clock menu more-arrow no yes

Filed under:

Twins 2018 Season Preview: What did Spring Training tell us?

New, 16 comments

Sure, 99% of Spring Training stats are meaningless. But let’s look at a couple that aren’t.

Toronto Blue Jays v Minnesota Twins
Has Byron Buxton been able to take a step forward in his plate discipline this spring?
Photo by Hannah Foslien/Getty Images

None of us put much stock in Spring Training results or statistics, and for good reason. The lineups are bloated with non-roster invitees; the schedule is littered with split- or intra-squad games; teams’ top-tier players are enjoying plenty of R&R; and a lot of players spend Spring Training on some far-flung backfield working on a newfangled pitch, or on retooling a broken swing, or trying to, say, tame a particularly wild heater.

Combine all the above reasons with an especially small sample size and a tendency to overemphasize the new and different, and you can see why Spring Training is a minefield to navigate for fans and analysts alike: it’s a few weeks of not-actual Major League teams not-actually trying to win.

A team’s Spring Training record is worthless (good grief, there are ties), and so are the vast majority of individual stats. But Spring Training stats aren’t entirely junk. As The Economist’s Dan Rosenheck showed in a 2015 presentation at the Sloan Analytics Conference (and a subsequent article on The Economist’s Game Theory blog), two Spring Training stats in particular — walk rate and strikeout rate — can help provide a fuller picture and prognostication of a player’s upcoming season. Walk and strikeout rate stabilize more quickly than most stats, so looking at these two numbers — even in the wacky context of Spring Training — can help predict the forthcoming season.

As Rosenheck found, Spring Training K-rate correlated pretty strongly with regular season K-rate.

The correlation between Spring Training strikeout rate and regular season strikeout rate. (Note: Rosenheck’s title — “Nothing to see here” — is facetious. Because, you see, there is something to see here.)
From Rosenheck’s article “Spring Forward” in The Economist.

When Rosenheck incorporated walk rate and strikeout rate into Dan Szymborski’s already-robust ZiPS projections, Rosenheck found that the new-and-improved projections performed slightly better in predicting statistics like ERA and OPS than the standard ZiPS numbers.

As Rosenheck explains:

Given two players with identical expectations coming into the year, spring training statistics can cause their projections to diverge by up to 60 points of OPS or ERA—gaps that equate to salary differentials of over $10m a year on the free-agent market. Put another way, players whose forecasts were most aided by their performance during spring training have tended to beat their ZiPS projections by substantial margins, whereas those whose expected value declined as a result of spring training have generally fallen short of their ZiPS forecasts by an equally large amount.

Rosenheck acknowledges that his findings “won’t revolutionize the sport,” but they do help give a fuller picture of what to expect from a player entering a new season. And, with where the Twins reside on the winning competitive curve, the difference between Jose Berrios posting a 3.45 ERA and a 4.15 ERA, or Eddie Rosario posting a .715 OPS or a .775 OPS could be the difference between the playoffs and going home.

With all that in mind, let’s take a look at the walk rates and strikeout rates of some Twins who have seen (at least relatively) significant playing time and determine what bright spots and red flags may be present as Opening Day fast approaches.

Detroit Tigers v Minnesota Twins
Miguel Sano and Jorge Polanco peruse the Spring Training stats. (Adalberto Mejia is not into it.)
Photo by Hannah Foslien/Getty Images

The spring state of the lineup

The Twins had 12 batters reach 40 plate appearances this spring. Two of those dozen — Zack Granite and Gregorio Petit — are not on the Opening Day roster.

