Are Billy Beane Believers still expecting 97 wins this year?

What do the 100+ angry baseball fans who have posted livid responses to my earlier postings about Billy Beane have to say about the new data that has been assembled since I made my first claims? The A’s record is now 14-20. The chances of a team that wins 60 percent of their games going 14-20 in the first 34 games is about 1 in 50. Even a team that is expected to win 50 percent of their games (as the market predicted the A’s should at the beginning of the season) will start this poorly only 20% of the time.

So far it looks like the market knows more than the “experts” who respond to my posts. Which is, of course, what we should all expect.

Maestro

You do realize that we're not dealing with a random sample, right? In fact, a cursory look at the A's schedule indicates they've generally played better than average opponents and a few more road games than home ones. Wow, that research took me all of 2 minutes to dig up.

vickie

hey you have a sn mine is sportygirl4849

Thomas

But if you wanted a general manager of a team that started 14-20, you could do a lot worse than Billy Beane. Right?

Anonymous

Steve Carr, I am using statistics exactly the way statistics are designed to be used. The binomial distribution says that the chances of a team that has a baseline win probability of 60% going 14-20 over a 34 game period is .02. What possible gripe can you have against this most basic use of statistics?Well, one possible gripe is that you picked this endpoint and made this post because the A's have been so surprisingly bad over that span. Your calculation might point to how surpising it is, more than how bad the A's are. Afterall, many really unlikely things are bound to happen in a baseball season. If you were to find them all and decide that they are all meaningful you'd be led way off track.Jared

vickie

visit my site and comet it its anthing cool

paul

Martin: I took your suggestion and only looked at the number of wins beyond 60 per dollar of payroll (had to exclude one team who had only 58 expected wins). Oakland was #15 in that analysis.

Anonymous

JC Bradburry, a professor of economics at University of the South, Sewanee has recently done a regression analysis trying to link performance to other bariables that create performance such as batted ball types. He says his formula has been able to explain a little over 80% of the differences in observed OPS. Here are the links for the AL and NL.http://www.hardballtimes.com/main/stats2005/alpropshttp://www.hardballtimes.com/main/stats2005/nlprops Check out how Oakland has faired compared to what would have been expected. Every single player has an OPS below what would be expected. There obvously could be an explanation here. The formula might not be any good going forward. The data might not be recorded that well or their might be some bias in recording Oakland's performance (although I seriously doubt the last one). They also might have just been really unlucky.There's also some players palying well below their estab;ished track records, even in underlying performance. I'm not sure you can blame the GM when Eric Chavez has been performing like on e the worst players in the major leagues.Dr. Levitt,It just seems your thinking on this subject is similar to that of your critics on the abortio-crime hypothesis. They point to the surface and say "but murder sky-rocketed!!" They don't look any deeper becasue they don't want to. You've looked at the A's win/loss record and said "but they're losing!!" It seems you also do not want to look any deeper.Also, isn't 97 games a bit of a strawman?

Read more...

Anonymous

BTW.JC Bradburry is the author of www.sabernomics.com. I think his methodology can be found there.

Very Disappointed

I'm shocked by how ignorant Mr. Levitt appears in his posts about baseball and Billy Beane, especially since I'm enormously intrigued by his groundbreaking work in the social sciences. It makes me worry that he's off track in areas that I know a lot less about. Because he's clearly not getting the point of Moneyball or digging very far into the numbers this year as to why the A's aren't performing very well.

Black and Blue Notes

Nothing you have said diminishes this truth: The A's made the postseason several years in a row, with a payroll at the lower end of the MLB rankings. This makes them the best performance-value in the league. Beane is the architectof this "performance-value" and must be given credit for the outcome. This year was an intentional step back. He acknowledged this would be a tought year before the season started.Given the economics of the game (the recent contracts given to mid-level performers), Beane rationalized that he would eventually be unable to maintainhis assets in their current form. Mulder,Hudson,Zito,Harden,etc. They would command at least 10 million each. He made a pre-emptive strike and traded them away before theri value diminished - knowing he gave away the best chance to win immediately.Instead, he chose to secure the A's future by acquiring excellent prospects. These prospects need a year or two to develop. Then we will see if this season is successful or not.It is not to be measured by the wins/losses this year. It is inthe development of the prospects that he traded away 2 potential hall of famers to acquire. You can not inject this season into the discussion, because it is a season for sowing - not reaping. If he's right with these prospect,he will have secured another 6-7year run of success, and also cemented the team's staying in Oakland. The essence of Moneyball. Think outside the box to attain your objectives.

Read more...

Matt

True, the A's have started out dismally a few times and won 95+ games, and of course the players on a team change over the season (due to trades, etc.), so they do have a chance.That said, to win 95 game, at this writing they will needs to go 80-43, i.e. they would basically have to transform into the bast team in baseball. Their current record can't be explained by fluke either- their Runs scored/Runs allowed is 145-205 (incidently, the worst run-producing offense in the AL), which by the pythagorean theorem of baseball predicts a .333 winning percentage, the worst in baseball. So to have a good chance at making the playoffs, the A's will have to go from the worst to the best. I like Beane as much as anyone, but that ain't happening.(For the non-nerds out there, the "pythagorean theorem of baseball" is that a team's winning percentage can be predicted by the formula RS^2/(RS^2 + RA^2). In a given year, usually 26-28 of the 32 teams in baseball will be within 3 wins of their "expected" total. A comparison between expected and acual records can be found here: http://sports.espn.go.com/mlb/stats/rpi)

Read more...

