April 13, 2017

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.

For 2017, I am just going to publish the Log5 hit averages and the NN probabilities with parks factored in. I am keeping track of the results here.

First, the Log5 Method picks:

0.298 — Corey Seager batting against Brett Anderson
0.286 — Justin Turner batting against Brett Anderson
0.278 — Adrian Gonzalez batting against Brett Anderson
0.278 — Ian Kinsler batting against Phil Hughes
0.276 — Asdrubal Cabrera batting against Wei-Yin Chen
0.273 — Starlin Castro batting against Matt Andriese
0.273 — J.T. Realmuto batting against Robert Gsellman
0.272 — Nick Castellanos batting against Phil Hughes
0.271 — Eduardo Nunez batting against Jonathan Gray
0.271 — Miguel Cabrera batting against Phil Hughes

Daniel Murphy gets the day off as Corey Seager moves into the top spot against Brett Anderson. In fact, the Log5 method likes the better hitters on the Dodgers against Anderson. Anderson gives up few walks without a high strikeout rate, so lots of balls in play can turn into hits. Phil Hughes is similar, and I’m not surprised to see a similar type of hitter like Ian Kinsler rank high against him.

The NN with Park produces this top ten list:

0.298, 0.716 — Corey Seager batting against Brett Anderson.
0.271, 0.695 — Eduardo Nunez batting against Jonathan Gray.
0.273, 0.695 — J.T. Realmuto batting against Robert Gsellman.
0.278, 0.692 — Ian Kinsler batting against Phil Hughes.
0.259, 0.691 — Dustin Pedroia batting against Chad Kuhl.
0.266, 0.690 — Dee Gordon batting against Robert Gsellman.
0.286, 0.690 — Justin Turner batting against Brett Anderson.
0.271, 0.688 — Miguel Cabrera batting against Phil Hughes.
0.246, 0.686 — Yunel Escobar batting against Yu Darvish.
0.257, 0.685 — Starling Marte batting against Eduardo Rodriguez.
0.276, 0.685 — Asdrubal Cabrera batting against Wei-Yin Chen.

The lists are in agreement with Corey Seager at the top of the NN as well. I find the inclusion of Yunel Escobar on the list quite surprising. Note that his expected Log5 hit average against Yu Darvish is just .246. Darvish is a high strikeout pitcher that does not give up many hits. Here is the list of parameters:

[‘0.202’, ‘0.196’, ‘0.249’, ‘0.286’, ‘0.217’, ‘0.227’]

All parameters are hit average, hits/PA. All are regressed toward the 2017 average where appropriate. At this point in the season, the 2017 numbers are highly regressed.

The first two parameters are Darvish in 2017 and 2015-2017 against non-pitchers. Three and four are Escobar in 2017 and 2015-2017. Parameter five is the 2017 MLB average for non-pitchers. The final parameter is the three-year park value for non-pitchers.

It looks to me like the three-year batter hit average dominates the NN. It’s much higher above the league average than Darvish’s numbers are below league average. On a day with few games, that moves him into the top ten.

As always, the best picks on a give day have about a 25% chance of not getting a hit. Good luck!

1 thought on “Beat the Streak Picks

Leave a Reply

Your email address will not be published. Required fields are marked *