Poll 2021: Which Group of Hitters Performs Better?
Yesterday, I asked you to vote on which group of pitchers you expect to post a better ERA over the rest of the season. One group was composed of the 10 biggest SIERA overperformers, while the other included the underperformers. For the first time, I’m going to take the same polling idea and use it for hitters. So let’s follow the same concept and compare two groups of hitters based on xwOBA overperformance and underperformance. We know that xwOBA isn’t perfect. Neither is SIERA. In fact, no estimated/expected/forecasted equation is going to be perfect because there will always be a player or multiples that figure out how to do something we have a difficult time quantifying or there’s simply bound to be players each year that fall into either end of the extremes for no reason at all except for randomness. So let’s keep that in mind when reviewing these two groups.
My initial population group consisted of 115 hitters who have put at least 200 balls into play according to Statcast. Group A is composed of the 10 largest xwOBA overperformers, while Group B is composed of the 10 largest xwOBA underperformers.
| Player | BA | xBA | SLG | xSLG | wOBA | xwOBA | Diff |
|---|---|---|---|---|---|---|---|
| Cedric Mullins | 0.314 | 0.275 | 0.541 | 0.453 | 0.391 | 0.350 | 0.041 |
| Adam Frazier | 0.330 | 0.296 | 0.463 | 0.393 | 0.375 | 0.336 | 0.039 |
| David Fletcher | 0.309 | 0.275 | 0.381 | 0.321 | 0.314 | 0.277 | 0.037 |
| Marcus Semien | 0.277 | 0.246 | 0.528 | 0.460 | 0.372 | 0.336 | 0.036 |
| Jared Walsh | 0.278 | 0.266 | 0.556 | 0.479 | 0.374 | 0.340 | 0.034 |
| Yuli Gurriel | 0.313 | 0.274 | 0.472 | 0.392 | 0.364 | 0.332 | 0.032 |
| Randy Arozarena | 0.251 | 0.213 | 0.400 | 0.352 | 0.321 | 0.290 | 0.031 |
| Xander Bogaerts | 0.321 | 0.287 | 0.545 | 0.492 | 0.396 | 0.366 | 0.030 |
| Raimel Tapia | 0.283 | 0.259 | 0.387 | 0.333 | 0.316 | 0.289 | 0.027 |
| Garrett Hampson | 0.253 | 0.224 | 0.415 | 0.377 | 0.308 | 0.282 | 0.026 |
| Group Average | 0.296 | 0.264 | 0.468 | 0.403 | 0.354 | 0.320 | 0.034 |
| League Average | 0.240 | 0.243 | 0.403 | 0.413 | 0.313 | 0.319 | -0.006 |
| Player | BA | xBA | SLG | xSLG | wOBA | xwOBA | Diff |
|---|---|---|---|---|---|---|---|
| Kyle Tucker | 0.271 | 0.319 | 0.503 | 0.602 | 0.351 | 0.409 | -0.058 |
| Aaron Judge | 0.282 | 0.327 | 0.526 | 0.639 | 0.386 | 0.441 | -0.055 |
| Juan Soto | 0.283 | 0.315 | 0.445 | 0.543 | 0.369 | 0.418 | -0.049 |
| Paul Goldschmidt | 0.265 | 0.288 | 0.432 | 0.546 | 0.335 | 0.383 | -0.048 |
| Elvis Andrus | 0.235 | 0.276 | 0.322 | 0.396 | 0.261 | 0.308 | -0.047 |
| Alec Bohm | 0.243 | 0.273 | 0.343 | 0.415 | 0.283 | 0.325 | -0.042 |
| Freddie Freeman | 0.274 | 0.305 | 0.489 | 0.582 | 0.371 | 0.413 | -0.042 |
| Eugenio Suárez | 0.175 | 0.211 | 0.372 | 0.438 | 0.275 | 0.314 | -0.039 |
| Charlie Blackmon | 0.261 | 0.288 | 0.364 | 0.436 | 0.324 | 0.360 | -0.036 |
| Kevin Newman | 0.210 | 0.246 | 0.273 | 0.316 | 0.233 | 0.268 | -0.035 |
| Group Average | 0.250 | 0.285 | 0.405 | 0.489 | 0.318 | 0.362 | -0.045 |
| League Average | 0.240 | 0.243 | 0.403 | 0.413 | 0.313 | 0.319 | -0.006 |
| Group | BA | xBA | SLG | xSLG | wOBA | xwOBA | Diff |
|---|---|---|---|---|---|---|---|
| A | 0.296 | 0.264 | 0.468 | 0.403 | 0.354 | 0.320 | 0.034 |
| B | 0.250 | 0.285 | 0.405 | 0.489 | 0.318 | 0.362 | -0.045 |
These are some interesting groups and unlikely to be the names you might expect to see on each of these lists. It certainly appears that Group A has more speed, but it also has its share of slowies in Walsh and Gurriel, while B features just one name in Newman who clearly owns above average speed. Supposedly, xBA has factored in a batter’s Sprint Speed since 2019, but I’ve been skeptical that that’s the case, or at least it hasn’t been factored in enough. That was from my own subjective feeling after working on an updated xBABIP equation, but Alex Chamberlain confirms this as well. So this is a weakness to be aware of, but it doesn’t mean xwOBA in its current form is useless as a batter could still overperform or underperform their mark ignoring the speed issue.
