clock menu more-arrow no yes mobile

Filed under:

Exploring Normalized Spin Rates

Using a funny sounding metric

Boston Red Sox v Tampa Bay Rays Photo by Brian Blanco/Getty Images

While spin rate is a seemingly simple tool for evaluating pitches, the complexity of it and its relationship with other measurements makes it challenging to use effectively.

All pitches have a spin rate measurement attached to them from Statcast, but only the spin where the spin axis is orthogonal (perpendicular) to the path of the ball contributes to the movement of the pitch. For example, we know that the backspin of a fastball contributes to its vertical movement, but for a slider, any “gyrospin”, or side spin, doesn’t contribute to the ball’s movement.

But, research done by Driveline Baseball has shown that fastball spin rate is strongly correlated with velocity for an individual pitcher. As a pitcher adds velocity to their fastball, we would expect the spin rate on the pitch to increase as well, which makes comparing spin rates on fastballs with different velocities challenging. Meaning, a fastball with 2400 rpm at 90 mph is much more impressive than a fastball with 2400 rpm at 100 mph. Looking at exclusively spin rate, however, makes them seem equally impressive.

In 2016, Driveline published two articles exploring the relationship between spin rate, velocity and pitch performance using both publicly available MLB data and data collected from their labs. From this study, they introduced the idea of "Bauer Units", or a velocity-normalized spin rate metric, which is calculated by dividing a fastball’s spin rate by its velocity. In doing so, this allows us to effectively remove the impact that velocity has on the spin rate, and focus purely on the spin itself.

In their study, they had a former professional baseball player take swings in a batting cage against fastballs with different spin rates while researchers recorded a variety of biomechanical measurements, including the area of the ball the batter made contact with. They concluded that the spin rate impacted performance as we would theoretically expect — the batter swung under the pitches with higher spin, and the point of contact rose on the ball as the spin rate decreased. In a game setting, having an average contact point closer to the bottom of the baseball would mean more fly balls, pop ups and whiffs.

While this relationship has been observed in a lab, I was interested in looking to see if the same relationship was present in data collected during actual games, and whether Bauer Units (BU) was a better measurement of fastball spin than spin rate alone.

To do this, I looked at the relationship between both fastball spin and Bauer Units and various batted ball performance metrics. While some impact would be expected on pitches across all locations, I suspected that the benefits would be most pronounced on pitches at the upper third of the strike zone.

Using 2016 data, I plotted the relationship between each pitcher's spin rate/Bauer Units and their fly ball rate, whiff rate and pop up rate in the upper third of the zone. Let’s break down the relationship between the spin metrics and each performance metric one at a time to see if Bauer Units gives us a better understanding of how a fastball would perform than pure spin rate.


Here, we see almost no correlation between BU or spin rate and fly ball rate in the upper third of the strike zone (U3 FB%). Looking at these small r-squared values, they suggest that BU explains U3 FB% ten times better than simply spin rate alone. I’m hesitant to conclude that Bauer Units is a great indicator of expected fly ball rate, but it appears to be significantly better than just spin rate.

While being able to identify which fastballs can generate fly balls is helpful, knowing how far these fastballs are expected to go is important as well. I think I could probably get on the mound and generate a lot of fly balls, but these fly balls would undoubtedly be landing far into the seats. However, as discussed on the July 14th episode of the Statcast MLB podcast, exit velocity is ~82% hitting and ~18% pitching, so we wouldn’t be able to get a strong read on the exit velocity of a fly ball even if we could predict the pitchers impact on it with perfect accuracy.

Whiff Rate

While spin rate has higher a correlation, it isn’t the better analytical tool because velocity is not accounted for. It’s obvious that faster pitches have a higher whiff rate, and because of the strong relationship between spin rate and velocity, it is velocity that’s driving the correlation in the above chart.

This situation presents a good opportunity to use Bauer Units – because the impact of pure spin rate is muddied by velocity, BU allows for a clearer picture of how the spin is impacting the pitch.

Pop Up Rate

Pop ups, while infuriating for the hitter and not particularly exciting for the fan, are excellent for the pitcher. They have a BABIP of effectively .000, so a pitcher who can generate a lot of pop ups can pad his stats in a similar way that strike outs would. Looking at the relationship between BU and pop up rate compared to the relationship between spin rate and pop up rate, there is a marginally higher correlation when using Bauer Units, but not large enough for me to comfortably conclude that Bauer Units gives us a better idea of how fastball spin impacts pop ups.

So many factors go into a high fastball becoming a pop up, fly ball, or whiff that I’m not surprised that these r-squared values are so low. But, when taking into account the physics behind the high spin fastball, these relationships give some credence to using Bauer Units as an analytical tool.

Knowing that, holding all else equal, BU gives us a little bit better of a picture of the expected performance of high fastballs, we can look at BU for Rays pitchers this year to get a better understanding of their fastballs and see who looks better or worse when using Bauer Units instead of spin rate.

