What We (and the M’s) Mean When We Talk About Spin Rates
One of the best things about this off-season – and admittedly there hasn’t been a whole lot to talk about yet – has been the introduction of the Wheelhouse podcast. It’s essentially a conversation between GM Jerry Dipoto and broadcaster Aaron Goldsmith, and covers the M’s, their focus for the offseason, etc., all with an analytical bent. As such, in addition to talking Shohei Ohtani, the last episode featured a lot of discussion of pitch spin rates, and what the M’s look for in a potential pitcher, and why, for example, Nick Rumbelow or Nick Vincent were guys they acquired. The M’s like high spin fastballs, and they don’t care who knows it.
Now, Nick Vincent checks out as a high spin guy, but if you go to the Statcast data, as Jake Mailhot did in this piece, you find something a bit odd: Nick Rumbelow sure looks like a (well) below-average spin rate pitcher. The podcast also namechecked James Paxton as a high spin guy, but again, he looks below average by Statcast. Andrew Moore, the guy with absolutely elite levels of vertical rise on his fastball – also below average in actual spin rate. What’s going on here?
First, spin and movement are correlated, as spin is what *creates* movement in the first place. No spin, no movement. But if you correlate spin rate and vertical movement of the type measured by pitch fx or Trackman/Statcast, you’ll find that the correlation is pretty darn weak. Again, Andrew Moore had one of the highest vertical movement averages of any pitcher in baseball, but that pure backspin wasn’t the product of elite spin – it was the product of an extremely over-the-top motion that allowed all of the spin he DID impart to work to provide rise. That is, Andrew Moore’s spin was exceptionally efficient – a huge proportion of it went towards moving the ball.
What’s INefficient spin? Doesn’t all spin move the ball? No, that would be too easy. There are two types of spin, as Alan Nathan wrote about in a seminal baseball article at BP, and which I essentially pilfered for this write-up of inefficient spin king Garrett Richards. Transverse spin results in pitch movement. The other type, Gyro or bullet spin (think of a spiral in football), does not. Any pitch has some of both, but the degree to which one pitcher’s fastball is gyro or transverse-heavy varies a great deal. The best example of the *anti-Andrew Moore* – a guy with remarkably LOW spin efficiency – was just in the M’s transaction wire last week: ex-A’s hurler Sam Moll. Moll’s fastball spin rate would’ve been the highest on the M’s, displacing 2017 champions Casey Fien (remember him?) and Ariel Miranda. Moll’s vertical movement is nearly 1 standard deviation *below* average for four-seam fastballs.*
If you imagine two axes – let’s put spin rate on the Y axis and spin efficiency or pure vertical movement on the X axis – you create four quadrants related to spin. In the upper right are the guys with high spin rates and high pitch movement/spin efficiency. Ariel Miranda’s way up in the corner in this hypothetical thought experimenty diagram, with Nick Vincent (whose efficiency isn’t all that remarkable) more towards the middle. In the bottom right are the Andrew Moores, Nick Rumbelows and Ryan Gartons – the spin efficiency specialists. Moll, Rob Whalen and Dan Altavilla are in the upper left, with high spin rates but INefficient spin leading to cutter-like fastball movement. And in the final quadrant, you get the low spin, low movement guys who are more ground ball focused: Felix, Tony Zych, and the pitcher with one of the lowest spin rates in the Statcast database, Jean Machi. It may be better to just show what I mean here:
As Mailhot mentions in his LL piece, it’s possible that Rumbelow’s placement in the Andrew Moore zone is the result of measurement error; maybe Dipoto mentioned him as a spin rate target because in the 2 years since he last pitches in the majors, he’s changed his mechanics and now actually DOES throw a high spin fastball. That’s certainly possible, but the point of this post is to posit that the M’s don’t seem to care so much about spin rate in and of itself. It matters, as many of their targets – from Vincent to Moll to Arquimedes Caminero – may have resulted from spin rate numbers. But it’s clearly not ALL they look at. Moll and Whalen – two high spin, low movement guys with nothing all that pleasing in their stat lines – may have been targeted BECAUSE of their inefficiency, something I mentioned RE: Whalen here. And the team clearly loved Moore (and Paxton, of course) for the movement he generates; they care about Moore’s transverse spin, and DON’T care about a relative paucity of gyro spin. They also went out and acquired Jean Machi last year, a guy with freakishly low spin. Goldsmith asked in the episode if it’s just a matter of being either high or low, and staying out of the middle, and I think he’s on to something, but it makes sense if you view spin as more than just the Statcasted RPM calculation. The M’s want someone with distinctive spin. It can be very efficient, very inefficient, very high, or very low, but they want it out of the middle of the above graphic.
