<$BlogRSDURL$>
Proceeds from the ads below will be donated to the Bob Wuesthoff scholarship fund.

Wednesday, August 24, 2005

Saber-Homerism? Or Just Lousy Algorithms?

The adoption of Oakland as the universal pet of the sabermetric crowd has its annoyances to be sure, and one of them is seeing this Baseball Prospectus Hit List for last week. The A's go 1-5 against the likes of the middling Orioles and the bottom-feeding Royals, and they still get an up green arrow? Either that's the most blatant example of saber-homerism I've ever seen, or it's one heck of a flawed algorithm they're using to place those little arrows.

Update: Jay Jaffe responds.


Comments:
They my have moved up solely as a result of the free-falling Sox. But even then, you've got wonder how the Indians don't shoot right past them. And how do the Devil Rays stay stationary?
 
Last year, Oakland was the near-unanimous pick of the Hardball Times, B-Pro & other saber types. It was totally hilarious.

Hey, who did you pick this year, by the way?
 
From Jay Jaffe:

Now, with regards to this week's rankings, I caught some flack from a few readers regarding the A's coming in at #3, rising a notch despite a 1-5 record. But the A's weren't the only high-ranked team to have a bad week. The White Sox had a much worse week, and those of the Angels and Braves, the two teams directly below the A's last week, were nothing to write home about. Here are the run totals of the four teams:
W-L RS RA
A's 1-5 17 24
Angels 3-4 30 27
Braves 2-4 33 36
White Sox 1-5 14 28

The A's, despite losing, at least did a relatively good job of preventing runs, which tends to have a positive effect on those Pythagorean calculations. The Angels actually allowed fewer runs per game, and they've got a better raw run differential this year than the A's, but once all of the adjustments are thrown in, the A's still come out ahead this week.

But not by much. In fact the A's, Indians, and Angels are separated by .0027, with Hit List Factors (the unpublished average of those four winning percentages) of .5660, .5656, and .5633, respectively. That's about 1/3 of a win this far into the season; another run here or there would have likely jumbled those rankings.


http://futilityinfielder.com/blog/2005/08/kiss-my-pythagoras.shtml
 
Well, that would be the A's.
 
> The A's, despite losing, at
> least did a relatively good
> job of preventing runs.

And as we all know, it's much better and far more satisfying to lose games by one run than by 10 runs.
 
Wow. The A's did a good job of preventing runs while losing a home series to a team coming in on a 18-game losing streak and getting swept at home by a sub-.500 club.

Strike up the band.
 
Well, the point of the hit list isn't to rank teams by their records, that's already done by most newspapers. It's an attempt to use other means to rank teams in the hopes of predicting future success. It'd be interesting to see a 3 or 5 year correlation of the hit list results against actual records. Basically track the two and see how well it works.

The real problem is that the A's have been built with formulas like that used by BP in mind. So, if there are biases or innaccuracies built into those equations (which there must be) they will tend to favor the A's. I think that's why the predictors always seem to overestimate the A's success while underestimating others in comparison.
 
I'll stick to using the W - L equations to evaluate how the teams are doing....
 
So, do you, like John Kruk, believe that the Cy Young should go to the pitcher with the most wins regarless of other factors?
 
Rob, I'm a little surprised you haven't mentioned Frank Robinson's latest comments regarding Raffy, knowing your love for the guy. Is there anybody more obsessed with his own place in the game than Robinson? As great a player as he was, he's really starting to bug me.
 
No the Cy Young does not belong to the pitcher with the most wins. Yet the numbers in the run column on the scoreboard at the end of the day determine whether or not one is in Hawaii or playing baseball in October.
Yes, one's record clearly indicates how hot a team is. The A's in the past week lost a series to Kansas City {Kansas City for god sake!!! maybe the worst team in baseball in the 21st century right now}, while the Angels split a series against the defending champions.
To seriously claim the A's are a stronger team than the Angels currently is utterly absolutely absurd.
How can one just use runs scored and allowed without factoring in the team they played to measure the strength of the team is ludicrous, if the BCS used this system, well I don't need to go there...
 
If you read Jaffe's post, he mentions that the rankings take into account factors such as strength of competition and the ball park the games are played in. The A's sucktitude the last few days doesn't completely negate the fact that they have been one of the hottest teams in baseball for the last few weeks.
 
Adam, the "s" at the end of "weeks" is vital here. I guess I'm still waiting for an adequate explanation of how, since those arrows are assigned over one week, the A's get a shiny, happy green.
 
Isn't the green arrow there simply because the A's moved up one slot, most likely because the white sox sucked even worse and the Angels and Indians were unable to pass them by a tiny margin. I think it's pretty obvious that if the next week is like this past one, there won't be a green arrow next to the A's. Like I said before, this week's sucktitude wasn't quite enough to knock them down a peg, but only by a tiny margin.

While the arrow reflects team movement week to week, the rankings are determined based upon the whole season.
 
Adam is absolutely correct.

Repeat after me: The Hit List is based on cumulative year-to-date stats, not just on what happened last week.

Again: The Hit List is based on cumulative year-to-date stats, not just on what happened last week.

One lousy week wasn't enough to narrow the gap between the A's and the Angels. Yes, the Angels have a better raw run differential. But once the run element, park, and quality of competition (EQA allowed, and opponents EQA) measures are taken into account, the A's come out ahead this week by a narrow margin. I'm tickled pink that you take the Hit List so seriously -- do ESPN's power rankings draw any comment? -- but some of you folks are missing the forest for the trees.
 

Post a Comment



Newer›  ‹Older
This page is powered by Blogger. Isn't yours?
Google

WWW 6-4-2