Saturday, 7 October 2017

MLB Data Revolution

The basic UMPO System suffered losses in the first two games yesterday with American League East rivals Boston Red Sox and New York Yankees both beaten.

The New York Yankees had an 8:3 lead after five innings, but lost in 13 innings. 

When post season games go into extra innings, the value is on the underdog as you might expect in what is essentially a 50 / 50 situation. I've explained before that the favourite is decided by the starting pitchers, and it would be rare for a starting pitcher to still be playing in the tenth innings.  

The MLB database I use only goes back to the 2004 season, so the sample size for play-off extra innings games is naturally small, but of the 36 games to date with an odds-against underdog, the split is now 18:18.

The Chicago Cubs were the first UMPO winner of the season in the day's third game, before the hot favourite Los Angeles Dodgers took care of the Arizona Diamondbacks with no trouble, helped by a four run lead after the first innings.

We may well be into a new era for baseball. The basic UMPO is improved by ignoring bigger 'dogs, and my suspicion, backed up by the improved performance of shorties in the regular season, is that this will now be the case in the post season. More to come on this topic.   
The second was a revelation born of a statistic that only recently came into existence—the launch angle. Radar and camera measurements of the angle at which balls leave the bat have shown that the optimal swing angle looks more like an uppercut than many hitters preferred. Hitters, in turn, have started swinging for the fences in droves. Home runs this season reached a record level.
The record number of home runs this season, and three 100+ teams certainly hint at that, and this article from the Wall Street Journal looks at the (negative) impact on the game of the "data revolution".

No comments: