Is Paul DePodesta.

Awful comparison.

and is on pace to do so now. You have to laugh about mentioning BABIP in media discussion?? Dude, get real. BABIP and WAR are the new batting average and RBIs. They are universally available stats that should be recognised when ignorant media types blather about average and are self explanatory in the case of BABIP and others.

Yes, the fact that Francoeur has shown a bit more pop than Church and doesn’t particularly favor the other two “true outcomes” means that there will be a sharper curve in the correlation between BABIP and productivity. This works both ways though. Frenchy’s BB% is just so brutally awful that unless he has a high BABIP and/or sustains the type of power he’s flashed capability of, then his utility is going to drop off sharply. Sure his BABIP was below expectation in 2008. But the point of a stat like BABIP is to examine deviations, its supposed to do that sometimes, if it didn’t, you’d have a very unique player or a very strange statistical improbability. His BABIP in 2005, his “good” year, was .341, when his LD% was 19.1, which would make his eBABIP by this method closer to .311 (if you factor for his highish FB%, I actually get it closer to .306, but that’s nit picking). So this evidence seems to support the initial hypothesis, that when his BABIP is extra high, he will be productive. But isn’t that just kind of a circular argument, in that its like saying “when there is a favorable [uncontrollable] outcome, the outcome [of the larger set which includes the controllable and uncontrollable] will be more favorable.”? In the largest set, that is, his overall career, where the law of large numbers becomes more relevant, even ignoring recent trends, his career wOBA is still only .313, so color me unimpressed that his wOBAs have fluctuated with strong correlation to his BABIPs.

Combine that with the fact that his HR/FB has dropped in every season he’s spent in the majors, and you have a pretty glaring flaw in process here. Sure, he’s developed himself, in exactly the way he’s always claimed he would, he’s hitting more pitches. His contact rate has improved substantially since his rookie and sophomore seasons. But that’s not helpful if his HR/FB is going down the tubes and he’s never drawing any walks. It brings everything back to his BABIPs, since the “true outcomes” are becoming devalued according to his career trends (and also according his own claims about the relationship between OBP, the scoreboard, and statistical evaluation).

Sure, its nice that he takes extra bases, that’s something the Mets could certainly use. Those numbers might indicate that the crazy small number that represents the likeliness Frenchy will miss third base is a teensy bit smaller than the crazy small number that represented likelihood that Ryan Church would have done that the time he did. It also means he’ll earn himself a few extra offensive runs every year on the basepaths. But all that can ever be is gravy to whatever production he gives you in more tactically meaningful situations, such as when he’s in the batters box.

As for PECOTA, you are right, Nate Silver’s system does seem to favor Francoeur there. But there is one number in PECOTA that can tend to get overlooked, but is really the basis for the entire PECOTA process: Similarity Index.

Similarity Index is a composite of the similarity scores of all of a player’s comparables. Similarity index is a gauge of the player’s historical uniqueness; a player with a score of 50 or higher has a very common typology, while a player with a score of 20 or lower is historically unusual.

Jeff Francouer: 47
Ryan Church: 65

What this means is that Church’s typology is very common. Combine that with a high Beta score, 1.21, and its pretty safe to assume that there is a high level of attrition and variable performance among players that fit this typology in their early 30s.

Francouer may not be historically uncommon, but he falls short of the “very common typology” tag. In fact, Church’s 20th most comparable player, Wally Westlake, generated a higher similarity score (48) than Francoeur’s top comp, Charlie Spikes (42). Ryan Church’s top comp, Bill Howerton, generated a similarity score of 67. So it seems even PECOTA recognizes that Francoeur is something of a statistical outlier, if perhaps only a mild one.

In essence, this doesn’t really mean anything of particular use, its just my way of saying that Francoeur’s PECOTA should be taken with an even larger grain of salt than most players’, including Ryan Church’s.

I believe Francouer’s lack of plate discipline will always make him susceptible to BABIP variance. His career Isolated plate discipline (IsoD = OBP-AVG) is .042. He’ll always struggle to have a decent OBP as long as he can’t draw walks. That might be acceptable for a slugger (like Wily Mo Pena 5 years ago) but it’s disappointing for a line drive gap hitter.

Compare that to someone who is great at taking walks, David Wright, whose career IsoD is .080. Wright will always get on base even in the unluckiest of hitting streaks. Similarly, Ryan Church’s is a respectable .072.

trading prospects for veterans is a good way to look good in the short run, since if the prospects flame out it looks like the other guy got fleeced. Even if they turn out to be average, it’s a lot easier to focus on a star like Santana and neglect to recognize the value of average performance for free in 3 positions (add the salary of a guy like Santana, and the winner ceases to be obvious). Only if the star flames out and the prospects emerge does the GM look like a dupe.

In the long run, of course, the value of having depth and flexibility with payroll begins to take precedence over having a few stars that restrict your financial options and put your team at risk of total disaster if one or more gets hurt. This is why guys like Colletti, Bavasi, Phillips, and Minaya might look good in the short-run, but unless they get lucky, they eventually undermine their organizations.

Duquette is an exception because his move was truly special: to trade your star pitching prospect for an actively bad pitcher.

Its actually one of the simplest and most necessary “new” stats to come around in a while. Its just that we can try to be more precise, especially with the recent advent of Hit F/X. The data on Hit F/X is just beginning to accrue so it will be a while, but until then, the information BABIP studies have generated is infinitely useful, there’s just too much statistical “noise” in historical BABIP trends to deny that there is a very uncontrollable element to it, both for pitchers and hitters, at least given the data available to us. As you’d figure, BABIP outliers are more sustainable for hitters (i.e. David Wright), and for pitchers there is a greater correlation with quality of defense. But in both cases the level of variance indicates information that is either:

A. Pure random chance
B. A function of data that’s just not available. And obviously, you can’t analyze data that’s not available. Maybe, probably, and hopefully Hit F/X will help solve this issue, but again, its so new and fresh that there’s just not nearly enough data out there yet to apply it. So, we works with what we gots.