I don’t subscribe to Baseball Prospectus for many of the same reasons I don’t subscribe to cable tv. There are only so many hours in the day and paying for constant content that far outstrips my ability to consume it rankles me. With cable my problem is also how much they charge, and that there are commercials. With BP my problem is how often they congratulate themselves for modest achievements.
But today a friend sent me a Will Carroll column (from today) that contain’s Will’s usual excellent listing of current injuries and their prognosises, and also includes an open letter to Jim Palmer, who recently claimed (based on no evidence at all, except the homers he hit) that Brady Anderson used steroids the year he whacked 50.
I wrote about this two weeks ago, pointing out that absent evidence this is scurrilous slander of the lowest most egregious sort because Palmer admits he doesn’t know anything. Today, Will enlist the help of the most valuable Keith Woolner to come up with a list of players whose homer total jumped by more than 25 home runs in one season.
While Brady’s was the second biggest jump in the history of the game, the first biggest was Davey Johnson’s surprise season in 1973, and among other players who show up on the list are Lou Gehrig, Harmon Killebrew, Dick Stuart, Rico Petricelli, Cito Gaston and Bob Cerv. Unlikely juicers, every one.
This is fun stuff. Will’s column is chock full of very useful information. If you have sampled it you know that. If you haven’t and you follow baseball and are just a little bit obsessed by it BP is worth checking out. Just don’t worship these guys as if they were the Christ. And let’s keep an eye out for some industrious soul to really analyze what is happening in all the projections that are out there.
Because apropos my earlier rant this week about the BP projections, Nate Silver can be found on the BP website, on the front page, in a free article (which is basically a sales pitch for the book and stats on the site) about projections, laying out why PECOTA (or really Pecota, since acronyms of five or more letters should be upper and lower) is the best projections system around.
That got me to thinking about things some more, and I have to admit that if PECOTA is a correlation point or two better than other systems consistently then that is genuinely valuable. So, I’m going to shut up about this after this post, until I come up with a way to measure accuracy in a meaningful way. And then we’ll take a look at what’s going on here, there and everywhere.