Archive for May, 2007

May 18

TV: CNBC’s Street Signs, 2:50 pm (EDT)

The Baseball Economist in Village Voice

Allen Barra uses my estimate of Roger Clemens’s worth to rationalize his signing by the Yankees.

Professor J.C. Bradbury of Kennesaw State University in Georgia tells the Voice that Roger Clemens isn’t worth it. In his new book, The Baseball Economist, Bradbury calculates the value of a win in terms of revenue created. He then credits each player for his contributions to those wins in monetary terms—sort of like Bill James’s theory of “Win Shares” converted to dollar signs. It’s complicated, but then so are both baseball and economics. The bottom line for Bradbury, economically speaking, is that Clemens’s worth is about $11 to $12 million, or about a third less than the Yankees will be paying him to start perhaps 20 games in what’s left of the season. (Actually, the cost to the Yankees is greater than Clemens’s reported $18.6 million salary. They’ll also be kicking in more than $7 million to Major League Baseball’s luxury tax fund.)

How is this rational? Read on.

Happy 40th John Smoltz

Jenny of Velcro Vernacular (and a new contributor to Talking Chop) describes John Smoltz’s 40th birthday presents from his teammates. Includes funny pictures and commentary.

Is Jeff Francoeur For Real?

The season has started well for Jeff Francoeur, who is batting .304/.358/.493. This is much better than his 2006 line of .260/.293/.449. While unexpected batting averages are quick to revert to the mean, the real change in Francoeur’s game has been his propensity to walk. This season he has walked 11 times for a walk rate of 6.79%—not good, but much better than his 3.35% of 2005. He’s also added a foot-tap to his swing to help with timing. Could he finally be getting it?

My perception from watching him at the plate is that he has improved as a hitter, or at least he has demonstrated that he is capable of improving. In the past, I could almost always predict where to pitch Francoeur to make an out—chasing balls up-and-in and low-and-away would normally do the trick. This year, I have noticed him laying off of these pitches to get into good counts so that he forces the pitcher to throw him a strike.

Sure enough, the data backs up my perception. He’s swinging at the first pitch in 44% of his plate appearances, compared to 52% last year. 14% of his trips to the plate go to a 2-0 count, compared to 11% in 2006. He’s also swinging at 56% of first pitches, compared to 61% the year prior. And he’s seeing 3.54 pitches per plate appearance compared to 3.32 in 2006. Granted, he could be doing better, but it is encouraging improvement. (This sweet data is now available at Baseball-Reference by clicking on Pitch Data Summary.) He deserves more than a pat on the back, because he was forced to learn this at the major league level. Let’s hope these changes are the beginning of something more.

Now for the bad news. First, the batting average: he’s not going to bat .300 for the year. At his best, I see him as a .275 hitter. His career numbers (minor and major league) are below this, and his current batting average on balls in play is .351—about 50-points above his career average. Next, his power has not changed: his isolated-power of .189 is identical to his 2006 performance. When you take this into account, and assuming he keeps his current walk rate, he’s about an .800 OPS player, which would translate to an OPS+ of 108. Again, this is an improvement from his OPS+ of 89 in 2006. It’s respectable, but not All-Star caliber play.

The good news is that Francoeur is 23, so we should expect the improvement to continue. And who knows, maybe the improvement is more than what we see here. He’s still young and learning. His personality is such that I think there is a reason to be optimistic. I do wish that the Braves hadn’t rushed Francoeur to the big leagues. You can work on things in the minors without consequences; there is no time for experimenting in the big leagues. Even if he never blossoms to meet the unreasonable expectations placed upon him, he will be a ballplayer that any major league team would be happy to have on its roster.

Is he for real? It depends of your definition of “for real.”

My Solution to Doping in Sports

If you haven’t been paying attention to the Floyd Landis doping case, you’re missing quite a spectacle. The battle between Landis and the World (U.S.) Anti-Doping Agencies (WADA and USADA) is full of…well, comedy. It would be funny to laugh at it if it didn’t involve the reputation of a man who may be innocent. After winning the Tour de France last year, a drug test following Landis’s amazing Stage 17 win revealed a high testosterone to epitestosterone ratio. Supposedly, further tests have revealed the presence of synthetic testosterone.

