Baseball Musings
Baseball Musings
October 28, 2004
The Value of Information

Seth Hackbarth at Sportblog.org asks an interesting question.


Does this championship vindicate sabermetrics?

My answer is no, sabermetrics was vindicated long ago by GM's like Branch Rickey and managers like Earl Weaver. For the last 10 years, the Yankees and the Braves understood sabermetrics in putting their teams together, in deciding who to trade and who to sign. They don't always force their managers to play sabermetric ball, but they do give them sabermetrically sound teams.

If this championship shows anything, it's the value of information. I'm an employee of Baseball Info Solutions (BIS). The Red Sox are one of our major league customers. BIS supplements Theo and his crew with fine details of the game. With this information, the Red Sox know the probability of Albert Pujols hitting a ball down the third base line. They know the probability of Jeff Suppan throwing a high pitch on a 3-1 count. They know what pitches a batter will chase, and what pitches he'll lay off. And more importantly, they appear to have found a way to communicate this to their batters and pitchers.

I don't know how the Red Sox use the stats they receive. Theo Epstein has stated that the Red Sox have many sources of information. My guess is that he uses one source to confirm another. I was watching the afternoon edition of Baseball Tonight yesterday, and the analysts were discussing Manny's HR off Jeff Suppan. They showed how the Astros were swinging and missing at Jeff's low pitches, but Manny laid off those. When Manny refused to swing, Suppan came up in the strikezone, and Manny walloped the ball. This is exactly the sort of thing that a scout can notice and data can confirm.

And consider this. The Red Sox played three teams in the post season. The Angels and Cardinals are not consumers of BIS information. They were swept. The Yankees are clients of BIS, and they nearly beat the Red Sox.

So while the World Series victory wasn't needed to vindicate Sabermetrics, it has vindicated an approach to the game where minute details matter. The Red Sox have a competitive advantage in information vs. most teams in the league, and they did an excellent job of exploiting that advantage this year.

Update: Check out this link, mentioned in the comments below. It gives you a good idea of how the front office, manager and players interact with the information available to them.


Posted by David Pinto at 09:33 AM | Management | TrackBack (1)
Comments

There was actually an interesting behind the scenes article in the Philadelphia Inquirer a day or two ago.
http://ballssticksstuff.typepad.com/balls_sticks_stuff/2004/10/the_inqys_jim_s.html

Posted by: TomG at October 28, 2004 09:49 AM

Another step in the football-ization of baseball. The game will never, unlike football, get too complex for ordinary fans to understand what's going on. But the drive for more detailed information is the same.

Posted by: Crank at October 28, 2004 11:10 AM

David that's really a cool insight, thanks for passing it along. It became clear to me after game 3 of the ALCS that the Red Sox pitchers had established a game plan against the Yankees and they stuck to it religiously and didn't make mistake pitches in the zone. It was a wonderfully disciplined, and I'm guessing, data driven approach and I'm convinced that it's what enabled them to beat the Yanks.

Posted by: steve at October 28, 2004 11:28 AM

I think at least the Sox's win blew another hole at the wall that separates the most stubborn minds in baseball's old-school and sabermetrics. As to the Cardinals, I remember reading George Will's Men at Work years ago, and that La Russa seemed to be the master of information. I don't know what happened, but it appeared this year his team was under prepared and failured to adjust in the World Series. Any thoughts on La Russa?

Posted by: wilson at October 28, 2004 11:32 AM

Re: Information

Many people will say the Sox were lucky, but as I believe Branch Rickey used to say, luck is the residue of design. Theo Epstein may not have relied on BIS or Bill James 100%, but he was wise enought to hire them both and make them available to his manager and staff, and therefore he was completely prepared for the postseason.

The odds of winning 3 straight games are 1 over 2 to the 3rd, or 1/8, or .125; the odds of winning three straight in the divisional series and also sweeping the world series are 1/2 to the 7th, or 1/128, which are also the odds of winning seven straight games.

The odds of winning eight straight games are 1/2 to the 8th, or 1/256, which means it's twice as hard to win 8 straight games as to win 7 straight games.

The odds of winning 3 straight, losing 3 straight, winning 4 straight, and winning 4 more straight, would be 1/2 to the 14th, or 1/16448.

So the odds of what the Red Sox did, their path to victory, are one in 16,448. To be fair, that's the odds for a particular team to travel that path; the odds for a team to travel the path are considerably better, since either the Sox or Angels can sweep, and then either the Sox, Angels or Yankees can lose three and win four, and then any of those can either sweep or be swept by the Cards, which means you have a number of branching possibilities which can be readily calculated. I won't do it here; it's an elementary probability problem.

But still, the Red Sox beat long odds. after beating the Angels, and after losing three straight to the Yankees, the odds of just winning four straight were 1/2 to the 4th, or 1/16 (.064 or roughtly 6.4%) based on straight probability. That would be 6 times in 100 World Series based on straight chance, and of course, you have to mimize those odds, because once a team gets up three games, it's because their probability of winning is probably more than .5 per game because they're better; and shifting that number even slightly makes the resulting likelihood of winning four straight almost infinitesimal.

Now shift that to winning 8 straight; winning 4 v. the Yanks and 4 v. the Cards; that's 1/2 to the 8, or 1/256. That's a likelihood of .0033 or .33%; or less than 3 times in 1000 sets of playoffs will you see 8 games in a row won. That's in random sets of playoffs. not once in a century--once in a MILLENIUM will you see 8 in a row won in playoff play.

So this really is one for the books.

--AJK

Posted by: Art Kyriazis at October 28, 2004 07:03 PM