Group A has posted a significantly stronger batting average than B, but has surprisingly posted an xBA 0.021 points lower. I’m guessing that a lot of that gap could be explained by the speed issue mentioned above, and perhaps once adjusting for speed, the two xBA marks might be quite close to each other. That said, it’s hard to imagine that speed alone would explain the gaps between each group’s BA and xBA.
Next is SLG, in which speed has less of an impact, but still an effect nonetheless. While speed has the greatest impact on infield ground balls that end up as singles, speedier hitters could also turn singles into doubles and doubles into triples, raising their SLG that xSLG isn’t fully reflecting. Again though, speed is likely to play a more impactful role in BA than SLG, so the gaps here shouldn’t be explained as much by it and therefore be more actionable. It’s interesting that Group A’s xSLG is actually all the way down at Group B’s actual SLG, while Group B’s xSLG is actually higher than Group A’s actual SLG.
Overall, we find that Statcast calculates Group A should be posting a wOBA well below Group B, which is not what we found during yesterday’s pitcher SIERA groups, where the underperformers still deserved a higher ERA, with the gap being significantly narrower by SIERA. Here, Group B actually looks like the significantly better group of hitters. And by simply scanning the names, it surely looks like a far better group as well.
So which group performs better over the rest of the season? Let’s get to the poll questions. Feel free to share your poll answers and why you voted the way you did.
Mike Podhorzer is the founder of ProjectingX IQ, an advanced fantasy baseball analytics platform that transforms projection data and in-season performance signals into actionable intelligence. He is the 2015 Fantasy Sports Writers Association Baseball Writer of the Year and three-time Tout Wars champion. He is the author of the eBook Projecting X 2.0: How to Forecast Baseball Player Performance, which teaches you how to project players yourself. Follow Mike on X@MikePodhorzer and contact him via email.
How does first half vs. second half wOBA look historically league-wide? With warmer weather, pitchers tiring, etc., I imagine it tends to be higher. (There’s probably an effect from September callups, too, which aren’t really a thing anymore, so I suppose you could leave Sep. numbers out of it.)
With sticky stuff enforcement, any upward trends might be more pronounced as well.
Unlike the pitchers, this one isn’t even close. The 2nd group wayyyyy outperforms the 1st group ROS.
Ha, you would certainly think so…if you believe in xwOBA!
In both of these exercises (yesterday and today) I think we are missing an important factor: has this player consistently outperformed (or underperformed) this particular metric? If the player has done so for several years, I would anticipate that there is something in the player’s profile that isn’t well-captured by the metric, resulting in this systematic error.
If Group A had a bunch of guys that consistently outperformed their xwOBA, while Group B had guys who sometimes did better than their xwOBA and sometimes did worse, then I would be more confident in voting for Group A to continue their performance for the second half.
In the future, I think the pool you use to do this comparison should consider this factor as well. You may have, but showing this type of information would be more useful for such a comparison.
I completely agree, but that’s the fun of it — it’s up to you to check the players and determine whether that consistent under over overperformance, if it has been occurring, will continue.
Off-topic issue Mike but have you ever looked into Z-Swing%-O-swing% as some sort of useful stat? I mean, yes Luis Robert swings a lot out of the zone but he swings a lot in the zone also. Conversely, yes Biggio doesn’t chase but he doesn’t swing a lot in the zone either. I was thinking it makes sense to look at it as another tool in evaluating plate discipline. It wouldn’t be perfect since obviously there are plenty of strikes you don’t want to swing at and some very hittable pitches an inch off the plate, but in general it captures swing at strikes and not at balls.
I’ve been playing around with it and pulled all qualified hitters and did a scatter plot against both wOBA and xwOBA and you don’t have to be a mathematician to see a pretty strong relationship, especially with xwOBA. I kind of think it should be a stat on FG pages myself.
BTW, the guy who got me looking into it is Kyle Tucker. He leads baseball this year in Z-swing-O-swing and by a good amount. He’s about average in chase rate but swings a lot in the zone.
I may have looked into it in an old xK% equation or something else. I know I’ve investigated all those plate discipline stats years ago but can’t remember anything besides what was included in my original equation.
I’m guessing the relationship is strong because of the relationship with strikeout and walk rates which will directly affect wOBA. I don’t know if batted ball quality could be explained at al by the two metrics.
The answer lay, I think, not in whether we have enough sample size to determine whether *player* X is likely to consistently beat or underperform xwOBA, but whether we have enough sample size to determine whether *group* X is likely to consistently beat or underperform xwOBA.
Each individual has 200+ balls in play, but each group has 2,000+ balls in play to analyze. I expect Group A will outperform Group B, though by a smaller margin than in the first half.
As for individuals, Fletcher, Bogaerts, and Tapia already beat xwOBA by a large amount last year, by 0.029 or more. In fact, 29 qualifying hitters had a wOBA-xwOBA of 0.026 (Hampson’s number for 2021 first half) last year. Gurriel is someone who went from negative differential to positive this year – one would expect him to be a better candidate for regression downward than the guys who have proven they can beat xwOBA.
Suarez, Freeman, and Goldschmidt were negative differentials last year who are still negative differentials this year. Blackmon was positive last year, negative this year. He’d be a candidate for regression upward.