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:

0.347 — Josh Harrison batting against Chad Bettis
0.340 — Starling Marte batting against Chad Bettis
0.332 — Daniel Murphy batting against Miguel Gonzalez
0.313 — Ryan Braun batting against Bartolo Colon
0.310 — John Jaso batting against Chad Bettis
0.303 — Gregory Polanco batting against Chad Bettis
0.303 — Jonathan Lucroy batting against Bartolo Colon
0.302 — Yunel Escobar batting against Ivan Nova
0.302 — Nomar Mazara batting against Collin McHugh
0.299 — Jung-ho Kang batting against Chad Bettis
0.299 — David Freese batting against Chad Bettis

Bettis is a good target. He has a high BABIP and a high home run rate. With low strikeout and walk rates, a lot of balls are put in play against him. Kang has been hitting well since his return from injury, and currently owns a six-game hit streak.

Here is the NN list:

0.292, 0.757 — Jose Altuve batting against Martin Perez.
0.347, 0.755 — Josh Harrison batting against Chad Bettis.
0.332, 0.741 — Daniel Murphy batting against Miguel Gonzalez.
0.340, 0.740 — Starling Marte batting against Chad Bettis.
0.302, 0.731 — Yunel Escobar batting against Ivan Nova.
0.292, 0.730 — Eduardo Nunez batting against Thomas Koehler.
0.243, 0.728 — Ben Revere batting against Miguel Gonzalez.
0.313, 0.728 — Ryan Braun batting against Bartolo Colon.
0.303, 0.725 — Jonathan Lucroy batting against Bartolo Colon.
0.299, 0.722 — Jung-ho Kang batting against Chad Bettis.

Josh Harrison, Starling Marte, and Daniel Murphy appear to be the consensus picks.

Just to demonstrate why Bettis does so poorly, here are the parameters for the Bettis-Harrison match-up. All the number represent a hit average, hits/plate appearance:

[‘0.283’, ‘0.269’, ‘0.300’, ‘0.285’, ‘0.229’]

The first number is Bettis in 2016. The second number is a three-year weighted average for Bettis. The third number is Harrison 2016. The fourth number is a three-year weighted average for Harrison. The final number is the 2016 MLB average for position players. Short term, both players are well above league average. Long term, both are well above league average, but about 15 points below their short term averages. So these are two players who give up/collect hits at a high rate in general, and more extreme this season.

Again, the best players on this list will not get a hit about 25% of the time, so choose wisely.