Rogers Centre- Credit: DaveMe Images

A primer on using spring training stats and some Blue Jays standouts


Often said to be meaningless, Spring Training stats are painting some interesting pictures for certain Blue Jays





From September 30th to February 21st, not a single Blue Jay registered a plate appearance or pitched an inning in an actual game. For months, fans had to make do with examining and reexamining old data, so it’s not surprising that fans give as much thought as they do to spring training stats produced in basically meaningless games. That said, while the games have no bearing on a team’s World Series chances, we know that they do have some predictive value, thanks to Dan Rosenheck.


Five years ago, curious as to whether or not Spring Training stats had any predictive power, Rosenheck conducted a study. He found that spring training performance (according to a handful of peripheral stats) was correlated with regular season performance. While interesting, that might’ve just meant that good players do well in spring training and during the regular season, while bad players do poorly in both.


To make sure that spring training stats could actually predict something that stats from previous seasons did not, he created player projections that mixed together Dan Szymborski’s ZiPS projections with spring training data. He found that, in fact, spring training stats do help predict future performance and are, thus, worth examining for a purpose greater than simply killing time until the real games start.


For hitters, the most predictive spring training stats are their strikeout rate, walk rate and the rate at which they generate extra bases when they put the ball in play (ISOcon). For pitchers, it’s much the same. Their strikeout and walk rates in spring training provide useful information, as does their ground ball rate (Rosenheck notes that GB% is not publicly available during spring training, so he estimates a pitcher’s GB% using the equation 0.8 multiplied by the number of ground outs surrendered divided by the sum of ground and air outs surrendered).


It is unsurprising that one’s strikeout and walk rates have some predictive power, given that they require the smallest regular season sample sizes to say something meaningful. A batter’s ISO and a pitcher’s ground ball rate need a bit larger sample size than their strikeout or walk rates, but aren’t far behind, taking on meaning far quicker than most other stats.


A recent Eno Sarris post on the stats he prefers to focus on, based on the amount of data one has, also corroborates Rosenheck’s findings, to some extent. With smaller samples, Sarris focuses on more advanced metrics that aren’t publicly available in the spring (like max pitch/exit velocity, pitch mix/shape, swing rates and Command+).


However, once one has about two months of data, Sarris finds strikeout, walk and barrel rates to be quite useful. While we can directly observe the first two stats in the spring, barrels aren’t publicly available until the regular season starts. Interestingly, in emphasizing ISOcon for hitters and GB% for pitchers, Rosenheck indirectly examined barrel rates for both — a higher barrel rate is correlated, to some degree, with a higher ISOcon for batters and a lower GB% for pitchers. That we don’t end up with two months of data over the course of spring training reminds us that while it’s valid to make observations using pre-season data, we shouldn’t get carried with how we adjust our expectations.


One other important stat to account for in the spring is the quality of opposition. In the regular season, it’s fair to assume that every plate appearance matches up two major-leaguers. In the spring, one might go up against a grizzled MLB vet in one plate appearance and a guy with only A-ball experience in the next.


Thankfully, Baseball Reference provides a stat called OppQual. This metric operates on a scale from one to ten, with MLB-caliber opponents given a rating of ten, Triple-A opponents a rating of eight, seven for Double-A, five for High-A, four for Low-A and so on.


In the first few innings of spring training games, major-leaguers are generally matched up against major-leaguers. As such, the OppQual of guys who expect to be on the 26-man roster is generally around seven to eight. On the other hand, those destined to start the season in the minors are going up against players in similar situations, so they tend to have OppQual below seven.


In that vein, if a lower-level prospect is posting crazy numbers in MLB spring training, check their OppQual, as it’s likely they’ve been going up against players at their own level, rather than major-leaguers. While it’s great they are producing well, those results do not suggest that the prospect should necessarily be fast-tracked to the big leaguers.


So, to sum up, a player’s strikeout and walk rates are useful predictors of their regular season performance, whether they’re a batter or a pitcher. It is also useful to examine a batter’s ability to generate power and a pitcher’s ability to limit it. Ideally, we’d be able to directly examine barrel rates in spring but, in their absence, it appears that a batter’s ISOcon and a pitcher’s GB% provide some meaningful information.


While not tested in Rosenheck’s study, these stats are likely more predictive for those who have faced off against relatively high-level opponents. Fortunately, we are now at the point of spring training where lower-level prospects are sent to minor-league camp, which should make the data more and more useful going forward (though this, too, has not been formally tested).