But all of this is debatable according to Landis, and it’s hard not to be sympathetic with the cyclist. (Read more about the case at Trust But Verify). The case has been bungled throughout the appeals process, and even the original samples appear to have been mis-handled. Any objectivity the lab testing the samples had was lost long ago, because it has every reason to defend its initial finding. Furthermore, WADA and USADA officials have a lot at stake as well. I won’t detail all of the happenings in the case, but the latest mishaps involves an arbitrator rebuking his fellow arbitrators (you know, those guys who are supposed to objectively decide the case) for leaving him out of a decision that went against Landis and Landis claiming that prosecutors offered a deal to help them get Lance Armstrong. The head of the WADA Richard Pound once commented that with Landis’s identified testosterone levels “you’d think he’d be violating every virgin within 100 miles.” And you thought leaks were bad in the BALCO case?

It’s been nearly a year since the incident, and the circus continues. This case contains numerous examples for economists to use in the classroom: the incentives for competitors to cheat (prisoners dilemma), the incentives of bureaucrats to justify their existence rather than seek the optimal (public choice), and the consequences of setting thresholds for hypothesis testing (the opportunity cost hypothesis rejection cutoffs). There is no way to develop a perfect performance-enhancing (PED) drug testing procedure in any sport. Athletes have every incentive to cheat, and the incentives are not right for monitors. It doesn’t mean we shouldn’t try, but I don’t think the current system is very effective.

I propose that we keep testing, but change the punishment. As a fan, I want to revel in the excitement of the sport in the present rather than wait for results like a high school senior anticipating the arrival of his SAT scores. The cost of using steroids is borne entirely by the participants. (Please, no sob stories about role models.) I don’t face any negative health consequences from PEDs, but the athletes I cheer for do. They ought to be the ones concerned that to succeed they must dope just to keep up with the dopers.

My solution: fine dopers 100% of their winnings and endorsement deals and redistribute the money to athletes who test clean. If you want to dope, go right ahead. And if it brings you glory, you get it. But, the additional fans you bring to the sport and take away from clean athletes, you have to give it all away to the ones you harmed. If you want to play clean, you might lose an edge on the competition, but you are rewarded for staying clean. The results of all of the tests will remain confidential, and the public has no fear that anyone will be stripped of a title. Dopers who win do get fame, but their fortunes will be limited. Garnishing endorsement deals will be difficult to enforce once athletes retire, but the financial payoff from cheating will be much lower than it is now. There is still the problem of labs, but I think they’ll behave better once they are removed from public scrutiny. If a finding is reversed, no one will know they got it wrong in the first place.

How would this work in baseball? This is something I discuss in my book. Basically, I think the players union needs to take charge of drug testing to alleviate the fear that owners will mis-use medical information contained in drug tests (e.g., THC levels), and punish players with fines that will be redistributed to players who test clean.

I’m tired of discussing steroids in sports. I think this solution will allow fans to focus more on the game, knowing that clean athletes are being compensated for their behavior.

Don’t Worry About Andruw Jones

Once again, Andruw Jones has gotten off to a slow start (.241/.379/.457) by his standards. Now, I don’t think it’s necessarily a pattern that means something. If fact, his career March/April line (.266/.356/.501) is nearly identical to his overall line (.266/.345/.504) (B-R PI splits). His batting average and power are down a bit, but it’s down league-wide. His OPS+ 126, which isn’t much off what he’s done during the past two seasons: 133 (2005) and 129 (2006).

So, I wonder why Joe Simpson won’t quit riding the guy for not “going the other way” during Braves broadcasts. Andruw responded in the press yesterday.

“I’m not a right-field-ball hitter,” Jones said. “I’m a pull hitter. That’s the way it is.”

The Braves haven’t hit a home run in six games, but Jones said he wasn’t worried about that or his own sluggish start.

