Published 2026-10-14 · Updated 2026-10-25
Information coefficient
The information coefficient (IC) measures how well a signal predicts forward returns — usually the correlation between a factor and the next period return. It is small in absolute terms, and that is normal.
On this page
What it is
The information coefficient is the correlation between a signal and the forward return it is meant to predict.
It is the standard yardstick for a factor: does the score line up with what happens next?
Why small is normal
A daily IC of 0.02–0.05 is often considered useful. That surprises people.
A small IC is not a weak result — it is what a real edge looks like. Because the bet is repeated across thousands of names and days, a tiny consistent edge compounds.
| IC | Reading |
|---|---|
| ~0 | No relationship |
| 0.02–0.05 | Typical useful range |
| > 0.1 sustained | Suspicious — check for a leak |
How to check
Ask whether the IC is out of sample, and whether the forward return was measured from a point where the signal was actually known. A high IC is a reason to look for look-ahead bias, not a reason to celebrate.