A cautionary look at some Cubs spring training stats

There is never a shortage of articles that will warn you about getting too excited about spring training numbers, and that’s generally wise. After all, I think we can all agree that despite a Cactus League stat line of .412/.464/.627 for Mike Freeman, it’s correct that hasn’t cracked the 40-man roster. However, it will be interesting to see if that performance piques the interest of any other club.

With all the obvious “small sample size” caveats in mind, earlier this spring Kurt Mensching wrote a piece for The Athletic that highlighted a few statistics that do seem to correlate with improved performance during the regular season. I was intrigued, so I decided to delve a little deeper and see if any of these numbers are relevant to the 2018 Cubs. The results were sort of mixed, but I’ll be curious to return to this post later in the season and see how these observations play out.

Spring stats that matter

First, some background from Mensching’s piece. He relays some information from The Economist’s Dan Rosenheck who demonstrated that contrary to popular belief, their are some statistics that correlate to improvements in the regular season. While traditional measures like ERA or batting average may be misleading, some peripherals stabilize faster:

There are good reasons that these numbers have so little predictive power. Batting average can vary wildly depending on whether a player happens to hit balls right at the fielders or between them—Kevin Costner’s character in “Bull Durham” offers a great speech elucidating this phenomenon. And ERA is largely determined by whether a pitcher happens to give up eight hits in a row, or scatters them at safe intervals over the course of a game. But baseball also provides many other statistics that primarily reflect a player’s actual skills, such as their frequencies of strikeouts and walks, or whether the balls they hit (or allow to be hit against them) tend to travel in the air or on the ground. These “peripheral” numbers tend to stabilise much faster: the year-to-year correlation for qualifying batters’ strikeouts from 2013-14 was .90. And sure enough, they also show a strong connection between spring training and the subsequent regular season (see scatter plot).

The entire slide deck is available online and I highly recommend it, particularly if you’re a sucker for great scatterplot graphs. However, for today, I’m most interested in two of these:

1. Strikeouts and walks. Rosenheck argues that if your K rate or BB rate is substantially different in the spring, it’s an indicator of what will happen in the regular season. Interestingly, that seems to be true for both pitchers and batters.
2. Stolen base attempts. The idea is that if a team is running more during spring training, they are more likely to run in the regular season. This makes intuitive sense, and Len Kasper and Jim Deshaies talked about this during a broadcast earlier this week. Running more in spring training would appear to be a prerequisite for running more in the regular season.

A caveat before I drill down on each of these, I used slightly different timelines for each of these sections. It was mostly an accident. I compiled a bunch of numbers based on timelines that made sense to me and then realized all three of them were different. I toyed with going back to add a year or two here and there, but ultimately liked my original breaks better, so I left them. I’ll address why each time frame is what it is in each section, and apologize for the inconsistency to the statistically obsessed.

Kyle Hendricks

The Cubs starting rotation has looked particularly strong this spring, however given the above about strikeouts and walks, I am pretty excited to see what Hendricks has in store for the regular season. In 19 innings this spring Hendricks struck out 23 and walked two. That is a marked improvement over his career numbers (which are pretty good anyway). It’s also a marked improvement over his best spring training campaign to date (it will shock no one that that was 2016). You can see the comparison below:

I want to be really clear, I’m not arguing that Kyle’s spring training numbers will translate to similar numbers during the regular season. I am mentioning that improvements in K/9 and BB/9 in spring may be correlated with improvements in the same stats in the regular season, and that’s excellent news for Kyle Hendricks and the Cubs if it’s true.

• I considered including 2015 here, but my gut instinct was to start with Hendricks’ breakout year of 2016 and move forward. Hendricks didn’t see a lot of “third time through the order” batters before 2016 and it just felt off to include stats that didn’t reflect that.

Movers and shakers

So, I almost didn’t include this section, because there is a lot of noise here and frankly it wouldn’t take a lot to get me to admit there’s nothing to see here. However, I spent a lot of time doing some math and making some charts, so I decided it would be a shame to not share those with you.

Caution: read these charts with a high degree of skepticism.

Like actually, all of the skepticism.

In fact, this is about the point where I wondered about this entire “maybe some spring training stats matter” thesis, but I was already committed. Plus, I figured I’d at least put something out there that could at least be refuted next year if this came up again.

As I told you above, that whole strikeouts and walks thing apparently applies to hitters as well as pitchers. If someone’s K or BB rate is substantially off career norms in spring training it could be an indicator that that someone may struggle/succeed in the regular season. With that in mind I compiled career and 2017 numbers for players with over 150 plate appearances in 2017 and compared them to this year’s spring training. The results are below:

The top five movers didn’t move in a good direction (you’d want a negative change in K rate and a positive one in BB, and well, none of the top five changes are in the right direction.) The largest movement in K rate, in order, from 2017 to spring training were Jason Heyward, Ben Zobrist, Kris Bryant and Anthony Rizzo. All of them were striking out more often in spring training than they were in 2017. The fifth largest mover was Tommy La Stella, whose walk rate dropped 9.1% in spring training vs. 2017. Again, this isn’t destiny and spring training sample sizes are small, but let’s hope these are anomalous.

• Why just one year of data and career numbers? I wasn’t trying to look at comparisons of every year and every spring training ever in this section. I just wanted to know if movement in this spring was dramatically different from the previous year, while recognizing that the previous year could have been odd statistically. Plus the table was already really busy, so I just let it go after one year.

Run away?

The last stat that is supposed to give us some predictive value out of spring training is how much a team tries to steal bases in spring training. So I went back and looked at the last three years of Cubs SB/CS numbers and compared it to this year.

I immediately found a bunch of stuff that doesn’t back up the running correlation, at least in the case of the Cubs. I thought about just chalking this up to Joe Maddon and calling it a day, but then I remembered the last section where I gave everyone a heart attack over K rates and decided it might be worthwhile to throw some shade at predictive spring training stats.

The 2017 Cubs didn’t run a lot. They had 62 stolen bases in 2017, which was good for 24th out of 30 major league teams. The 2017 spring training Cubs were tied for fourth-most stolen bases in spring training.

The 2016 Cubs ran slightly more, 66 stolen bases which put them 20th overall. But if you were basing this on spring training data you would have thought they’d run a lot more. The 2016 Cubs ran more than anyone in spring training.

Interestingly, the 2015 Cubs were maybe the worst team at stealing bases in spring training. They were in a three-way tie for 28th in the majors with 13 stolen bases, but they also had the dubious distinction of having the most CS (also 13, clearly not lucky) of that group. Ergo, probably worst team at stealing bases in spring training. So I was a little surprised to check out their regular season numbers and see that of the three years I was looking at, 2015 was by far the best year for Cubs stolen bases — they stole 95 bases, good for 9th overall in the major leagues.

I’ve heard some people comment on the 2018 Cubs running more in spring training, and that’s true, but I’d take it with a grain of salt for right now. Cubs team running numbers in spring training and the regular season don’t seem to have a lot of relationship in the Maddon era.

I do think it’s worth mentioning that the original study was based on individual players and not team philosophy. It’s possible that a lot of the stolen bases that were driving spring training totals in all of these years were people who didn’t wind up making the club.

• Why three years rather than two like Hendricks or one like the hitters? Well, three years is the Joe Maddon-era Cubs and it seemed like a better break point for a statistic that is largely driven by game strategy.

Takeaways

I’m still pretty skeptical of spring training stats, but after reading Rosenheck’s work, I’ve been more than a little curious about what, if anything, it meant for the Cubs. The results were sort of mixed, but if they persist into the season they give us a pretty interesting set of stats to keep an eye on during next year’s Cactus League games.