In my previous post I used my baseball simulator to look at the best lineup for the Kansas City Royals. Up at the plate this time are the Pittsburgh Pirates. In this exercise I will use my simulator to play actual baseball games, in fact over 2 million games per lineup to find out which lineup is the most productive vs both a RHP and a LHP. Each lineup is compared in wins per 162 games against the default lineup from MLBDepthCharts. Player projections are taken from Bill James.
Below are the results...
Default lineup from MLB Depth Charts.
1. Starling Marte
2. Neil Walker (S)
3. Andrew McCutchen
4. Garrett Jones (L)
5. Pedro Alvarez (L)
6. Russell Martin
7. Travis Snider (L) << Gaby Sanchez vs LHP
8. Clint Barmes
9. Pitchers Spot
Best Lineup vs RHP
1. Russell Martin
2. Andrew McCutchen
3. Neil Walker (S)
4. Garrett Jones (L)
5. Pedro Alvarez (L)
6. Travis Snider (L)
7. Clint Barmes
8. Pitchers Spot
9. Starling Marte
Wins/162 Games Better Than Default Lineup: 0.84
Best Lineup vs LHP
1. Russell Martin
2. Andrew McCutchen
3. Garrett Jones (L)
4. Neil Walker (S)
5. Starling Marte
6. Gaby Sanchez
7. Clint Barmes
8. Pitchers Spot
9. Pedro Alvarez (L)
Wins/162 Games Better Than Default Lineup: 1.48
Skinny: The thing that pops up right away is that the simulator likes Russell Martin and his low batting average and decent OBP at the top of the order. The top four lineup spots are almost the same vs both RH/LH pitchers with Jones and Walker flopped at the 3rd and 4th spots. Marte who is projected by MLBDepthCharts to leadoff finds himself batting 9th against RHP and 5th against LHP. I am a little surprised at Marte batting 9th against RHP but there are a slew of left handed batters that bump him down, so it really comes down to batting 7th in front of the pitcher, or batting 9th where he comes in to play as a table setter (leadoff man) at the bottom of the order, kind of like Alvarez does against LHP.
And if you haven't noticed, batting the pitcher 8th and not 9th pretty much helps every single NL team.
Excellent, thanks!
ReplyDeleteMore Pirates-specific comments will go back at BucsDugout, but this made me think of a general issue -- do you take into account vulnerability to LOOGYs? The RHP lineup seems like it'd be very vulnerable to a LH reliever -- Jones, Walker, and Alvarez all have huge platoon splits. I wonder how you'd program that in -- maybe simulate the vs. RHP lineup facing a LH reliever an inning every other game, if that reduces the run expectancy.
Also, I'd like to see this using a non-Bill James projection, but that's assigning you a lot of extra work. Thanks for the post
All good questions. The simulator does have an algorithm for changing pitchers and part of what it looks at is how many left handed or right handed hitters are coming up. But it also can pinch hit to get the platoon advantage back for that at bat. I'm not sure how much using a different projection system would change things, but that's a good point.
ReplyDeleteThanks for the thoughtful comment!