Bauer Units - Rays Pitchers

Name Spin Rate Velocity BU Indexed BU Indexed Spin Difference
Name Spin Rate Velocity BU Indexed BU Indexed Spin Difference
Sergio Romo 2302 85.4 26.96 112.31 104.64 7.68
Chih-Wei Hu 2431 93.5 26 108.33 110.5 -2.17
Matt Andriese 2395 92.2 25.98 108.23 108.86 -0.63
Blake Snell 2429 94 25.84 107.67 110.41 -2.74
Tommy Hunter 2464 96.2 25.61 106.72 112 -5.28
Austin Pruitt 2342 92 25.46 106.07 106.45 -0.39
Jake Odorizzi 2283 91.5 24.95 103.96 103.77 0.19
Chase Whitley 2229 90.1 24.74 103.08 101.32 1.76
Brad Boxberger 2288 92.5 24.74 103.06 104 -0.94
Ryne Stanek 2425 98.1 24.72 103 110.23 -7.23
Dan Jennings 2266 92.1 24.6 102.52 103 -0.48
Chris Archer 2255 95.5 23.61 98.39 102.5 -4.11
Jacob Faria 2158 91.8 23.51 97.95 98.09 -0.14
Alex Cobb 2086 91.5 22.8 94.99 94.82 0.17
Alex Colome 2120 95 22.32 92.98 96.36 -3.38
Jose Alvarado 2094 98.2 21.32 88.85 95.18 -6.33

Here, we see immediately that much of the Rays’ pitching staff has Bauer Units over the league average of 24. Ideally, a pitcher has both high velocity relative to league average, and high spin relative to that specific velocity, giving him weapons to generate fly balls and whiffs. Let’s take a look at a few specific Rays pitchers of interest.

Blake Snell

Blake Snell stands out as someone with high indexed BU and also a fastball that’s clocked at above league average velocity. This doesn’t come as a surprise, as the pitch consistently received high praise in the minors, landing a 60/65 grade from Dan Farnsworth at Fangraphs before the 2016 season. Despite this, Snell hasn’t had great results on the pitch thus far — he’s generated fewer whiffs and fly balls than the league average fastball would.

The difference between what we would expect to happen and what has happened can be partially explained by the swing rate on Snell’s fastball. This year, only 38% of the fastballs Snell has thrown have been swung at, much fewer than the league average rate of 47%. Hitters are swinging at the fastballs he throws up in the zone, but while his fastball shape and spin suggest that the pitch would be the perfect thing to throw in these spots, he has instead mainly kept it away from hitters and stayed away from the top of the strike zone. Below are heat maps of Snell’s fastball locations against right handed hitters and left handed hitters, as well as heat maps of the swing rates on these pitches.

The Rays tend to pitch away from hitters in general, but I think Snell might be working the ball away so often because it’s a safer option. If he tries to put a fastball at the top of the strike zone and misses low, then it’s grooved down the center of the plate. It’s certainly possible to groove a pitch when trying to work away, but Snell gets a little more wiggle room by working outside – he can miss to some extent in either direction and the margin of error is larger. I’d expect that Snell is still getting some of the benefits from his fastball spin, but perhaps not as much as he theoretically could.

Sergio Romo

Sergio Romo’s elite BU has likely helped him keep his fastball serviceable at its low velocity. Bronson Arroyo, for example, also operates at a similar fastball velocity but has less spin and usually sits at around 23.5 BU. His fastball is also hit for a home run once every 20 swings, compared to Romo’s, which hasn’t left the park yet this year. Romo’s spin rate of 2300 rpm is the 14th highest of all pitchers with a fastball velocity less than 90 mph and while it certainly isn’t ever going to be an even average pitch, it prevents hitters from hunting it and taking advantage when he throws it.

Jake Odorizzi

Jake Odorizzi’s fastball doesn’t have the highest Bauer Units or most spin, but he is the epitome of the high-spin fastball principle. Odorizzi’s fastball generates above average whiffs and fly balls, and he consistently throws it at the top of or above the strike zone. Paired with a split-change, Odorizzi features a devastating difference in vertical movement between the two pitches. In the overlay below, the line with the squares is Odorizzi’s fastball and the line with the triangles is his split-change.

Throughout Odorizzi’s career, he has been steadily increasing the average vertical location of his fastball, which has coincided with an increase in whiffs on the pitch as well. This could be the result of a few different things, but with a generally consistent fastball shape over his career, I’m inclined to believe that his fastball location played a part.

But, there’s an upper limit to how much he can keep increasing his average fastball location before hitters simply stop swinging at it. I’m sure he will adjust accordingly, but it will be interesting to see if he hits that threshold over the next few years.

Finally, average BU helps us better understand how a fastball should perform, but like all average statistics, it can hide the underlying distribution in the data. Looking at the distribution of Bauer Units for Odorizzi’s fastball, we see a similar picture for both 2016 and 2017.

Thus far in 2017, it looks like Odorizzi has eliminated some of the skew in his distribution from 2016 and shifted fastballs closer to his average BU. This is encouraging, as it suggests he is continually increasing the repeatability of his fastball and eliminating the pitches with low spin.

Overall, Bauer Units have shown to be marginally more effective in analyzing fastballs than pure spin rate. The result on a specific pitch is determined by so many independent and inter-related factors that the small r-squared values from comparing BU and performance statistics are understandably low. More important than the actual values themselves is the comparison between Bauer Units and spin rate – small r-squared values aside, if we are going to use spin rate to judge pitches and Bauer Units has been shown to have a stronger relationship, it might make more sense to include Bauer Units in the discussion.

After leading the league in high fastball rate in 2015 and 2016, the Rays have dropped to fourth overall this season, perhaps as a result of changes in personnel and in the way the strike zone is being called. As the rest of the league, notably the Dodgers, adopts a similar strategy, it will be interesting to see if the Rays stick to this approach or make some sort of adjustment to regain a competitive edge. Even with a fly ball heavy staff and rumors of the ball being juiced, the Rays are still able to post a below average home run rate and keep their team ERA down. Here’s to hoping that everyone gets and stays fully healthy for the final stretch of the season so we are able to see this in action.

Spin rate data from Baseball Savant.