Clutch Hits - Forecasting Pitchers (January 20, 2004)
For those who have not followed the PECOTA thread completely:
Here's a little study I did on pitching forecast. I've got to believe that the reason that the forecasters did so poorly with the pitching forecast is their lack (or little) of regression towards the mean.
Next time you see a forecast, ask the forecaster this question: "how much do you regress?". Chances are, they regress way too low.
--posted by TangoTiger at 02:11 PM EDT