Can We Draw Conclusions from Bullpens Thrown at Submaximal Intent Levels?

(Note: this was a post I wrote in March 2019 for True Grind Systems, a sports performance training facility in Austin, TX. You can find the original post here. All media included is my own or was extracted from Rapsodo Cloud. Because I've learned a considerable amount since this post was written, there have been some added notes and corrections italicized in bold throughout.)

Of the many new trends throughout the baseball training world, one that I keep seeing pop up is the role of “intent.” Intent can be defined in several different ways, but for the sake of specificity within this post, I’ll simply describe it as the overall level of effort within each repetition. In this case, let’s say our repetitions are individual throws off of a mound.

High intent can help us gauge how our training outcomes will translate into game scenarios and can allow us to better facilitate preparedness for our next outing. Generally, we can have a good idea for what type of arsenal we’ve equipped ourselves with before we ever face a batter if we’ve already simulated everything we can outside of an actual game. For most pitchers, throwing all of their pitches off of a mound at game intensity at least once, if not several times, before ever stepping on a mound helps with this feel.

However, it is not plausible, or at least not intelligent, to throw at maximal intent off of a mound every day. And, with the rise of the tech revolution in baseball, many pitchers are wanting to utilize tools such as Rapsodo as frequently as possible. This leads to pitchers throwing bullpens often at less than 100% RPE, or Rate of Perceived Exertion, while monitoring the pens with a pitch tracking tool.

At first, this seems like a great idea. See what your pitches are doing and adjust as needed while keeping your workload low. Maybe even design a new pitch without taxing your arm. However, what we need to remember as players and as coaches is that data from bullpens thrown at submax intent may not accurately reflect who our pitchers really are. How do we know how much our repertoire might differ on an indoor turf mound at 75% RPE intent compared to their 5th start of the season in the middle of March?

Luckily for me, I have a 17-pitch bullpen report from Rapsodo during which I threw a bullpen at 75% RPE while rehabbing shoulder surgery over a year ago. At the end of that pen, I “let loose” on a couple fastballs at roughly 90-95% RPE to see where I was at. Here’s the fastball report:

The sample size is very small, but right away you should see some major differences in the first five fastballs (one outlier omitted) thrown at 75% RPE and the last two fastballs thrown at 90-95% RPE. Aside from the obvious uptick in velocity, we see:

• avg spin rate go from 1927 to 2253 RPM resulting in a 326 RPM increase
• a very slight increase in tilt for spin axis
• avg horiz. break goes from 7.7 in to 10.9 in resulting in a 3.2 in increase
• avg vert. break goes from 16.12 in to 14.05 in resulting in a 2.07 in decrease
• avg release height goes from 4.66 ft to 3.9 ft resulting in a 9 in decrease

Again, the sample size is very small, but we do see some significant changes in pitch characteristics just because I threw the ball a little harder.

Now let’s analyze these points. I want to combine the first two, my 326 RPM increase in spin rate and the small tilt in spin axis, with the fact that spin efficiency was relatively constant throughout the pen. If we only looked at my first five FBs, we see a below average raw spin rate of 1927 RPM (MLB average is roughly 2200 RPM for FBs) with a high spin efficiency and a pretty vertical spin axis. As a general rule of thumb, a pitcher with a spin rate sitting well below 2200 RPM means he should be living in the lower half of the strike zone with his FB, especially with such high spin efficiency and an almost vertical axis. This is a pitch that has some “sink” from the batter’s viewpoint and does not generate much “lift” from spin against gravity.

*something I overlooked when this was originally written was the effect of spin relative to velocity. To normalize this relationship for comparison between pitchers, we use the Driveline Baseball-created Bauer Units. BUs equal total spin divided by velocity. 

(BU = spin rate (RPM) / velocity (MPH)

The MLB average Bauer Units are roughly 24. So, for my first five FBs at submax effort, I averaged 76.78 MPH with an average spin rate of 1,851 RPM, equaling an average of 24.1 Bauer Units. 