Twins Batters Spring Training Walk/K-Rate

Name PA BB K K% BB% BB/K
Name PA BB K K% BB% BB/K
Eduardo Escobar 56 4 9 16.1 7.1 0.44
Robbie Grossman 54 5 7 13.0 9.3 0.71
Byron Buxton 54 1 12 22.2 1.9 0.08
Max Kepler 54 3 12 22.2 5.6 0.25
Ehire Adrianza 52 4 5 9.6 7.7 0.80
Zack Granite 50 5 4 8.0 10.0 1.25
Brian Dozier 50 4 10 20.0 8.0 0.40
Gregorio Petit 46 2 6 13.0 4.3 0.33
Ryan LaMarre 45 4 11 24.4 8.9 0.36
Miguel Sano 46 8 9 19.6 17.4 0.89
Joe Mauer 43 5 7 16.3 11.6 0.71
Mitch Garver 40 4 8 20.0 10.0 0.50
Data via MLB.com

Outfielder Ryan LaMarre, who beat out Granite for the final bench spot, was the Twins’ biggest surprise inclusion on the 25-man roster. By the above plate-discipline numbers, Granite and LaMarre’s spring performances were polar opposites: LaMarre struck out in nearly one-fourth of his preseason plate appearances, while Granite was the only Twin to walk more than he K’ed in Spring Training.

As always, context is important, so I’ve included these 12 batters’ preseason ZiPS projections next to their Spring Training plate-discipline numbers.

Twins Batters’ ZiPS vs. Spring Training Numbers

Name ST BB% ZiPS BB% +/- BB% ST K% ZiPS K% +/- K%
Name ST BB% ZiPS BB% +/- BB% ST K% ZiPS K% +/- K%
Max Kepler 5.6 8.5 -2.9 22.2 20.1 2.1
Joe Mauer 11.6 11.3 0.3 16.3 15.7 0.6
Brian Dozier 8.0 10.4 -2.4 20.0 20.6 -0.6
Eduardo Escobar 7.1 6.4 0.7 16.1 18.9 -2.8
Zack Granite 10.0 6.5 3.5 8.0 12.9 -4.9
Mitch Garver 10.0 9.4 0.6 20.0 25.7 -5.7
Ryan LaMarre 8.9 6.7 2.2 24.4 30.9 -6.5
Robbie Grossman 9.3 13.2 -3.9 13.0 19.8 -6.8
Byron Buxton 1.9 7.2 -5.3 22.2 29.4 -7.2
Gregorio Petit 4.3 4.0 0.3 13.0 20.2 -7.2
Ehire Adrianza 7.7 7.5 0.2 9.6 16.9 -7.3
Miguel Sano 17.4 11.9 5.5 19.6 35.4 -15.8
Data via MLB.com & Fangraphs

And here’s a visual representation of these batters’ plus/minus this spring versus the ZiPS projection for walk and strikeout rates.

Data via MLB.com & Fangraphs; graph by Louie

As you can see, most Twins hitters performed well, cutting down their strikeouts and adding some walks. Miguel Sano, in particular, performed extraordinarily — especially considering the fresh titanium rod in his leg — coaxing eight walks and striking out nine times in 46 plate appearances. Both Sano and Byron Buxton struck out less than ZiPS projects them to this season — perhaps bolstering the case that Sano and Buxton are both poised for big seasons — though Buxton’s anemic 0.08 BB/K ratio doesn’t augur well for any potential change in approach.

Max Kepler is the lone resident of the “Crap Quadrant”; Kepler struck out 2.1% more often during Spring Training than ZiPS projects him for in 2018 and walked 2.9% less than projected. Not a great omen for a player who ended last season so poorly.

Minnesota Twins v Milwaukee Brewers
Taylor Rogers enjoyed a great spring in the strikeout and walk departments.
Photo by Stacy Revere/Getty Images

The spring state of the pitching staff

On the pitching side, the Twins had 14 pitchers who faced at least 40 batters this Spring Training.

Twins’ Pitchers Spring Training Walk/K-rate

Name BF K BB K% BB% K-BB%
Name BF K BB K% BB% K-BB%
Alan Busenitz 41 11 2 26.8 4.9 21.9
Taylor Rogers 63 15 3 23.8 4.8 19.0
Addison Reed 40 7 0 17.5 0.0 17.5
Gabriel Moya 46 12 4 26.1 8.7 17.4
Ryan Pressly 45 11 4 24.4 8.9 15.5
Kyle Gibson 52 9 1 17.3 1.9 15.4
Jose Berrios 79 16 6 20.3 7.6 12.7
Phil Hughes 74 12 3 16.2 4.1 12.1
Trevor Hildenberger 58 10 4 17.2 7.4 9.8
Tyler Kinley 57 13 9 22.8 15.8 7.0
Tyler Duffey 60 7 4 11.7 6.7 5.0
Aaron Slegers 43 4 2 9.3 4.7 4.6
Adalberto Mejia 41 3 3 7.3 7.3 0.0
Jake Odorizzi 40 4 4 10.0 10.0 0.0
Data via MLB.com & Fangraphs

Of the 14 pitchers who faced at least 40 batters, five — Busenitz, Mejia, Duffey, Slegers, and Hughes — won’t be starting the regular season with the big-league club.

