top of page

Astros: The xwOBA Test

For years, I have written about the xwOBA gap and I have said that in April, I think it is a better predictor of full season performance than wOBA or anything else.


The fun thing about this is that what xwOBA CURRENTLY indicates about some of the Astros is pretty different than what OPS, wOBA, or other metrics are showing. Which is right? Well, here is the early data from 2022 for these core players that we have data for THIS year.

The xwOBA for the team 22-24 games into the season for the Astros team was 0.3 percent too low. The AVERAGE error for these eight players was o.6 percent too high.


How does this compare?

  • xwOBA 0.6% too high

  • wOBA 13.7% too low

  • OPS 14.7% too low

  • wRC+ 34.7% too low

The xwOBA is clearly better but some players like Pena were almost 20% too high (xwOBA was the best method for predicting Pena's season stats.) Individual player xwOBAs at this point can have high variability but as a group it is better than ANYTHING else.


Today, let's not argue that any further. Let's just document 23 games into the season what xwOBA and wOBA reflect about our Astros and in October we will revisit this and see which was right in predicting where these Astros batters end up.

The current xwOBA of the Astros team in 0.327 vs. a league average 0.323. The Astros are 14th in the MLB.


Let me pause before this next section and remind you what wOBA and xwOBA are. Here is the MLB.com explanation of wOBA and xwOBA.

The thing to understand as we think about these expected stats going forward is that xwOBA takes into account the actual exit velocity and launch angle of the balls that HAVE ALREADY BEEN put in play. It indicates that based on the profile of the batters contact what results one should have expected. It is NOT a projection stat it is a expected results algorithm. If the exit and velocity and launch angle change, the xwOBA will change.


We should also consider BABIP. League wide BABIP is 0.296. If a batter is significantly higher in BABIP than that they are likely to have been lucky so far. If a batter is significantly lower in BABIP than that they are likely to have been unlucky so far.


Let's play "what if" though. What if the current xwOBAs are correct for each player? I color coded the players that would get hot (RED, pink) and the players that would get cold (DARK BLUE, light blue).

  • Tucker would have an INSANE season- think Judge and Alvarez last year.

  • Bregman and Alvarez would have great years.- Bregman is doing well AND has a BABIP of only 0.235 and a wOBA gap of 0.051- look out league.

  • Dubon would be better than league average.

  • Julks would be marginally below average- it's been fun to see Julks get off to a hot start. His 0.391 BABIP is warning you that it is a lot of luck so far. Expect him to level out at a little below average this year. That's good for a backup OF in his rookie year.

  • Pena projects to truly have the sophomore slump at xwOBA of 0.295- It's not terrible but I don't want to believe that. The projected wOBA is still not as bad.

  • xwOBA hates Chas McCormick and Jake Meyers

    • McCormick's numbers show he has been lucky- 0.346 BABIP and a -0.100 wOBA GAP

    • Meyer's 0.375 BABIP would tend to validate the projection of only 0.285 xwOBA.

    • Our two starting CF have the most negative wOBA gaps and are both expected to drop.

    • The projected wOBAs for both are still not as bad.

    • IF I truly 100% based every move based on this data, one of these two should be traded today.

  • Diaz would hit at below league average- the wOBA Gap and BABIP are telling us that Diaz has been unlucky so far but he might be below average for the first time in his career. His xOBA would be o.284 which is better than the starter. Diaz's projected stats still remain more positive than that and see a near average hitter.

  • Abreu would be better than he has been but it would end up worse than the nine players before. It is concerning. Let's hope the projected wOBA that is holding at above average 0.335 is right.

  • Here is the data point the Maldy fans SHOULD be latching onto.

    • THIS is the data-based argument they COULD make, but I have heard it from NO ONE.

    • If Maldonado actually matched his 0.266 xwOBA the rest of this season, he would end up VERY close to his 2022 season (0.264 wOBA).

    • His BABIP is 0.179 so he has been terribly unlucky, and this validates the wOBA gap prediction he will get better. No one can sustain a 27 wRC+

    • I don't see it. I think he is done.

    • I believe his xwOBA will drop quickly with all of the weak contact he is currently making.

  • Hensley would be better than he is now but still be the worst on the team. He has the biggest wOBA gap on the team and a super low BABIP. Will he even get the opportunity for the luck to change?

It could be a wild ride ahead for some of these guys. Don't doubt the xwOBA predictor too much or you might get burned. But do remember that even this method does a weak job at predicting some indvidual players.


So here is the xwOBA test. Will these early xwOBAs hold true? it is good news for some and bad news for others.



76 views0 comments

Recent Posts

See All

Comments


Post: Blog2_Post
bottom of page