clock menu more-arrow no yes

Filed under:

Spring Training Stats: Finding the Signal in the Noise

New, comments

Not all spring training stats are meaningless, and some of them are telling us a lot about a few players.

MLB: Spring Training-Washington Nationals at St. Louis Cardinals Jasen Vinlove-USA TODAY Sports

At some point in your life, you’ve surely heard someone tell you not to stress over spring training stats. In general, that’s great advice. In spring training, players are playing in strange ballparks that they won’t see in the regular season, facing a lot of hitters and/or pitchers they won’t face during the regular season, operating on different sleeping schedules with different nutrition, all while trying to introduce new swings or pitches in their repertoires. The 26th through 35th players in each organization frequently see a disproportionate amount of playing time. That leaves a lot of room for variance in performance at the player and team level. It’s no wonder there’s a lot of random noise when translating spring stats to the regular season. However, a few spring stats are meaningful. We can try to separate the signal from the noise.

A few years ago, Dan Rosencheck of The Economist detailed a few spring training stats that were worth following. He presented his info at the MIT Sloan Sports Analytics Conference. His original article can be found here and it’s well worth the read. You can also see his slideshow from that presentation here.

The gist of Rosencheck’s work is that strikeout and walk rates stabilize very quickly, so much so that spring training rates can be predictive for the regular season. That’s true for both hitters and pitchers. Similarly, a hitter’s isolated slugging on contact (ISOCON) and batting average on contact (tasty, tasty BACON) are also predictive.

With that information in mind, we can compare 2018 stats for Cardinal hitters and pitchers against their 2019 spring training using K% and BB% for both hitters and pitchers, and ISOCON and BACON for hitters.

Here’s how pitchers have improved their K-BB% this spring compared to last season. Note that this graph won’t include Hunter Cervenka, Tommy Layne, Chris Beck, Ryan Helsley, Brett Cecil, Giovanny Gallegos, and any pitcher who either didn’t have enough innings this spring or had no MLB record last season. You’ll see Alex Reyes, but keep in mind that his K-BB% from last season is based solely on four fateful innings in Milwaukee. All of this data is through Wednesday’s games.

Jack Flaherty’s gains look modest compared to others, but his 11.3% increase in K-BB% is on top of an already sparkling K-BB% last season. It looks like he’s about to jump to light speed. His K-BB% this spring is 31.3%. If he had done that last season, he would have had a better K-BB% than every other pitcher (140 inning minimum) except Chris Sale. It’s better than Max Scherzer, Justin Verlander, Jacob deGrom, Gerrit Cole, Trevor Bauer, Blake Snell, and... the entire league after Chris Sale.

Dakota Hudson’s jump is just as exciting. He’s turning his electric repertoire into fewer walks and more strikeouts. His rates in both categories even surpass his minor league rates, and they dwarf his 27 inning 2018 Major League sample. He has expanded his K-BB% by a whopping 19.5% this spring. We already knew he could bully hitters into groundballs. Piling more strikeouts and removing walks from that mix makes Hudson a very dangerous pitcher.

As for the others, Tyler Webb and Ryan Meisinger have forced their way into the conversation when the team invariably needs an extra bullpen arm. In Webb’s case, he has separated himself from the other lefty options and could eventually end up with a full-time role.

It’s nice to see John Brebbia, Adam Wainwright, Michael Wacha, and John Gant holding serve. You’d like to see better from Mikolas and Leone, while performances by Poncedeleon and Gomber should leave you concerned about the starting pitching depth past Gant and Hudson. Chasen Shreve’s rates have been alarmingly poor- he’s walking more than two hitters for every one he strikes out. With Shreve struggling and Cecil still a massive question mark on the IL, the door seems open for Layne, Cervenka, and Webb.

Now, we’ll move on to hitters. Below, you’ll see K-BB% differences from last regular season to this spring. The hitter data is through Thursday afternoon’s game. The graph won’t include Adolis Garcia, Jedd Gyorko, Francisco Peña, Edmundo Sosa, and any hitter who either didn’t have enough playing time this spring or had no MLB record last season. I’ve included Yadier Molina and Matt Wieters, though both have limited numbers of plate appearances.

The biggest improvements in K-BB% have come from three young hitters who need to improve their plate discipline- Tyler O’Neill, Harrison Bader, and Paul DeJong. That’s very encouraging. O’Neill especially has made enormous strides in reducing his strikeout rate and increasing his walk rate. If there’s one hitter you should be most excited about based on spring training, it’s the dense pillar of meat. His strikeout rate this spring is under 30% (just barely, at 29.3%) while his walks are up to 13.8%. Along with his speed, defense, and prodigious power, that’s a valuable player.

