Just to be clear: As our regular second-baseman is injured, the Giants are planning to start Emmanuel Burriss as our interim regular second baseman. And have released Mike Fontenot.
Now, stat-heads will have looked at those two pages, noticed that there isn’t anything that Emmanuel Burriss does better than Mike Fontenot, and concluded that, wait for it, perhaps Jeff or Christopher would like to chime in here, Manny Burriss is swinging the bat well this Spring.
It’s worth, I think, taking a moment to look at the Burriss/Fontenot choice in order to talk about the way we humans make decisions. In addition to the common fallacies that stat-heads complain about, there’s a more fundamental thing that stat-heads have been driving at for a decade now, that this choice brings into high relief.
The most persuasive fantasy here is the Small Sample Size illusion, and in particular the version that states that the most recent sample is the best one for predicting future performance. While at-bats are not entirely independent events, we have plenty of baseball history to show that a good month does not indicate a good year. In the rest of life, we also have a tendency to think that a streak of good performance (or bad performance) will continue. If you have a good couple of months for your (f’r’ex) cholesterol numbers, or your grocery bills, or your mood in the morning, it may well be that there is a reason for the improvement, but it may also just be that the coin came up heads a bunch of times in a row.
The next fantasy is the Audacity of Hope, or in this case just clinging to hope. We (Giants fans, but particularly the Giants organization, I would think) had high hopes for Manny Burriss; we wanted him to be another success story of the farm system. He isn’t. But when he has a good Spring Training, there’s the pull of Hope, to say that he really will be that success story. Mike Fontenot, on the other hand, has already done everything we hoped for and more. Nobody expected him to be an All-Star when we picked him up. Nobody was really hoping that he would be good enough to be a regular starter at second, since we had a second baseman, even if he was a trifle brittle. We wanted him to be a replacement, which he was. But Manny Burriss still carries with him some hope. It’s not unlike the sunk costs problem, except instead of an expensive contract burden, it’s hope. This—both the sunk costs problem and clinging to hope—is obviously applicable to a bunch of real-life problems. In particular, I’m thinking of support for politicians that are clearly not going to win the primary or the general election, or even continued support for politicians that have clearly shown themselves not up to the elected task, but I support them in a primary for a higher office, because I retain such fond hopes.
I also think that the front office connected the Hope with some idea that Mr. Burriss is younger than Mr. Fontenot. Which, you know, he is. Four and a half years, in fact. If Mike Fontenot were 27 and Manny Burriss were 23, I might say, what the hell, keep the kid, you never know. But in basebally years, Manny Burriss is not a kid anymore. He is probably done improving, which is particularly nasty because he hasn’t actually done any improving for years, now. Mike Fontenot is done improving, too, by the way, but what with having been better than Manny Burriss for a couple of years already, my money would be on Mike Fontenot continuing to be better than Manny Burriss, even if neither of them are very good. This isn’t the sort of thing that’s widely applicable, except to the extent that I think we are our current cultural moment thinks of twenty-somethings as still in their youth, unready to settle to a task—and thinks of middle-age as starting at sixty.
But the big point I wanted to make is that Mike Fontenot is a Replacement Value utility infielder. And Emmanuel Burriss is not that good.
I think the big stat-head thing over the last decade or so was this idea of measuring value against potentially available replacements. In the past, it was more common to compare people to average players, or even to good players—when Batting Average was the most important thing, people compared anyone’s average to .300 (good) or to .250 (average), not to replacement level. Similarly with home runs, the question being whether your guy could hit a ton of dingers, or hit an average number, not whether it was easy to pick up somebody cheap who could hit just as many.
Mike Fontenot is a replacement level player. Any team, caught all of a sudden without a second baseman, should be able to pick up somebody who is just as good as Mike Fontenot. Actually, today any team can pick up Mike Fontenot, which goes to show. He is an excellent example in my eyes of what we’re talking about when we talk about Replacement Level. Is he good? No, he’s not good. Is he average? No, he’s not average. Well, not everyone can be average; that’s why they call it average. Is he within shouting distance of average? Yes, he probably is within shouting distance of average. Is there somebody out there being cut this week by some other team who is just about as good? Without actually looking at the transaction log, I would guess that there is. But Emmanuel Burriss is worse than replacement level. A team with a replacement level player would win about a game a year more than a team with Emmanuel Burriss.
One thing that statheads really have been pushing is that Major League teams really should not have any players on the roster that are worse than Replacement Level. Unless of course there’s some powerful reason—your terrible player is beloved by the fans, at the end of the career, and for one more season and (one always hopes) a handful of at-bats, it’s worth losing a roster space to sell more season tickets. Or there’s the argument that such-and-such a player is half a coach, already, transmitting wisdom to the young ’uns. Or—hey, he had a great Spring Training! He’s in the best shape of his life! </headdesk>
The thing is, while this insight is widely applicable, it’s actually very difficult to apply properly. Baseball statisticians argue about what is the appropriate line for replacement value, and they have an absurd amount of data. The lesson to learn is not really that you should cut loose from your workplace, classroom, community theater production, rock band, D&D campaign or household anyone who isn’t pulling their weight. The lesson to learn is that replacement value is a tricky thing. A CEO, for instance, probably doesn’t have a good handle on the replacement value of a truck driver, or even of a subcontracted transportation service. She may think she does, but she probably doesn’t. A constituent probably doesn’t really have a good handle on the replacement value of the town’s mayor. You may think you know what a replacement level house painter is, and you may be right, but you may not be. And if you feel really, really sure about it—remember that the Giants feel really, really sure about Emmanuel Burriss and Mike Fontenot.
Tolerabimus quod tolerare debemus,
-Vardibidian.
Does the concept of replacement level apply in college admissions? If so, how?
Other than in recruiting for college athletics, of course, where the application is obvious.
Chris—do you know, I was just the other day, thinking that if the entire freshman class at Harvard, or Swarthmore for that matter, were bodily assumed between the acceptance letters going out last week and the actual enrolment, and needed to be replaced with the next few hundred people on the list, that nobody would be able to tell the difference.
…and that probably someone should have told us that.
Thanks,
-V.