Fixing My Fantasy Weakness: Hitter Evaluation
I was recently asked the following question by Werthless.
Jeff, what are you trying to accomplish here? Are you trying to estimate the volatility in an individual player’s projection? That’s an interesting question, and directly related to the risk of the player. Are you trying to do better than Steamer at predicting performance? That’s a big endeavor. Are you trying to predict injuries? Might be better to do that directly. Are you trying to better estimate number of plate appearances by estimating job security? Might be better to do that directly.
Then, you can combine the models to perhaps better quantify a player’s risk of meeting preseason performance objectives. You can apply your model onto a different year’s data to see how well your predictions match reality (ie. Do the higher risk players actually underperform more often than lower risk players).
I do have a plan I’m implementing but it wasn’t known to my readers. Sorry. I want to understand which hitter traits to concentrate on. If they don’t exist, I created some.
For a few season’s now, my hitters have steadily outperformed my pitchers. In my three main leagues, here are the pitching-hitting splits from this past season.
Read the rest of this entry »