This is essentially average. Without going too in-depth on this, pitchers with above average Bauer Units (assuming high spin efficiency), should concentrate on throwing more fastballs up in the zone. Their spin rate is high relative to their velocity, meaning the magnus effect (more on that below) is creating the perception that the ball is rising as it gets to the hitter. Pitchers with below average Bauer Units might want to focus more on throwing down in the zone, as they have low spin relative to their velocity. These are pitches that appear to sink, utilizing more of gravity's effect than magnus force due to spin. 

In correcting what I suggested about a pitcher with a spin rate under 2200 RPM, we have to take into consideration my average Bauer Units of 24.1. With this set of fastballs, I would not necessarily want to live down in the zone. What can be inferred from my average Bauer Units is that this set of fastballs do not create either the perception of rise or sink. In these instances, it is generally recommended that the pitcher look to develop a two-seam fastball or utilize breaking pitches more often in their gameplan.*

However, when looking at the last two FBs, you’ll see pitches with league average spin rate, a tilted, non-vertical axis, and a spin efficiency that is still high. This type of FB is one that you don’t necessarily want to exclusively feature on the lower half of the zone. The spin rate is not low, which means the same type of “sink” presented on the first five FBs is no longer present. We also don’t see this pitch intentionally thrown up in the zone very often due to the fact that it doesn’t present much “rise,” either. The tilt will create horizontal movement, and a small, less vertical amount of sink (more on this below), which is where we can start to be creative in regards to our plan for attacking hitters.

*again, we have to take Bauer Units into consideration. In these last two fastballs at near max intent, I average 25.6 Bauer Units. This is considerably higher than the first set of five fastballs. Here, I have an above average Bauer Unit measurement with a high spin efficiency. Knowing this, I can plan on featuring my fastball up in the zone, as it will appear to have more rise or "hop" to the hitter. 

This is another great example of the role that intent plays in bullpens. Based off of normalizing the relationship between velocity and spin in these sets of fastballs alone, we start to create an entirely different fastball profile due to a single characteristic that changed drastically with intent.*

Based off of this contrast alone, these are two different pitches that carry different characteristics for getting hitters out.

Secondly, let’s analyze the change in spin axis combined with the changes in horizontal and vertical break from these sets of FBs. While the spin axis tilts very slightly (imagine the face of a clock), going from an average of 00:50 to 1:14, or 24 “minutes,” this can significantly change pitch characteristics. When we consider the 326 RPM increase in addition to this tilt in axis, it makes sense to see the increase in horizontal break and decrease in vertical break. The increase in horizontal break can be explained by a couple of things: one, the axis shifting from almost vertical to a very slight rightward tilt generates more arm-side run on the pitch due to magnus force acting in that direction on the ball.

Secondly, the big 326 RPM spike would magnify this arm-side run magnus effect even more. The decrease in vertical break is a little more difficult to explain due to the fact that spin rate increased but the ball did not generate as much lift – it decreased by roughly 2 in. There are two factors in play here: one, the tilt in axis reduces some of the vertical lift that was generated by magnus force when the pitch had a much more vertical axis. However, some of this reduction is negated by the fact that spin rate increased so much. Essentially, we are seeing the tilt in axis reduce lift while the increase in spin rate fights some of that tilt’s effect. This creates less vertical separation between the two sets of FBs than we would see if the spin rate remained the same throughout.

The last difference noticed between these two groups of fastballs was a 9 in decrease in release height. This is somewhat unexpected and is presented with less evidence as to why this might occur. One reason could be that an increase in effort might change mechanics to a point where we see that drastic of a change. Another reason could be that I tend to feel more in-sync with my delivery at full intent and, as a result, a more natural arm slot was exposed in contrast with a more reserved, accuracy-focused delivery. Whatever the case may be, this significant of a difference in release point has the potential to create a vastly different FB profile.

All in all for the FB report, we can see that the changes in data on pitches thrown at 90-95% RPE create a completely different pitch. We see a faster moving ball, a ball that generates more horizontal movement, a ball that is released much lower, and a ball that is most likely perceived to have more lift from the batter’s viewpoint (even with net vertical break actually decreasing).
While I didn’t throw any 2-seam FBs or cutters at more than 75% RPE, there is still something to be learned from the data on these pitches, as well. Again, sample size is very small.