One reason why high spin pitchers may be better than the pure spin efficiency guys is that high spin breaking balls are pretty clearly better. Garrett Richards pairs his cutter-like ultra-inefficient fastball with the game’s highest spinning slider and curve, and batters simply can’t hit them (Richards can’t stay on the field, but that’s another matter). Similarly, Dan Altavilla’s slider has much more break than average, and it was unhittable last year, even as his fastball struggled. Nick Vincent and Rob Whalen also feature high-spin breaking balls. This contrasts with Moore, whose curveball has one of the lowest spin rates on the team. Ryan Garton’s breaker was at least efficient, but neither Rumbelow nor Moore get much movement on their curves, something that helps explain why Moore’s curve was annihilated in the big leagues (in an admittedly tiny sample). With the discussion about the combination of fastball and curve and how one pitch plays off another, a high spin, high movement pitch helps create the separation that Dipoto mentioned.
While they’re ecumenical about spin, they clearly have a favorite outcome, and that’s a fastball up in the zone with high vertical movement. Sorry Sam Moll – the pitch they love is Nick Vincent or Andrew Moore painting the black at the top of the zone with a FB with 10-12″ of vertical rise. That pitch plays into the strategy that I ascribed to Dipoto back in July and that he lays out in detail in the podcast: they want fly balls, and pitchers with lots of movement who can pitch up will generate a ton of them (along with whiffs). That helps the pitchers post consistently low BABIPs, and thus opens a gap between a pitcher’s/team’s FIP and their actual runs allowed. The problem, as we’ve seen, is the home run. Safeco was supposed to ameliorate the flaw in this strategy, but it hasn’t been up to the job, and thus the M’s ultra low-ground ball rate pitchers have hemorrhaged long balls. A good spin rate *is* correlated with improved whiff rates, but at least for the M’s last year, it had ZERO correlation with wOBA on fastballs (it was 0.039). Andrew Moore’s (Ariel Miranda’s) vertical movement wasn’t enough to stave off home runs, and Dan Altavilla’s LOW movement fastball (but high spin rate!) didn’t fare any better.
The M’s need to use this data for more than potential transactions. Most importantly, they need to figure out what kind of HR/FB rate Safeco’s likely to support. In 2017, it’s become too risky to go out and court fly balls, no matter how low of a BABIP you’re able to run. The M’s strategy worked perfectly last year, and they STILL gave up far too many runs. With Mike Leake in the fold, and – fingers crossed – the return of Felix Hernandez, the M’s may give up fewer fly balls in 2018. Now they need to use spin rate information to help tailor their pitchers’ approaches – tweak Andrew Moore’s breaking balls, or see if Miranda can either drop his arm angle a bit (he does this a lot against lefties) or throw a sinker, etc.
* For more on spin and what they call “useful” spin – another way to frame spin efficiency – see this article at Driveline Baseball from Michael O’Connell. It goes into great detail about the differences between pitch types and the specific axis each pitcher’s ball has. They’re able to measure spin directly and isolate transverse spin with Rapsodo technology, which they use to tailor training to each pitcher. Andrew Moore’s a client there, I believe.
A note on data: I pulled statcast spin rates for the 2017 M’s and compared it to movement data from Pitch Info/BrooksBaseball. Whiff rate was calculated from Statcast. All of the data are for four-seam fastballs, so many pitchers had very few (or no) four-seamers to measure. The correlations are NOT weighted by playing time/number of pitches. That means Casey Fien’s handful of pitches counts as much as Miranda’s couple-thousand. The alternative would be to account for pitches, but that just introduces a different kind of bias. I don’t think the way I did it was great, but I didn’t like the alternative. Besides, we’re looking at pitchers’ spin rates and movement, not a game/league average. The low n just means those specific measurements have wider error bars. The correlation of spin and V-Mov was 0.243 and the correlation with wOBA was, as mentioned, 0.039.
Comments
4 Responses to “What We (and the M’s) Mean When We Talk About Spin Rates”
Leave a Reply
You must be logged in to post a comment.
I’ve enjoyed the first two episodes of the Wheelhouse as well.
Marc, thank you for writing this – it’s quite an interesting topic.
I love USSMariner. But it’s at times like this that I wish, off to the side somewhere, there was a permanent open thread for comments about whatever’s in the air, while awaiting the next official post.
…you mean like super mega Dee Gordon trades?
I was expecting Marc to title this article “What We Talk About When We Talk About Spin Rates”.