Outsmarting Pythagoras
4
|
The Hardball Times found this 9/17/2007 on www.hardballtimes.com [flag] |
Tags:
MLB
Comments (7)
-
-
salb918 Now doesn't that feel better?
-
-
JinAZ There's an interesting parallel here to a debate in biology. Steven Jay Gould, himself a baseball fan, wrote a paper along with Richard Lewontin in the late 70's called The Spandrals of San Marco and the Panglossian Paradigm. They argued against a trend that they saw in which evolutionary biologists seemed to only be considering adaptationist explanations for how organisms got the traits that they have. They were not, he said, considering alternative explanations involving chance events, correlated changes with other traits, developmental constraints, etc.
In a general sense, that's pretty much what you seem to be arguing -- that folks need to stop trying to explain everything as being due to some decrete and repeatable cause, and that chance events play a larger role in how things got to be the way they are than we have a tendency to think.
The problem with Gould's arguments, however, is that he often seemed hell-bent against considering any adaptationist explanation...and that's clearly not an optimal approach to trying to understand evolutionary history.
I think the same problem could surface here--most deviations from Pythagoras is likely due to chance events, but that doesn't mean that more discrete and repeatable explanations cannot also be sought. In other words, studies like Gassko's (which you correctly cited) that attempt to evaluate and quantify the extent to which other factors might be coming into play are certainly still worthwhile. It's just that we need to be cautious about overstating the importance of such factors until we have objective evidence of their effect sizes.
Anyway, thought it was worth mentioning. Always fun to blend professional life (I'm a biologist) with hobby life. :)
-j-
salb918 That's a very interesting parallel.
It's just like in chemical engineering when the famous Prandt...no, I guess there aren't any parallels to my real life. Oh well.
You're right that my objections can be taken too far. It's worth noting when deviations occur for a reason, and whether the deviations can be suitably explained by certain factors. Of course, my primary objection, though perhaps not clear, was the use of those factors to predict a continuing deviation.
-
JinAZ Reading my post now it looks like I was critiquing your article a bit. I do think we need to be cautious with stating that everything is almost certainly random, but at the same time, I also completely agree with just about everything you said in your article. In fact, this is something that has come up recently in some of my own discussions, and you did a great job of articulating the problems with those arguments. Kudos on a very nice article. -j-
salb918 I agree, then, that we agree! Your critique, as it were, is a fair one.
-
-
-
-
GuyM Sal: The leverage part of David's article was flawed -- see the Ballhype discussion that followed. So, as far as I know, we still don't have any solid evidence that strong pen = pythag overperformance.
Links (4)
Another look at run differential
Published 9/17/2007 by David Just at Just Baseball
... The Hardball Times’ Sal Baxamusa “ranted” today about run differential. If you read anything today, make sure his Outsmarting Pythagoras article is on your list. ...
THT: Outsmarting Pythagoras
Published 9/18/2007 at BBTF's Baseball Primer Newsblog
THT: Outsmarting Pythagoras Sal Baxamusa says! As an example, let s look at what happened in the AL West. Here are the standings since Sheehan s column ran: On July 19, 2007 Team W L pW pL RS RA Seattle 53 39 47.0 45.0 456 445 Oakland 45 50 48.7 46.3 402 391 Since July 19, 2007 Team W L pW pL RS RA Seattle 25 31 24.3 31.7 270 311 Oakland 29 27 27.0 29.0 299 308 Wouldn t you know it? The A s have gained four real games on the Mariners since then and have played almost three games better in the Pythagorean world. Furthermore, Seattle s record is just under its Pythagorean projection. Whatever enabled the Mariners to outdo their Pythagorean record earlier in the year seems to have worn off in mid-July, despite Sheehan s observation that the Mariners are the better team and will continue to be so. It s time to stop trying to outsmart ourselves. While we ma ...
Smells Like Padres in Here - Tuesday Edition
Published 9/18/2007 by Winfields Ghost <info@gaslampball.com> at Gaslamp Ball: Front Page Posts
... Those of you who like all your formulas and statistics and crap I don't understand may enjoy this as it references Arizona and how they have defied logic so far. ...
rockies in 6
Published 10/8/2007 by lboros <info@vivaelbirdos.com> at Viva El Birdos: Front Page Posts
... , although his colleague sal baxamusa begs to differ. one thing all parties would agree on: the diamondbacks can't score runs. they finished 14th in the league in scoring and dead last in park-adjusted ops+. also dead last in batting average and on-base percentage. they did finish middle of the pack in homers, dead even with the rockies (171 apiece), but their inability to get on base doesn't bode well. they won't get any help from the colorado staff, which yielded the 3d-fewest walks in the league. nor do the individual matchups offer any encouragement: three of the colorado playoff starters --- fogg, jimenez, and jeff francis --- posted a combined 2.73 era vs the dbacks in 9 starts this year, yielding just 3 hr in 56 innings. all 3 pitchers had era's under 3.00 vs arizona. manny corpas had a 1.64 era vs them in 11 innings; brian fuentes (now the setup man) held the dbacks scoreless in 9.2 innings. ...
OK, and here's my pet peeve: when baseball analysts say that a pythagorean variance is liable to "regress to the mean." Pythagorean variances don't regress to anything, because pythagorean records aren't a "natural" stat. They're a calculation.
Put another way, if the Diamondbacks had a +10 variance in July (just an example), the most likely outcome at the end of the season for them would be +10--not something closer to zero.
That's because the projected pythagorean variance for a team going forward is zero, which doesn't impact its previous variance. When something regresses to the mean, the overall average falls back to the mean.
Phew. Glad I got that off my chest!