« Time Fades Away | Main | Feliz! »

July 03, 2009

Who You Callin’ Lucky?

Evan Grant has taken umbrage to ESPN’s Rob Neyer and Fangraphs’ R.J. Anderson regarding their skepticism of Kevin Millwood’s success. Millwood is the “benefactor of an unsustainable amount of stranded runners [presently 86.8%, second best in baseball], thus keeping his ERA at a comfortable, and easily overrated, 2.64,� alleges Anderson. “Don’t be fooled by Millwood’s ERA,� asserts Neyer, citing his .261 BABIP.

Grant tells Neyer to “wake up� and further demurs:

It is very probably true that Millwood’s ability to strand runners will dip in the summer heat. After all, 86 percent is an astronomical number. But why is this viewed as a Millwood shortcoming.

Nobody is claiming Millwood’s exceptional strand rate is a shortcoming per se. Indeed, it’s a huge part of what has made him so effective. And that’s the problem. Millwood’s career strand rate is 71.5%. His best season is 79.1% in 2005, when he led the league in ERA. His present rate of 86.8%, if it holds up, would be the best in baseball since Pedro Martinez in 2000. Kevin Millwood is having a terrific season, no doubt, but Pedro Martinez he is not.

But attributing Millwood’s success to luck is asinine, first, and plainly against the crede of the stats analysts, second. Stats analysts don’t believe in luck. They believe in trends.
Frankly this statement simply isn’t true, or at least it’s misguided. “Luck� is Statspeak shorthand for a series of atypical occurrences; for example, stranding 87% of opposing runners for nearly half a season. Statheads attribute success to luck all the time. They do believe in luck in the sense that all players get lucky (or unlucky) sometimes. They don't believe in trends in the sense that luck is unsustainable in the long run.

Back to the broader point, Millwood’s performance is intriguing. His walk rate is typical for him. He’s not preventing homers exceptionally well. His strikeout rate is 14.8% the worst of his career. So, what is the source of his success?

Defense – Texas’s defense is unquestionably far better than last year.

BABIP, for better and worse – Yes, Neyer noted that Millwood’s .261 BABIP is the lowest of his career (his career BABIP is .306), which is to say, unsustainable, but I actually think Millwood might be on to something. He’s allowing fewer line drives (nearly automatic hits), more popups (automatic outs), and more grounders. Coincidentally, or perhaps not, he’s relying more heavily on his slider and changeup than ever before. Is the adjusted mix of pitches causing the beneficial change in his batted ball pattern? I can’t say, but it certainly isn’t hurting. The same applies to Millwood’s conditioning, as noted by Grant.

Luck – Yes, luck. There’s no way Millwood will continue to strand runners at such a high rate. Opposing hitters are batting .096 with two outs and runners in scoring position compared to .261 in other situations. Even if Millwood can sustain his low BABIP, he’s not going to keep retiring 90% of batters in two-out RISP situations.

Look, Millwood’s been a revelation, a pleasure to watch, and arguably the team MVP. But I’d bet his second-half ERA will be much closer to 4.00 than 2.64, and Texas will suffer a little as a result. I don’t see this as controversial or offensive. It seems pretty obvious. And I’ve watched him throw.

UPDATE: Hmmm. I googled "Kevin Millwood BABIP" and found that Yahoo's Scott Pianowski offered similar opinions, including the possible connection between his pitch selection and batted-ball outcomes, on June 27th. I didn't discover his article until after I posted mine. Scout's honor.

UPDATE 2: Lone Star Ball's Adam Morris posted a lengthier examination of Millwood that again pre-dates mine. And again, I didn't discover his article until after I posted mine. Scout's honor. I really was a Scout.

Posted by Lucas at July 3, 2009 04:07 PM