Tuesday, September 23, 2008

A-Rod sucks

So say Michael Kay and Al Leiter. It's more of the same "You gotta watch the games" garbage that always gets thrown around.

People, get this through your skulls: perception is not reality. Just because you perceive a pattern does not mean it exists. If I could communicate just one idea to people when it comes to analysis, this would probably be it.

A-Rod hasn't hit with men on base this year. So what? He'll probably finish 25 RBIs off his excellent-even-by-his-own-lofty-standards 2007. The Yankees are going to finish nearly 200 runs shy of their total from last year. Were you expecting Alex to drive in 300+ runs?! Surely there must be some other explanation. In fact, not only is it impossible to find any real evidence that hitting with men on base is a separate skill from simply hitting in general, but the difference between A-Rod's so-called "clutch" numbers and his overall numbers is vanishingly small.

What's the difference between A-Rod's pathetic .232 average with RISP and two out and a .300 batting average?

Five hits.



What could we expect those hits to net the Yankees? Five runs? Seven? Ten? If A-Rod's five missing hits give the Yankees ten more runs, they'd likely be only five game behind the Red Sox instead of six. With six to play. Yeah. That's a huuuuuuuge difference.

With RISP regardless of outs, A-Rod is hitting only .261, a mammoth seven hits away from .300. Yeah. Seven whole hits. A truly gigantic difference.

With runners on base, regardless of scoring position, A-Rod is hitting a pedestrian .274, still seven hits away from .300.

Notice a pattern? Not only is the difference about as far from significant as possible, but as we increase the sample size, A-Rod's numbers become closer and closer to his overall numbers.

Do you want to know something else perhaps even more telling? A-Rod's OBP with RISP and two out is .424. That's higher than his overall OBP. His OBP with RISP regardless of outs is .398, still higher than his overall OBP. Perhaps no one wants to give A-Rod a pitch to hit when he can really do damage. Note that again as we increase the sample size, his OBP retreats to his normal level. This regression to the mean is the hallmark of random statistical noise.

Does this look like the profile of a choker or someone who had "it" last year and lost "it" this year? Or does it look like random statistical variation perhaps coupled with some very cautious pitchers?

Of all the Yankees' problems this year, A-Rod's performance is probably the very bottom of the list (OK, it's more of a problem than Mike Mussina). There is no reason, there is no reason, why we should ascribe A-Rod's "struggles" to anything other than the capricious whims of Lady Luck.

I'm sorry if that makes a worse story.


rklllama said...

Thank you John for stating everything so clearly. A-Rod for the win.

(Moose for twenty!)

D.Cous. said...

Are you channeling Walter Sobcheck at the end there?

John Lynch said...

Re: Walter

Not intentionally, though I can say that that was the frame of mind I was in.