Just Say No

by Charlie Wilmoth on Aug 5, 2005 12:44 AM EDT in News

The Pirates are considering bringing Jose Mesa back next year. Just like last offseason, this is a really bad idea. The Pirates will pay $2.5 million for Mesa's services this year - $2 million plus a $500,000 buyout. Despite Mesa's junk peripherals and the fact that he looks like he's about fifty (I don't mean to play the often unfair Latin-player-guess-the-age game, but just look at Mesa's face and try to convince yourself he's still in his 30s), he has posted a 3.76 ERA this year.

That's mediocre for a reliever, but it's only a half run higher than last year's, so expect the Pirates to pay him about $2 million again. That is too much to pay a player who is very likely to collapse (in the baseball sense and in the literal "I've fallen and I can't get up" sense), not very effective in the best of circumstances and only good for about sixty innings a year.

The management's reason for their desire to resign Mesa is depressing: "Management will be reluctant to have [Mike Gonzalez] open next season as the closer without some experience in the role."

Closing is not magic. This year, the Pirates' suddenly smarter cousins in Milwaukee flipped their established closer, Dan Kolb, for a good starting pitching prospect, then handed the closer's job to Derrick Turnbow, whose only previous claim to fame was testing positive for steroids. Turnbow has posted a 2.09 ERA for the Brewers this season, while Kolb lost the Braves' closer job. Turnbow's salary this year is $322,000.

The Pirates wouldn't necessarily be able to find a Turnbow if Mesa left, of course, but Mesa is no Turnbow himself. And anyway, the class of pitchers who are capable of closing is much larger than the class of pitchers who actually close, so dropping another $2 million for another 60 innings of work isn't the best idea. $2 million isn't a lot to pay a closer, it's true, but if you spend $1 million here and $2 million there, it adds up quickly for a team like the Pirates.

Also, if experience is the issue, Mark Corey is having a fine year as the closer for Class AAA Indianapolis. If Mike Gonzalez or someone else can't be trusted with the job, Corey could probably do a reasonable imitation of Mesa for the league minimum salary.

Tweet Comment 2 comments | 0 recs |

Do you like this story?

Comments

Minor league predictiveness

Not sure where else to ask this:

You've mentioned the predictiveness (sp/use?) of minor league stats for ML performance. I wonder, is there any predictabilility for who is likely to become a AAAA player, like, say, Karim Garcia?

I could be misreading his ability, but he put up good minor league numbers, in the PCL but also in the IL at age 24. Was he never really given an opportunity?

Is it something else, like a poor glove, that holds these guys back? Is the 4A player a myth?

Would Adam Hyzdu not be considered because he didn't really dominate the minors until older? Is age a key?

Feel free to point me to a website if that's easier.

azibuck

by azibuck on Aug 5, 2005 12:11 PM EDT reply actions

MLEs

http://www.baseballthinkfactory.org/btf/scholars/czerny/articles/calculatingMLEs.htm

This article explains how MLEs work pretty clearly. If you google "Bill James" "MLE" you'll find lots of other stuff, too.

There are lots of things that hold people back from believing in MLEs but that do not, I don't think, disprove them. The first is that using one year's stats to predict the next is often full of variability, but this is true when using major league stats as well. For example, if a player puts up a .900 OPS in the majors one year, he may do much worse the next year. If a player puts up a translated OPS of .900 in the minors, he may also put up something different in the majors the next year. People see evidence like that and assume there must be such thing as a AAAA player, when it's possible the player just had a career season or peaked. However, MLEs are just as broadly predictive as major league stats.

Another thing is that MLEs have to be translated, and this can cause problems. Obviously, the playing in AAA is not as good as it is in the majors, so you have to adjust for that. Also, the PCL is a very, very extreme hitters' league. This means that sometimes a player like Garcia can put up ridiculous numbers in the PCL, but his MLE would look nothing like those numbers.

Finally, players going from AAA to the majors often have to compete for playing time, so the temptation to draw conclusions from small sample sizes is very strong.

These factors lead people to conclude that there are 4A players. And in a sense, there are: someone like Hyzdu can hit 30 homers a year at Altoona, which seems great, but when you consider that he was doing that when he was a very old player there (ie, not likely to get better) and the pitching at AA is much, much worse than the majors, you can guess that Hyzdu's not going to do much more for you than be a functional 4th outfielder, which is basically what he is. However, there's nothing magic about it, I don't think - it's a difference in the quality of the league, nothing more.

I looked at Garcia's numbers. The thing is, Garcia played in the PCL when he was young, and he played in Tucson and Albequerque, which are two of the craziest hitters parks in all of pro baseball. If you adjust for park and league factors, they're crazier than Coors Field. So if you just look at those numbers without translating, it's no surprise that his major league numbers didn't resemble those.

When Garcia moved to the International League (Toledo, Buffalo, Columbus) his numbers levelled off a lot, which we'd expect. His career minor league OPS (and btw, MLEs only work for AAA, and, to a lesser extent, AA), including the nutty seasons in the PCL, is .865. His career major league OPS is .703. I bet if you adjusted for the quality of the pitching at AAA and the ballparks he was playing in, those numbers would actually end up about the same.

by Charlie Wilmoth on Aug 5, 2005 2:35 PM EDT reply actions