“This road trip coming up, we might hit 50 homers,” he said. Of his skid, he said, “A slump’s a slump. It’s a long season. I think I had an 0-for-40 once. I’m not worried. At the end of the year, I’m going be where I want to be.”

As I’ve shown before, Andruw succeeds when pulls the ball (also see here), and his slumps are the product of poor plate discipline. Of course, last night and the night before Andruw had hits to “right” field—meaning to the right of second base, but they were more towards center—and Simpson attributed these hits to his working with Terry Pendleton and his “go the other way” approach. Well, one thing we know is that Andruw ignores Terry Pendleton’s advice on this. He said as much in the quote referenced above, and he’s said so before. He hasn’t changed his approach in terms of where he hits the ball, and after examining those recent hits on the Tivo I believe he was trying to pull the ball but was just late in both cases. But, that’s not important.

The reason Andruw Jones is struggling has nothing to do with him turning outside pitches into weak grounders to the left side. Andruw is striking out more: 23% this year versus 19% last year. Might he be striking out because he’s trying to pull pitches and ends up missing and getting in bad counts? That is possible, but Andruw seems to have improved his plate discipline. He’s walking more—16.55% in 2007 versus 12.25% in 2006—and he’s seeing more pitches–4.1 compared to 3.9. In fact, while Andruw is about 25 points below his career batting average his on-base percentage is about 35 points above his career mean. His isolated power is down, but there is more to the story here. First, his iso-power of .216 isn’t bad—he’s third on the team behind Chipper Jones (.341, wow!) and Kelly Johnson (.222). Second, he’s on a pace to hit nearly 50 doubles, well over his career best of 36 (2000). I expect some of those doubles will turn into home runs will turn into home runs, especially as summer temperatures arrive.

When I look at the big picture, I am optimistic about Jones. He started the year with the plan, telling the AJC that he intended to up his walk total quite a bit this year. So far, he’s on a pace to walk 117 times—well above his career hight of 83 (2002). I give him a lot of credit for staying the course and not paying attention something that is probably the result of a random run in the data. His plate discipline will pay off, and I expect that Andruw will finish the season with an OPS over .900.

More on Clemens

Tim Lemke writes in The Washington Times on the value of Roger Clemens: The $18,500,000 Man. He includes my take, which I discuss here.

J.C. Bradbury, economist and professor of sport science at Kennesaw State University and author of “The Baseball Economist: The Real Game Exposed,” said four months of Roger Clemens should be worth about $12 million to a team. But he also acknowledged that Clemens may offer the Yankees things that his statistical analysis doesn’t take into account, such as the value of veteran leadership or the unique characteristics of the New York market.
Analysts who have studied the deal also said determining Clemens’ value is tricky because there are so few pitchers to draw comparisons with. Nolan Ryan, a power pitcher like Clemens, went 5-9 with a 3.72 ERA in 27 starts for the Rangers in 1992 at age 45cq-stats. A season later, Ryan went 5-5 in 13 starts with a 4.88 ERA. Phil Niekro and Charlie Hough has success past age 45 but threw the knuckleball.
“On the face of it, this is a lot more than what I’m stating that he should be paid,” Bradbury said. “But there’s probably nobody who has his kind of starpower who is also that good.”

Marcus Giles Returns

Marcus Giles returns to Atlanta tonight, and he seems to have shaken off his 2006 slump. His OPS of .828 is right on target with his 2004 and 2005 performances. The problem is, Marcus is once again overperforming his PrOPS (predicted OPS) of .744—80 points below his actual OPS.

PrOPS predicts OPS based on the way players hit the ball and does not focus on outcomes. As I have argued before, Marcus is someone who has been lucky throughout his career. Now, maybe Giles is an example of a player that casts doubt on PrOPS’s usefulness; however, I’m confident in it. If I owned him on my fantasy team, I would attempt to dump him now. But hey, that’s just me. Use at your own risk. 🙂

Overpaying Roger

Yesterday, the Yankees signed Roger Clemens to a deal with the Yankees for a pro-rated share of $28 million. Depending on how quickly he gets ready, he should make between $18-19 million. Based on his 2006 performance, I estimate he’ll be worth about $11.9 million if he starts 23 games at six innings a piece.