Anonymous

The iid assumption makes such probabilities easy to calculate, but a even a cursory look into baseball history will show some degree of correlation in consecutive game outcomes. Consider this year's Yankees. Assuming iid and that the Yankees are a .500 team (which is almost exactly what they are to this point), what's the chance that they'd start 11-19? A small 5.09%. Yet, that's exactly what happened. What's the chance that they'd follow with a 10-game win streak? A miniscule 0.098%. Yet, that's exactly what happened. Granted, I'm carefully selecting 1) the team, and 2) the time span, to make the numbers sound as impressive as possible, but, I think it's fair to say that the iid assumption is a poor one here.A much better way to judge the chance of Oakland winning 60% of their games after a 14-20 start would be to look at baseball history. How many teams with a 14-20 start went on to win 60% of their games?Jasonhttp://people.csail.mit.edu/~jrennie/baseball/

Read more...

Jonathan Hall

It's amazing that the word "injury" doesn't show up once on this entire blog. Let's glance at the A's players who are injured: Rich Harden, whom I would say is obviously the best pitcher in the AL. Kiko Calero, the best setup man. Octavio Dotel, one of the best closers. Bobby Crosby, their regular shortstop. Nick Swisher, a monster prospect slated to start in the outfield. Juan Cruz has been injured also, along with monster prospect Justin Duchsherer. Am I the only baseball fan here who actually watches what's going on on the field? It's not at all unreasonable to say that what the A's have undergone is extremely unusual, and therefore, they probably won't win 100 games. They probably will, however, play .550 or better baseball once their players are back - what more can we ask of Billy Beane?

Anonymous

Joe Sheehan wrote an interesting essay on Baseball Prospectus earlier this week comparing the A's record at this point this year to their record at this point in 2001 - basically identical. They went on a tear in the second half and were the wild card. They might or might not get there this year, and their immediate success has been against sad-sack teams (Giants, Mariners), but they are not out of it yet.

jmac

Well, it's August 7th and there should be no more arguments given what's happened the last 2 months with the A's.I wanted to point out that if yo uput the anonymous post about iid together with Mr. Hall's injury comments, you will conclude that the problem with using the binomial distribution in this case is that it wasn't a team with a 60% winning percentage which started the season 14 and 24, it was a team decimated by injuries, more like a team with a 40% win pct. Anyhow, the REAL team began playing together as a unit at the end of May when Street, Duchsherer, Johnson, and Crosby showed up. Plus, new catcher Jason Kendall finally got comfortable with AL pitchers and opposing hitters tendencies and now explpoits the situation much more effectively. Becasue of injuries, and trades, and teams with deep pockets willing to buy players to improve themselves mid-season, it is much harder to predict baseball than other sports.

Read more...

Maestro

Haven't the A's started out poorly most of the past 5 years? One year, I think '01, they were truly awful for a month or two, and still won over 100 games. This doesn't prove anything, but neither does their record so far this year. For the record, I expected them to take a step or two back this year, and I still expect 81-85 wins.

Anonymous

So 34 games prove that you are right? Unbelievable.

Martin

Wow, that's an astonishingly rude and obnoxious way to engage you readership.

I just reread the thread you're citing and it's full of heated, honest dissent. I didn't see a lot of lividity there.

Being one of the top 3 or 4 teams over a 5-year stretch is extremely difficult if you don't have the resources of the Yankees. The A's have been a top team for 5 years or so, and it's quite possible their glory days are fading. This proves very little. Beane's method is sound, but the league's caught up with him a little bit, and also he may be in a down cycle that occurs when the free agents leave. If he's half as good as his reputation, the A's will be very good in about two, three years. Or then again, maybe not.

The funny thing is, if you were to peg Beane on something legitimate, a lot of intelligent baseball fans, even your commenters, would be eager to agree -- Beane is overhyped, you have a point. But you are clearly not all that conversant with baseball (not as much as some of your commenters, who understand Lewis better than you do), or else your method isn't very strong. You seem almost desperate to pin Beane on some false straw-man version of Beane. Beane lost two of his top pitchers, and his third ace is maybe not as good as he appeared three years ago. This makes Beane's prowess at running a baseball team a mirage? Hardly.

Have some respect for your audience and come with something really interesting about Beane. This amateurish nonsense won't cut it.

You're treating your readers almost as well as you treated that restaurant a few posts ago.

Read more...

Anonymous

Hey people, don't confuse the coauthors. The restaurant piece was by Dubner, not Levitt.

Steve Sailer

Isn't it fascinating how much controversy Dr. Levitt's opinion's on Billy Beane have generated, while most commentators have been happy to take his seeminly more important abortion-cut-crime theory on faith, even though the latter fails dismally most of the obvious tests of plausibility (e.g., the first generation born after legalization of abortion went on the worst murder youth violence spree in American history)?