As reporters relayed Monday (Adam Wexler's Tweet below), Dusty Baker shared two things:
He planned to play Yainer Diaz in Game 3 at DH in place of Michael Brantley
He claimed Gray has been better vs. LHH than RHH as the REASON Diaz would start in Game 3.
So, is this true?
Has Sonny Gray actually been better vs. lefties than righties and is this a rational argument for starting Diaz over Brantley for the Astros?
To answer this, I am going to invite you all once again to understand the difference between expected stats and actual stats.
In fact, the difference between actual stats and expected stats also impacts what we can expect from Sonny Gray in general in Game 3.
As a reminder, expected stats are based on the quality of the batted ball (exit velocity, launch angle, etc.). Actual stats, even full season stats STILL have a degree of luck built into them.
To illustrate the expected stat conundrum for Gray:
- Sonny Gray's ERA was 2.79 in 2023. (THIRD best of all qualified SP out of 44)
- Sonny Gray's EXPECTED ERA was 3.69 in 2023. (Eleventh best of all qualified SP out of 44)
A 0.90 ERA difference is massive.
- Sonny Gray's wOBA was 0.270 in 2023.
- Sonny Gray's xwOBA was 0.297 in 2023.
Here are some numbers that show one how great Gray has ACTUALLY been vs. both LHH and RHH.
As the heat map colors indicate Gray has ACTUALLY been significantly better vs. RHH but great against both.
So Dusty was wrong? Well, it depends on what he REALLY meant.
As shown in my ALDS Guide, the expected stats show a slightly different story.
Here are Gray's xwOBA stats with each pitch type vs. LHH and RHH.
- Sonny Gray's overall xwOBA vs. LHH is 0.291.
- Sonny Gray's overall xwOBA vs. RHH is 0.303.
Therefore, IF Baker has said:
"Based on his stuff, we expect that RHH can actually hit Gray better than LHH,"
THAT would have been true. The Key issue here is Gray's terrible fastball xwOBA of 0.381 vs. RHH and the fact that he throws it 63% fastballs.
- For all of the radio hosts that chose to slam me about xwOBA and expected stats
- For all of those on Astros Twitter who make fun of me when I talk about xwOBA
- For all of those I have said that teams BASE DECISIONS on EXPECTED STATS
HERE is your evidence.
Brantley WOULD NOT be sitting, and Dusty would be WRONG based on ACTUAL LHH/ RHH stats.
Brantley is sitting, and Dusty would be RIGHT based on EXPECTED LHH/ RHH stats.
Now there is the general truth of those statements because THAT is what Dusty SAID,
Is Starting Diaz over Brantley vs. Gray the right decision?
Diaz is one of those unique players with reverse splits. His expected wOBA platoon split is significantly closer than his actual split. Diaz's xwOBA vs, RHP is actually incrementally better than Brantley's 2022/2023 xwOBA vs. RHP.
This would be true every night though. What about vs. Gray's pitch mix? You can also see this summarized in the ALDS detailed matchups.
These tables are available on the Patreon page.
Based on Gray's NORMAL pitch mix vs. RHH, Diaz is a SLIGHT lean to start over Brantley 0.364 to 0.360 xwOBA.
The issue is I would EXPECT vs. Diaz that Gray would throw curves and sweepers way more than fastballs and given Diaz's 0.268 xwOBA vs. breaking pitches, that would make Brantley the better start.
The table screams out the FAR BETTER decision, however. This was the universal response from Astros Twitter. Diaz SHOULD be starting at catcher, not Maldonado!
Astros fans, is Sonny Gray Better vs. LHH? It depends if you believe in expected statistics or not.
By expected stats Sonny Gray has been better vs. lefties than righties and this is a rational argument for starting Diaz over Brantley for the Astros.
Folks, the data in the table above and on this site in the ALDS Guide and in the ALDS Matchup Breakdown is based on xwOBA data. I KNOW you think it is stat nerd stuff. Matbe when you understand the decision tomorrow - Brantley vs. Diaz- you will understand WHY I build the analysis the way I do. Because THEY do.
Support the Site
Do you want to support this site? Do you want to make sure you continue to get real analysis like this?
It takes time and it does cost money, I really would appreciate your support of the work. Here is your opportunity to do that at a nominal cost. This support will help me continue to provide you what I do- objective analysis of your teams. It will also help the site improve and expand our reach.
If you don't think our work is worthy of financial support, I will accept your non-action feedback humbly.
The details are available at Patreon.com. It is VERY easy to sign up.