Information Coefficient
The rank correlation between predicted scores and forward returns — how well a quant signal predicts the future.
Definition
The information coefficient (IC) is the Spearman rank correlation between a signal's predicted ranking of stocks and their actual forward returns over a chosen horizon. An IC of 0 means the signal is no better than random; an IC of 1.0 would mean perfect ranking; an IC of -1.0 would mean inverted-perfect ranking.
In practice, robust quant signals produce ICs in the 0.02–0.10 range. That sounds tiny, but it adds up: an IC of 0.05 sustained across thousands of stocks and hundreds of rebalances produces meaningful aggregate alpha.
The IC matters because it isolates predictive power from execution. A signal can have a high IC but generate poor returns if implemented with high turnover or in capacity-constrained names; conversely, a signal with mediocre IC can produce good returns if it's implemented cheaply at scale.
How QScoring uses it
IC is part of QScoring's public validation commitment. Before subscription billing turns on, the methodology page will publish IC values for the QScore against 1-month, 3-month, 6-month, and 12-month forward returns, plus a rolling-window IC analysis to show stability over time. As of today the IC has not been published; the QScore is presented as a transparent synthesis of well-known factor research, not as a validated predictive signal.