record, including per-component ICs.

Why it matters to investors

IC is harder to fluke than a hit rate - it grades the whole ranking, not just the loudest calls. A modest but persistent positive IC on real out-of-sample nights is worth more than any single spectacular call.

Related terms

See Information coefficient in the live AI chain.

THE ENTITY maps every constraint onto one live model - which part is tight now, who owns it, and who gets squeezed when it moves. Plain-English reads you can check.

THE ENTITY is an educational read on the AI supply chain - research, not investment advice. It explains how the chain works and who sits where, never price targets or buy/sell calls.