New Run Estimator?
Tom Tango links to a massive paper detailing a new run estimator designed by a reader of his.
I've only skimmed it, and it's probably over my head, but he's looking for peer reviews, so if anyone out there is down to look it over and post comments (here and there), there's your link. Tango says he has one huge reservation he'll talk about tomorrow.
6 months ago
philkid3
6 comments
0 recs |
Comments
Some of his readers mentioned in the comments
that it was “too detailed”, meaning that it included the base situation when a batter was hitting. I guess that is kind of like cheating, as run estimators are really just supposed to use easily accessible stats.
St. Louis Cardinals... defying win expectancy since 2008
by vivaelpujols on May 12, 2009 1:10 AM EDT reply actions 0 recs
Yeah, that really amuses me.
It also uses RBI, which seems like a lot of cheating. I mean, why not just use runs to “estimate” runs?
Beyond the Boxscore // Calling BJ Upton lazy is lazy.
by Sky Kalkman on May 12, 2009 7:56 AM EDT up reply actions 0 recs
that would help
St. Louis Cardinals... defying win expectancy since 2008
by vivaelpujols on May 12, 2009 11:00 AM EDT up reply actions 0 recs
So it may be correllating better with run production by giving hitters credit for things out of their control?
by philkid3 on May 12, 2009 1:58 PM EDT up reply actions 0 recs
The guys is arguing that "timely hitting" is actually a skill
but nevertheless, the base/out situations that the hitter comes up with is completely out of his control.
St. Louis Cardinals... defying win expectancy since 2008
by vivaelpujols on May 12, 2009 3:13 PM EDT up reply actions 0 recs
It's a cheat...
…and a poor one at that. Lord knows if I had the dataset that he’s using for this, I could come up with a run estimator with a MAE below 10 at the team level.
He’s using runs to estimate runs and misapplying it at the individual player level. It’s a mess.
by cwyers on May 12, 2009 8:50 PM EDT up reply actions 0 recs









BtB on Facebook

















