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).

First, the Log5 Method picks:

0.334 — Starling Marte batting against Bartolo Colon
0.328 — Daniel Murphy batting against Erik Johnson
0.320 — Josh Harrison batting against Bartolo Colon
0.319 — Xander Bogaerts batting against Tyler Wilson
0.299 — Wilson Ramos batting against Erik Johnson
0.292 — Danny Valencia batting against Colby Lewis
0.292 — John Jaso batting against Bartolo Colon
0.290 — David Freese batting against Bartolo Colon
0.287 — Dustin Pedroia batting against Tyler Wilson
0.286 — Jung-ho Kang batting against Bartolo Colon

It looks like a good day for the Pirates against Colon after a tough game against Noah Syndergaard. Marte looks head and shoulders above the rest in this calculation.

Here is the NN list:

0.320, 0.753 — Josh Harrison batting against Bartolo Colon.
0.334, 0.747 — Starling Marte batting against Bartolo Colon.
0.319, 0.746 — Xander Bogaerts batting against Tyler Wilson.
0.328, 0.738 — Daniel Murphy batting against Erik Johnson.
0.272, 0.734 — Eduardo Nunez batting against CC Sabathia.
0.292, 0.732 — Danny Valencia batting against Colby Lewis.
0.253, 0.728 — Ben Revere batting against Erik Johnson.
0.286, 0.725 — Jung-ho Kang batting against Bartolo Colon.
0.256, 0.721 — Miguel Cabrera batting against Danny Duffy.
0.243, 0.720 — Lorenzo Cain batting against Justin Verlander.

This is about the best agreement I’ve seen between the two systems. The top four are the same, but in a different order. The two systems agree, it’s a good day to be a Pirates hitter.

As always, the warning, a .75 probability of one hit means about a .25 probability of no hits. Don’t be surprised when the perfect match-up yields a zero.

from baseballmusings.com http://ift.tt/23duazd