Two of those fellas, Busenitz and Hughes, were two of the top strike-throwers for the club in Spring Training: Busenitz led the team in strikeout rate and strikeout rate-minus-walk rate (K%-BB%) this spring, which is a boon for the bullpen if/when he’s called upon later this season.

Though he was put on the DL to start the season, Hughes’ spring feels like a return to the extreme strike-throwing ways that saw him set the big-league K/BB ratio a few years back.

Again, comparing these strikeout and walk rates to preseason projections provides some valuable context. (Note: ZiPS doesn’t predict pitchers’ strikeout/walk rates, so I’ve used the Steamer projection system for the pitching staff.)

Twins Pitchers’ Spring Training K/BB rates vs. Projections

Name ST BB% Steamer BB% +/- BB% ST K% Steamer K% +/- K%
Name ST BB% Steamer BB% +/- BB% ST K% Steamer K% +/- K%
Alan Busenitz 4.9 8.6 -3.7 26.8 20.0 6.8
Tyler Kinley 15.8 11.0 4.8 22.8 19.5 3.3
Taylor Rogers 4.8 7.8 -3.0 23.8 21.1 2.7
Ryan Pressly 8.9 8.7 0.2 24.4 22.6 1.8
Jose Berrios 7.6 8.3 -0.7 20.3 21.9 -1.6
Phil Hughes 4.1 4.8 -0.7 12.1 14.8 -2.7
Trevor Hildenberger 7.4 7.0 0.4 17.2 21.5 -4.3
Addison Reed 0.0 6.9 -6.9 17.5 22.0 -4.5
Aaron Slegers 4.7 7.3 -2.6 9.3 15.0 -5.7
Jake Odorizzi 10.0 8.2 1.8 10.0 19.4 -9.4
Tyler Duffey 6.7 6.1 0.6 11.7 21.6 -9.9
Adalberto Mejia 7.3 8.5 -1.2 7.3 18.7 -11.4
Data via MLB.com & Fangraphs

And the return of the “Crap” and “Quality” quadrants.

Data via MLB.com & Fangraphs; graph by Louie

On the plus side, a handful of relievers had terrific springs: Taylor Rogers, Gabriel Moya, and Busenitz well out-performed their preseason Steamer projections for both strikeout and walk rate, and new-addition Addison Reed didn’t walk a single batter. Ryan Pressly also performed well, keeping his walk rate roughly in line with preseason projections while adding more strikeouts than expected.

Less hearteningly, Trevor Hildenberger, whose projections are a bit conservative compared to Twins’ fans’ expectations, couldn’t meet even Steamer’s lowered bar; Kyle Gibson struck out batters at a higher rate than Hildenberger, who was only one percentage point ahead of Hughes. This may be a bit concerning, considering that Hildenberger endured a pretty heavy workload during his time with the Twins last year.

As far as starters go, Kyle Gibson’s strikeout rate held steady, and he decided to not walk anyone, which is a good recipe for success, while Jose Berrios’ spring hewed closely to Steamer’s 2018 projections. Jake Odorizzi certainly didn’t impress in the strikeout/walk department, notching only four K’s and allowing four walks. He also only faced 40 batters and, as a veteran joining a new club, I’m more willing to view it as noise than signal.

There’s about to be a whole lot of real, actual baseball that counts starting Thursday, so it’ll be easy to overlook the seemingly inconsequential goofiness of Spring Training. But keeping some of this data in the back of the mind may help inform what we’re seeing in these first few weeks of the season and give us that little extra bit of context.

Baseball’s back!