Beat the Streak Picks

Here are the top picks my programs produced for use in Beat the Streak. This post mostly explains the ideas behind the calculations. In addition, this post shows tests on the Neural Network (NN). This post discusses an NN that includes the ballpark. I recently update the models, and the results of those tests are here.

First, the Log5 Method picks:

0.376 — Jose Altuve batting against Mike Pelfrey
0.361 — Daniel Murphy batting against Matt Cain
0.334 — Wilson Ramos batting against Matt Cain
0.334 — Starling Marte batting against Matt Garza
0.314 — Corey Seager batting against Patrick Corbin
0.311 — James Loney batting against Chad Bettis
0.309 — Ian Desmond batting against Dillon Gee
0.308 — Rougned Odor batting against Dillon Gee
0.307 — Adrian Beltre batting against Dillon Gee
0.307 — Yoenis Cespedes batting against Chad Bettis

It appears some of the top hitters in the league are going against pitchers who allow hits. That’s a pretty good combination. The NN list looks like this:

0.376, 0.792 — Jose Altuve batting against Mike Pelfrey.
0.361, 0.773 — Daniel Murphy batting against Matt Cain.
0.334, 0.746 — Starling Marte batting against Matt Garza.
0.314, 0.734 — Corey Seager batting against Patrick Corbin.
0.334, 0.732 — Wilson Ramos batting against Matt Cain.
0.277, 0.724 — Xander Bogaerts batting against Tyler Skaggs.
0.280, 0.717 — Paulo Orlando batting against Lucas Harrell.
0.311, 0.716 — James Loney batting against Chad Bettis.
0.288, 0.715 — DJ LeMahieu batting against Noah Syndergaard.
0.307, 0.715 — Adrian Beltre batting against Dillon Gee.

I like the agreement between the two models at the top of the list. I find this a good confirmation of the Log5 method. The features that the NN detect are subtle. It’s also easy to add features and just build a new model, as with the park effects.

The NN with the ballpark factored in looks like this:

0.376, 0.795 — Jose Altuve batting against Mike Pelfrey.
0.361, 0.777 — Daniel Murphy batting against Matt Cain.
0.334, 0.751 — Starling Marte batting against Matt Garza.
0.334, 0.740 — Wilson Ramos batting against Matt Cain.
0.314, 0.738 — Corey Seager batting against Patrick Corbin.
0.277, 0.724 — Xander Bogaerts batting against Tyler Skaggs.
0.307, 0.724 — Adrian Beltre batting against Dillon Gee.
0.280, 0.722 — Paulo Orlando batting against Lucas Harrell.
0.311, 0.722 — James Loney batting against Chad Bettis.
0.304, 0.720 — Josh Harrison batting against Matt Garza.

Those are very high probabilities for Altuve and Murphy. It looks like a good day to double down. Even with probabilities that high, however, there still is a 1 in 5 chance of the batter not getting a hit in the game. Good luck!

from baseballmusings.com http://ift.tt/2ar0pFC