THE BOOK cover
The Unwritten Book is Finally Written!
An in-depth analysis of: The sacrifice bunt, batter/pitcher matchups, the intentional base on balls, optimizing a batting lineup, hot and cold streaks, clutch performance, platooning strategies, and much more.
Read Excerpts & Customer Reviews

Buy The Book from Amazon


SABR101 required reading if you enter this site. Check out the Sabermetric Wiki. And interesting baseball books.
MOST RECENT ARTICLES
MAIL : You ask | We say

Advanced


THE BOOK--Playing The Percentages In Baseball

Filter posts by...

 

Tuesday, February 23, 2010

Is PECOTA overzealous?

By Tangotiger, 01:40 PM

The most interesting thing Sky found was this:

Unbiased predictions would cause a regression of PECOTA to WPCT to have a slope of 1 and no intercept. However, using just 2005-2009 data, we see that this is not the case. We see a quite significant intercept of .10 (p-value of .02). Meanwhile the coefficient for PECOTA is .8, where it should be 1.

If you didn’t quite follow, this is the most important thing in his whole article.  What he is saying is that even after you regress individual player performance to create a PECOTA forecast, you STILL have to apply yet another regression to forecast team wins.  Why would you need to do this?  Well, either PECOTA is over-zealous (not enough regression toward the mean) in its player evaluation, or it is not properly modeling a team-season.  Every year, there’s about 10% of PA or IP that go to unexpected players, guys that you didn’t forecast to play much if at all.  And then you have another 10% of PA or IP that goes to guys whose roles changed alot for whatever reason (injury, etc).  All of these things could basically count as noise, that no specific team is in a position to take advantage of the unexpected or unknown (even though you might think that the Yanks might be in a best position and Royals in the worst).

If this is the case, then you need to regress the sum of individual players an addition 20% toward the league mean, when looking at team totals.  That is, while you may want to keep the individual player forecasts as-is to minimize individual player forecasts, you may want to regress the sum of their totals by 20% to minimize team forecasts.

I know VegasWatch has the data from Chone and MGL and the others.  I’d like to know if this is something that applies to all the forecasting systems, or is it PECOTA specifically?  If Chone and MGL are also in the same position, then I may be right.  If they are not, then PECOTA may be over-zealous.

(8) Comments • 2010/02/24 • SabermetricsForecasting
Page 1 of 1 pages

Latest...

COMMENTS

May 26 01:11
Largest demonstration in Canadian history?

May 25 23:40
“Why Kickstarter works”

May 25 19:41
What sabermetrics is NOT

May 25 19:41
Pete Palmer’s new book: Basic Ball

May 25 16:59
Howard Stern

May 25 15:12
Do pitcher’s reach back for velocity when needed?

May 25 12:51
Chad Curtis

May 25 11:26
Lack of hustle during a game

May 25 10:58
Rooting for laundry

May 25 02:38
NFLPA lawsuit against collusion

THREADS

February 23, 2010
Is PECOTA overzealous?