The newest 2020-2024 Cardinal, Paul Goldschmidt, has improved his strikeout and walk rates, although his spring numbers are more in line with his career plate discipline numbers than his 2018 regular season. In other words, Goldschmidt is taking walks and striking out as if he’s Paul Goldschmidt again.

The majority of players fall in the category with a negligible difference, within 5% (either positive or negative) of their 2018 numbers. Marcell Ozuna is striking out more and walking more. Kolten Wong and José Martínez have seen their strikeout and walk rates slip, but the difference is minor. Yairo Muñoz isn’t walking as much but has shed some strikeouts.

The two alarming profiles belong to Dexter Fowler and Matt Wieters. Fowler’s walk rate is way down. Ordinarily, you might ignore it. Given his 2018 season, though, it’s a troubling sign. Wieters has had limited playing time but he’s striking out like he’s 2018 Tyler O’Neill without doing the other 2018 Tyler O’Neill-y things.

What about isolated power on contact (ISOCON)? I’m hesitant to put too much weight on batting average on contact (BACON) because batting average is such a noisy stat even in its best form. I’m including plate appearances to help understand sample sizes.

Isolated Slugging on Contact, Spring vs. 2018

Name Spring PA 2018 ISOCON Spring ISOCON ISOCON Diff
Name Spring PA 2018 ISOCON Spring ISOCON ISOCON Diff
Tyler O'Neill 58 0.427 0.515 0.088
Harrison Bader 50 0.235 0.028 -0.208
Paul DeJong 51 0.265 0.243 -0.022
Paul Goldschmidt 50 0.351 0.161 -0.190
Jose Martinez 41 0.187 0.133 -0.054
Dexter Fowler 46 0.156 0.212 0.056
Yairo Munoz 41 0.184 0.097 -0.088
Matt Carpenter 33 0.381 0.429 0.048
Marcell Ozuna 40 0.187 0.231 0.043
Kolten Wong 38 0.167 0.143 -0.024
Adolis Garcia 25 0.100 0.000 -0.100
Matt Wieters 24 0.169 0.077 -0.092
Yadier Molina 27 0.201 0.000 -0.201
Edmundo Sosa 19 0.000 0.333 0.333
Francisco Pena 18 0.101 0.077 -0.024
Jedd Gyorko 10 0.194 0.000 -0.194

This is a good time to remind everyone that Roger Dean Stadium is one of the most extreme pitcher’s parks in all of baseball. Per Stat Corner, it had the third lowest HR rate for lefties in 2018 and seventh lowest for righties. For doubles, it’s fifth worst for righties, and a little below average for lefties. It was the sixth toughest ballpark for overall run production. That helps explain some of these ugly ISOCON figures.

O’Neill and DeJong seem to have found another gear, particularly when you adjust for ballpark (DeJong’s park-adjusted ISOCON difference goes from -.022 to .099; O’Neill’s goes all the way up to .392). Wong also moves from a negative ISOCON difference to a positive once we adjust for ballpark.

Fowler and Ozuna see their positive ISOCON difference expand after adjusting for ballpark, with Ozuna increasing from .043 to .141 and Fowler zooming up from .056 to .179. What a difference a late March spring training game can make- both were in the negative, non-ballpark adjusted, as of yesterday morning. The biggest concern comes from Bader. His lack of power is troubling, somewhat short-circuiting his plate discipline gains.

Conclusions

Spring training stats always have to be taken with a grain of salt, even in this case where we’re using data that has a higher level of predictability. However, it’s time to be very excited about the improvement in Tyler O’Neill, Jack Flaherty, and Dakota Hudson. That trio seems poised to at least beat their projections, and each has some breakout potential. Paul DeJong looks like he’s another step closer to establishing a stable, positive hitting baseline.

It’s satisfying to see Michael Wacha, Matt Carpenter, John Brebbia, and Adam Wainwright performing in line with their 2018 seasons. It appears that Webb and Meisinger have shown they can provide solid depth, if not take on enhanced roles as need arises.

Fowler’s plate discipline is a colossal red flag, although Thursday’s power show and this week in general have helped alleviate concerns a little bit. In the pitching staff, the Gomber and Poncedeleon combination has raised new worries although Hudson’s gains are a fine antidote.