Since you’ve read this far, I would feel bad not to highlight some statistical performances from the first half of spring training. As sample sizes remain small at the moment, it’s necessary to take these stats with a very big grain of salt.


Danny Jansen‘s 17.6% BB rate, 5.9% K rate and .692 ISOcon

When Danny (17 PA, 6.9 OppQual) was coming up through Double-A and Triple-A, he excelled at producing relatively high walk rates and relatively low strikeout rates. As a major-leaguer, he’s been fairly average in both regards, so it’s positive to see him run a good BB/K mix so far. His power generation is another area of strength this spring (he owns a career .205 ISOcon).


Joe Panik‘s 19% BB rate, 9.5% K rate and .400 ISOcon

Panik (21 PA, 8.1 OppQual) has consistently produced a strong BB/K mix throughout his MLB career, so it’s good to see it’s still intact. On the other hand, his production on batted balls has been a roller coaster. Over his last two seasons, during which he posted a 76 and 77 wRC+, Panik was limited to a .095 ISOcon, a key driver of his below-average hitting production. That he’s produced a bit of power against solid opposition is a welcome development, which will bolster his chances of making the Opening Day roster as backup infielder.


Randal Grichuk, Derek Fisher and Teoscar Hernandez‘s collective 17.9% K rate

The Blue Jays outfield is full of guys that generate a lot of power, as well a lot of strikeouts. In 2019, Grichuk (26%), Fisher (34.1%) and Hernandez (33%) each produced a high strikeout rate. This spring, in contrast, they’ve all kept things under control: Fisher has faced a reasonably high level of competition (25 PA, 7.6 OppQual) and maintained the lowest strikeout rate of the bunch (16%); Teoscar (18 PA, 7.1 OppQual) is not too far behind him (16.6%); and Grichuk (24 PA, 7.9 OppQual) has built on the progress he made over the last few months of 2019, posting a 20.8% K rate.


Anthony Alford‘s 56.5% K rate

It’s fair to say that every Jays fan has been hoping to see Alford (23 PA, 6.8 OppQual) break out this spring, as he is a great person with serious skill. Plus, he is out of minor-league options and would need to pass through waivers if he didn’t make the 26-man roster and needed to be sent to Buffalo. While this gives him an edge in the final roster decision, he has played so poorly this spring that it seems distinctly possible that he could clear waivers.


Travis Shaw and Rowdy Tellez‘s collective 46.2% K rate and .600 ISOcon

Based on their showing this spring, the team’s primary first basemen will produce a lot of strikeouts and a lot of power, roughly in line with expectations. Shaw (27 PA, 8.0 OppQual) and Tellez (25 PA, 6.8 OppQual) won’t strikeout nearly half of the time or earn an extra base every other time they put the ball in play, but these extremes are a reasonable window into their near-futures.


Hyun-Jin Ryu and Sean Reid-Foley‘s 0% BB rate

In terms of walk-limitation, these two pitchers could not be polar opposites. When I was reviewing the free-agent starting pitching market this winter, one of the main reasons why Ryu (6.1 IP, 8.3 OppQual) stood out to me was the 3.8% BB rate he posted last season (5.4% career BB rate). While he’s only faced 25 batters so far in spring, it’s nevertheless impressive that he has yet to walk one of them.

That said, it’s a minor miracle that Reid-Foley (5 IP, 7.3 OppQual) has pulled off the same feat thus far, given that he walked more batters (42) in the majors over the last two seasons than Ryu (39), despite Ryu pitching four times as many innings as Reid-Foley (265 vs. 65)!


Matt Shoemaker‘s 42.3% K rate and 4.8% BB rate

Shoemaker (6.2 IP, 8.5 OppQual) was one of the most positive stories from the early days of the 2019 season, which made his knee injury particularly unfortunate. Heading into the 2020 season, his ability to rekindle his 2019 form was a big question mark, so the fact that he has produced such eye-popping K/BB numbers (against relatively strong opponents) over his first two starts is extremely positive.





*Featured Image Courtesy Of DaveMe Images. Prints Available For Purchase.



Support Jays From The Couch On Patreon





Jeff Quattrociocchi

I'm an economics professor in the GTA whose lifelong love for the Jays was reignited by that magical August of 2015 and the amazing moments since.