Al Yellon
Editor
It's Gotta Be The Hair - May 15
Grant Brisbee
Editor
Tarasco Sauce - May 16
Jeff Sullivan
Editor
Mission Accomplished - May 16
Rob Neyer
National Baseball Editor
Baseball is Unpredictable - May 12
Dan Moore
Contributor
Joey Votto And The Twilight Of The Giant First Basemen - Apr 14
Jason Brannon
Hot Corner Scribe
Len Kasper Has Been Over Into The Future... - May 15
Jim Baker
Contributor
eBay Item Of The Day - May 15
Marc Normandin
Contributor
The Royals And Pitcher Development - May 15
Wendy Thurm
Contributor
On Being Young And Brash In The Social Media Age - May 12
SB Nation World Series 2011: Cardinals vs. Rangers
Oct 23, 2011 - Hey, did you hear the one about the guy who came in with terrible statistics and hit three home runs?
Yeah, huh? That actually happened.
Two facts to roll around inside your noodle:
1. Before Saturday Night's Game 3, Albert Pujols had played in 11 World Series games and batted .222 with exactly one home run;
2. In Saturday night's Game 3, Albert Pujols hit three home runs.
I suppose it's possible that nobody ever accused Albert Pujols of choking in the World Series. It's possible, because Pujols had played so brilliantly in Division and League Championship Series. Still, knowing how these things work, I can't hep suspecting that "some" at least mentioned Pujols's World Series statistics, and at least implied that those statistics said something more than, "Hey guys! Anything can happen in small sample sizes!"
I got to thinking about this not because of Pujols's three home runs, but because of Michael Young's one home run, which for a while seemed like it might actually matter.
Shortly after Young homered against Kyle Lohse in the fourth inning, Buster Olney tweeted thusly:
Waiting for someone to write again that Michael Young's three homers in 30 ABs against Lohse before tonight wasn't predictive in any way.
Gosh, you would almost think that Buster was trying to start a Twitter War, because of course referencing small sample sizes is like catnip for a certain youngish segment of the Twitterverse.
Oh, take our own Jeff Sullivan for example. He fired off three tweets in quick order:
One of those home runs was in 2004 and the other two were in 2005 - When else do we care about what happened six or seven years ago? What about the eight strikeouts? - There might be something there, but the overwhelming likelihood is that it's noise, no?
One thing about Buster Olney ... He doesn't consider himself above a Twitter War. Immediately, Olney responded to Jeff:
He's got four HRs in 32 ABs against Lohse, and you don't think that's at least a clue he sees the ball well against him?
You might guess that the youngish segment does not think that's at least a clue, at all. That segment weighed in, and Olney weighed back.
But then there's our friend Carson Cistulli, a gentle soul who doesn't use curse words and really, down deep, just wants everyone to get along.
@cistulli
Carson Cistulli Can someone w/ a copy of The Book please answer this?// @Buster_ESPN How many at-bats are needed before it's not too small of a sample size?
FavoriteRetweet
Why, yes. Someone can. From The Book:
You see, we're not saying that it doesn't matter which pitcher is facing which pitcher. It most certainly matters. Every person is different, and there's no reason to think that two overall equally talented pitchers, but talented for different reasons, will necessarily have the same success level against the same hitter. However, you can't tell by looking at the numbers from twenty-five or sixty PA. There is simply too much noise masking the truth under those numbers... For you to say that a certain hitter owns a certain pitcher, you have to go beyond the numbers. You have to look at the very specific traits of the players...
Among "the noise" is this profound truth: Young's three home runs against Lohse could almost as easily have been two home runs. Or one, or none. Think about it. Couldn't Lohse almost as easily have thrown three slightly better pitches? Couldn't Young have taken three just slightly worse swings?
We cannot know that Young's previous 30 at-bats against Lohse weren't instructive. We can know that the statistics, in isolation, tell us virtually nothing that will help us predict what will happen in the next at-bat. Just as Albert Pujols's World Series struggles before Game 3 told us virtually nothing about what would happen in Game 3.
Is this an apples-to-apples comparison?
No, not really. Call it instead tangerines-to-oranges. In both cases, the numbers just aren't big enough to drown out all the truth-masking noise. We all would be just slightly better off if everyone stopped reporting these numbers. Yes, they're now easy to find (thanks a lot, B-R.com) and easy to tweet (thanks a lot, Twitter). And depending on your particular bent, they might well be sort of fun. But none of that makes them meaningful.
7 comments
Rob Neyer:
How Did Albert Pujols Become A Great World Series Hitter?
7
Oct 23
16
Oct 20
5
Oct 20
2
Oct 19
2
Oct 19
Mickey Hatcher, Angels Hitting Coach, Relieved Of Duties
Vance Worley Has 'Soreness', Will Miss Wednesday Start
Justin Morneau Will Be Activated From DL Wednesday
Manny Ramirez Rehab Assignment Will Start Saturday In Albuquerque
Jon Jay Heading To Disabled List; Shane Robinson Recalled
Troy Tulowitzki Day-To-Day With Deep Leg Bruise
Matt Kemp Goes On 15-Day DL With Sore Hamstring
Jeff Niemann Likely Missing 'A Few Months' With Leg Injury
VIDEO: Bryce Harper's First Career Home Run
Torii Hunter Placed On Restricted List Following Son's Arrest
More News »
Comments
Blink
I’m with Malcolm Gladwell on this one. We “know” we’re seeing something long before we have enough facts to prove it to ourselves. By twenty AB’s I feel we have enough info to tell wether or not you walk or pitch around a certain guy.
by David Connor on Oct 23, 2011 2:32 PM EDT reply actions
I'd say 20 at-bats in a year or two might be somewhat predictive
but when you’re going back to the mid-00’s, well that dog’s just not gonna bark.
Time to update this: 2009 Did Not Happen. Neither did 2010. Or 2011. Sports suck.
by cjmulrain on Oct 23, 2011 3:16 PM EDT reply actions
It does depend on the time span that we are speaking of.
Also, The % of success matters; 15/20 is more instructive than 3/20.
by Eastwindquinn on Oct 23, 2011 4:24 PM EDT reply actions
Like Dotel vs Braun says something.
Lohse vs Young doesn’t say much.
by Jon S. on Oct 23, 2011 4:34 PM EDT up reply actions
Grrr
The title of this article was enough to get me stewing.
Kicking knowledge in the face.
by BlackOps on Oct 23, 2011 7:29 PM EDT reply actions
I say it can be as low as 8.
For example, I think 8 PAs is sufficient in the case of Jimmy Gobble v. Paul Konerko (5-6, 4 HRs, 2 BBs).
8-32, 4 HRs, 1 BB…because of the larger sample size, I feel more comfortable making an assertion on this. And it is likely not predictive.
by dejezeus on Oct 23, 2011 10:06 PM EDT reply actions
It depends how well the hitter performed in those PAs.
If a hitter is 6-6 with 6 HRs against a pitcher, then that’s probably enough.
If he’s 0-15 with 15 Ks, that’s probably enough.
It’s not how many PAs, it’s how many SDs.
What’s the standard deviation of performance, given the number of plate appearances there were?
I’d be quite happy with going down to the 90% level of certainty that the variation isn’t just chance (rather than the standard 95%), but apply that taking into account the number of pitchers that batter has faced, or else you’ll just prove that 10% of the time, the hitter does better than the 90% level.
by po8crg on Oct 24, 2011 9:52 AM EDT reply actions
Comments For This Post Are Closed