I think it is interesting that the Red Sox supposedly offered him $10 million less and to start pitching later in the season. If he pitched the time frame he pitched last year that would put his value at about $9.8 million, which is similar to the amount my system values him (see The Baseball Economist).

Does this mean the Yankees “overpaid” Clemens, as I insinuated in this post’s title? Not necessarily. There are many things that Roger might bring to the table that my system doesn’t value: star appeal, veteran leadership, unique characteristics of the Yankee market, or how much much George Steinbrenner values attention. However, I do think it’s a sign that the Yankees are hitting the panic button. I don’t think the Red Sox were serious about Roger, and I doubt Houston was willing to go this high. Maybe that’s the premium to get him to play in NYC.

Racist Referees Again

The response to the Wolfers-Price study finding evidence of racial bias in the calls of NBA referees has been big. This paper is everywhere. Much of the criticism is based off a misunderstanding of what the authors did: “But most of the players in the leagues are black” or “The tops foulers in the league are white.” The authors use statistical methods to control for these and similar problems. I won’t defend the study, but even though there are some potential holes in the paper, I’ve yet to see a truly damning criticism.

The main legitimate criticism of the study is that the authors use game-level data (looking at total fouls within a game) as opposed to more granular foul-by-foul data. From the box scores, the authors cannot determine which referee called which foul. This is true, but it doesn’t condemn the study. Just because a study is imperfect doesn’t mean it’s wrong or that it should be ignored. Heck, the best study would involve implanting neural sensors on referee brains to interpret subconscious feelings of racial prejudice. The denial from imperfection fallacy is one of the most annoying critiques that researchers have to put up with.

The granular data is not available to the public, so we have to choose to live in a world where he accept to deny the inference of this study…except that the data could be made available by the NBA. The main argument coming from the NBA is that they have done a study with granular data, and that they do not find the same result.

“The biggest flaw is it’s wrong,” Joel Litvin, NBA president of league and basketball operations, said. “They didn’t have sufficient data to do the study they purported to do. They were trying to prove that individual referees, black vs. white, white vs. black, showed bias. But you can’t look at a box score and know which referee called which foul.”

But this is not the point. Wolfers admits that the data the NBA has is superior.

“Of course it would be better if the data recorded who blew the whistle,” he said. “Even in the absence of that, and this is not an interpretation, it’s a fact, when an individual player plays, he has fewer fouls when the majority of the referees are white than the same player on a night when the majority of the referees are black.”

The problem is that the NBA won’t share its data. And no wonder, the league is in a great position. The suits can claim the high-ground that their study is superior, yet not bear the burden of exposing the study to further scrutiny. I once had a researcher do this to me. I’ve been waiting two years for this person to send me the study that “refutes” my findings. Needless to say, it’s infuriating enough just to have the information to be withheld. But the frustration increases when people buy into this trick.

David Stern has not handled this properly. He has every right to defend his league, and his agitation is understandable. However, Wolfers and Smith aren’t out to paint the NBA as a bunch of racists. They are just using sports data to examine theories of human behavior. What Stern should do is go to the president of the American Economics Association (Thomas Sargent) to select a group of independent qualified researchers. Supply them with the NBA’s data with the stipulation that the economists not reveal anything more than their findings. This removes the privacy concerns. After all, the NBA was willing to share its data with the private firm it hired to do the study. The data stays secret and we get an independent evaluation of the conflict.

Right now we have one study that may or may not have problems that we can see, and one that we cannot. In the end, the paper will go through the standard refereeing process, and it will almost certainly be published. Following referee suggestions the authors may even determine that there is no evidence of racism. Unfortunately, the refereeing process is long, and it will be 1-3 years before this paper is published. And even then, scholars may still disagree over the findings. If I have to defer to a study that exists or one that doesn’t, it’s an easy decision. I think it is in the NBA’s best interest to investigate this further, because the issue isn’t going away.