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:

0.312 — J.T. Realmuto batting against Luis Perdomo
0.311 — Dee Gordon batting against Luis Perdomo
0.297 — Martin Prado batting against Luis Perdomo
0.292 — Xander Bogaerts batting against Kevin Gausman
0.292 — Starlin Castro batting against Ivan Nova
0.292 — Christian Yelich batting against Luis Perdomo
0.291 — Mookie Betts batting against Kevin Gausman
0.291 — Marcell Ozuna batting against Luis Perdomo
0.289 — Yunel Escobar batting against Marcus Stroman
0.288 — Cesar Hernandez batting against Mike Foltynewicz

I suspect Luis Perdomo is the pitcher that appears most in the Log5 lists. He gave up a ton of hits in his rookie season of 2016, and continues to do so this year. He is also just coming off the disabled list. It’s a good day to be a Marlins player.

The NN with Park see it this way:

0.311, 0.732 — Dee Gordon batting against Luis Perdomo.
0.312, 0.730 — J.T. Realmuto batting against Luis Perdomo.
0.278, 0.719 — Daniel Murphy batting against Zack Wheeler.
0.284, 0.717 — Jose Altuve batting against Matt Andriese.
0.280, 0.716 — DJ LeMahieu batting against Jeff Samardzija.
0.297, 0.714 — Martin Prado batting against Luis Perdomo.
0.291, 0.712 — Mookie Betts batting against Kevin Gausman.
0.292, 0.711 — Xander Bogaerts batting against Kevin Gausman.
0.275, 0.707 — Corey Seager batting against Shelby Miller.
0.292, 0.706 — Starlin Castro batting against Ivan Nova.
0.289, 0.706 — Yunel Escobar batting against Marcus Stroman.

Dee Gordon comes out on top for the second day in a row. It took him until the 11th inning Saturday night to get a hit, but it still counts. Just to be specific, the NN answers the question, “If a batter is starting a game and a particular pitcher is starting in opposition, what is the probability of the batter getting a hit in the game. While Log5 is specifically looking at the batter-pitcher match-up, the NN takes into account the fact that the number of times a batter faces the starting pitcher is unknown. It might happen zero times. That’s a big reason the batter parameters dominate. The starting pitcher is important, but batters usually face other pitchers in the game. Realmuto owns the better 2017 hit average, while Gordon is better over three seasons. The NN appears to put more weight on the long-term numbers, so Gordon edges out Realmuto.

As always, your best picks have around a 25% of not getting a hit. Good luck!

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