This morning, Paul Swydan posted our Staff Predictions for the 2017 season. Because most of the people who write for FanGraphs, RotoGraphs, or The Hardball Times are at least a little bit data-oriented, the picks end up being pretty similar to what our projections say. The Cubs have the best team in the NL Central, and statistically, they have the best chance of winning the division, so they’re the logical choice for everyone to predict as the NL Central winner. Mike Trout is the logical pick for AL MVP, now that voters have shown they’ll give him the award even if his team doesn’t win, because he’s the best player in the league. And so on and so on.
But as we’ve noted before, projections are not predictions. Projections are the mean outcome in a probability distribution, and aim to identify the area in the middle of a bell curve. But the most likely outcome in of a series of possible outcomes may itself still be quite unlikely. A curve with probabilities of 10%, 15%, 25%, 25%, 15%, and 10% would result in a projection around the 25% probability marks, but you wouldn’t want to confidently predict that the 25% outcome is likely to occur, because 75% of the time, your prediction would be “wrong”.
So, if you look at the staff predictions table, you’ll notice that I made a few non-traditional picks. I went with the Yankees in the AL East, for instance, and I picked Lance McCullers to win the AL Cy Young. I don’t think the Yankees are the best team in their division, nor do I think McCullers is the best pitcher in the American League, but the fun thing about baseball is that, in one season, the results aren’t governed by the bell curves. Weird things happen, and since our predictions are just meaningless guesses, we might as well have fun with them and try to give ourselves a chance to say “I told you so!” in six months.
But I didn’t pick Lance McCullers to win the Cy Young just to be contrarian. While it’s pretty likely that a Chris Sale or a Corey Kluber has higher odds of winning right now, a healthy McCullers might have better odds than his reputation would suggest.
Read the rest of this entry »