On Millwood

One of the great things about the statistical boom and the internet is that you can have much more detailed arguments about a player.

10 years ago, vis-a-vis Kevin Millwood, the argument would have gone like this:

"Millwood sucks."

"No, he doesn't."

"Yes, he does, look at his ERA."

"Yeah, but he led the league in ERA a few years ago."

"Yo momma."

Nowadays, with all the new information that is readily available, it is more like this:

"Millwood sucks.  Look at his ERA last year"

"No, he doesn't.  His FIP is more relevant, and his FIP was pretty good last year.  He just had an unusually high BABIP, which isn't likely to be repeated."

"His tRA was terrible, though."

"That's because his LB% was higher than it has been historically."

"Exactly.  And that's why his BABIP was so high, and why he isn't as good as his FIP indicates, because he's throwing meatballs up there that the other team is driving all over the park."

"Yo momma."

So...recapping the data that's been kicked around this offseason, Kevin Millwood's Fielding Independent Pitching (his FIP, which is based on his peripheral rates) was pretty good last season -- a 4.02. 

His ERA was dramatically higher than that because his Batting Average on Balls In Play (BABIP) was .366 -- the highest in the majors last year among qualifying pitchers. 

Doctrinaire DIPS/McCrackenian theory says that Millwood can't really control whether balls in play go for hits or not, and thus it is a product of luck and defense.

The counter to that is that Millwood also gave up the highest rate of line drives in the majors last year, and that the high line drive rate was the result of him throwing meatballs up there for batters to hit, and thus the high ERA and high BABIP are to be expected, and are more indicative of his ability.  And thus, you should look at his tRA, which incorporates LD% (among other things), and reveals that Millwood's ERA last season was about what should have been expected.

The question then really becomes...how much control does a pitcher really have over his line drive percentage, and is it something that is that variable or tells that much about a pitcher?

Looking at the 10 pitchers with the highest line drive rates last season, Millwood, of course, is #1...but #2 on the list is Josh Beckett, a guy who isn't exactly known for throwing meatballs at the plate, and who had a pretty good year last year.  Matt Cain and Johan Santana also crack the top 10, and they are pretty good pitchers...but then you also have guys like Bronson Arroyo, Jeff Suppan, and Paul Byrd, who fall in the meatballing-strike-thrower category, and ciphers like Ian Snell and Barry Zito.

Nevertheless, my feeling is that if Millwood were just throwing BP pitches up there and that's why his line drive rate is so high, he'd be allowing a higher rate of home runs than he has.

Last year Millwood had the highest line drive rate he's had of his career -- his career percentage is 22.6%, and his previous two seasons in Texas his LD% was around 21%.

So what does all this mean?  I'm not sure, exactly.  Millwood's K and HR rates have been remarkably consistent since coming to Texas, and while his walk rates have fluctuated some, his walk rate last year wasn't that much higher than in 2006 (and was much lower than in 2007).

Bill James, CHONE, and ZiPS all expect him to be good, not great, this season, in large part, I would wager, because of solid peripherals over the past few seasons.  The projection systems expect him, in fact, to be worth more in 2009 than he's being paid, strange as that may seem.

But the projection systems also don't expect him to have a BABIP close to .370 again.  Now, the BABIP should come down anyway, because of the better defense behind him, but it really is a question of how much.

So whether or not you think Millwood will be productive (aside from any potential health issues) largely depends on whether you think the line drive rate and BABIP last year was an aberration, a random spike that will level out this year, or if you think he's going to allow a 25% line drive rate and a BABIP north of .350 again.

If you think that the line drive rate will come back down to the 21-22% rate, you probably expect he'll perform about like he did in 2006.

If not, you are expecting a repeat of 2008.