sure, but charting bad decisions/throws (or "should have been" interceptions) would be more predictive than just counting interceptions and reduces variance/luck in the data. in the long run it should even out, but in 8 games, one guy could pretty easily luck his way into significantly fewer (or more) picks than he should have. but you know this. Sure. But absent that data, actual interception totals are a better proxy than actual INTs plus remembering some bad throws. But you know this.