Back in September, I devoted a few hundred brief words to the concept of Minor League Equivalancies ("MLEs"), as applied to a couple of upper-level prospects in the Nationals' organization, Kory Casto and Larry Broadway. Today, I will take a far more comprehensive look at the MLEs of players in the Nats' chain, though this is owing to no labor or ingenuity on my part. Instead, Jeff Sackmann has kindly transmitted this data to me, for me to disseminate as I will.
Before I do so, which will appear in successive posts for each affiliate (from Triple-A New Orleans to Low-A Savannah), I want to cover a couple of bases. The first is to commend you to become a regular reader of Jeff's material. Jeff makes this task quite easy, because he is nearly everywhere these days. He maintains the excellent Brew Crew Ball (and contributes to Packer Backer Blog, which makes me think he has a Wisconsin connection). He writes for Beyond the Boxscore, the only sabermetrics site where you can vote on your favorite projection system. (The choices are Bill James, "CHONE," Marcel the Monkey, "PECOTA," Ron Shandler, and "ZiPS.") He also contributes to The Hardball Times ("Baseball. Insight. Daily.") and Heater, a fantasy baseball magazine. And, last but not least, Jeff is the mastermind behind the Minor League Splits website, in which capacity Jeff provides this data today. What a great website. Not only would we have not really known Kory Casto's struggles against lefties in '06 without it, but it's also the inspiration for this barrage of imagery:
So there. At any rate, I don't want to make my praise too effusive, but to the extent my praise is too effusive, I do have cause for it. Jeff does great work, and I thank him for his generosity.
Now, for the second thing. What's an MLE? Essentially, it is a translation of a minor league season's statistics into a big league context. I provided a longer explanation in my September post, but that was way back in September, and Jeff's explanation is much more cogent:
All MLEs are based on park-adjusted stats. . . .
Multipliers for MLEs are generated as follows. To discover the relationship between leagues, find the population of players who played in both leagues in a certain year. If there are enough players, you'll get a general idea of how those players performed at each of the two levels. Weight their contributions for playing time, so that a player who had 100 plate appearances at each level has more impact on the calculations than one who had 500 PAs at one level and 4 at the other. Since players rarely play at more than two levels in a season, the relationship between, say, the Carolina League and the Major Leagues is determined by finding intermediate relationships that we do have data for--say, Carolina League/Southern League; Southern League/International League; International League/Majors. Obviously the math here gets a bit hairy; this is just intended as a sketch to convey the general idea.
To wrap up, a couple more notes:
- Jeff does not produce MLEs for short-season-A and Rookie League players; thus, there the Vermont guys won't be found in the next few posts, nor will the Wave of the Future (i.e., Marrero, Williams, King, Englund).
- MLEs are not projections, they are just translations of what happened last year.