Let’s start with the 2-seam FB report:

Now, before analyzing any of these numbers, let’s take a look at the horizontal & vertical break plots for these pitches compared to the 4-seam FB:

(2S is above in blue, 4S is below in red. Don’t think of these charts as looking at quadrants of a strike zone, but rather a simple way to measure amount of break on each pitch.)

At first glance, these are all basically the same pitch. If we want to look closer, we’ll see that the 2-seam actually has less horizontal movement or arm-side run. In fact, these three 2-seams averaged 4.33 in of positive horizontal break, which is a little over half as much as my 4S moved laterally at 75% RPE, and two and a half times less than the 4S at 90-95% RPE.

This pitch also fails to create vertical separation from my 4S. Essentially, we have a pitch that, horizontally, does the opposite of what I want and, vertically, doesn’t do anything different from my 4S FB (which, again, is pretty much the opposite of what I want).

So, the question needs to be asked: is this a bad pitch?

Well, this pitch was only thrown at submaximal intent. At face value it sucks. After seeing what happened to the 4S FB as intent changed, however, we can’t be sure of anything. Mechanics might change. The ball might come out of my hand differently. It could completely change the pitch characteristics of a ball held with the same 2S grip.

There’s no way of knowing until it’s thrown hard. If, by chance, I were to come back one week later and throw this pitch at max intent with a Rapsodo behind the plate, and the pitch is faster but still terrible, then we can decide to can it completely or find a grip that works.

Lastly, let’s take a look at the cutter, but not without some background info first. This is my primary pitch behind my 4S FB. I happen to think it’s pretty decent because it gets a lot of swings and misses and soft contact barrels, and as a result: outs. I enjoy throwing it. I’ve always called it a cutter, but I’ve had catchers actually get irritated with me because I call it a cutter. They see it as a slider, and to be honest, I do too. I’m just too lazy to rebrand the pitch. There are some things that pass the eyeball test. The shape of this pitch to the naked eye is a hard, laterally sweeping pitch with less drop to it than horizontal movement.

You can see that’s not what happens during video of one of the cutters from that bullpen:

And here's what that pitch looks like on paper: 

This pitch is very cutterish/sliderish, staggering the line between both pitches. It has zero added horizontal movement, and very little vertical lift against gravity.

*to clarify, the amount of vertical break as well spin axis in this pitch is consistent with a cutter profile, while the amount of horizontal break and spin efficiency is typically more consistent with a gyroscopic type of slider.*

Not what I’m used to seeing out of this pitch.

One thing I noticed is a low spin rate of roughly 2100 RPM for a cutter/slider. At max intent, I would guess that spin rate would increase significantly. If I were to make another guess, I would say that the spin efficiency would drop to somewhere between 15-25% at max intent. It is plausible to see myself getting under the ball with my fingers at low intent compared to getting around the ball with my fingers at max intent. This change would diminish the backspin more natural to a cutter and create more low-efficiency gyrospin that’s more typical of a slider profile. We might even see negative horizontal break. These are the things that I believe would create the shape of the pitch that is described to me through catcher feedback and what I’m used to seeing with my mind’s eye.

*knowing what I know now, a lot of the paragraph above is just wrong. While the raw spin would probably increase with intent, I would say that spin efficiency would actually increase dramatically in order to create more of a laterally-sweeping pitch with very little vertical break. In order for this to occur, the spin axis would have to change to somewhere between 8:45 and 9:15, drastically different from the 12:20ish axis shown.

It is reasonable to believe, however, that my eyes had actually been deceiving me my whole career and my max intent slider was indeed more of a gyroscopic slider with little to no amount of vertical or horizontal break, in which case my assumptions would have been unintentionally correct. I did always have a low 3/4 arm slot before starting this rehab, so this is very possible in retrospect. Yet another example of what we don't know if we're not throwing at max intent.*

The huge issue with what I’m describing, however, is that I really have no idea. Not without a higher intent level. I’m 100% talking out of my rear end. Here, the eyeball test actually helps me discern the differences between what I’m creating with submaximal intent to what I would be seeing at high intent. Here, I know that 75% isn’t helping me. Sometimes we need to play devil’s advocate and remove our tech biases. That’s okay.

To wrap this up, it is my opinion that a pitcher will have a hard time drawing definitive conclusions from their pitch arsenal if they’re using data from submax intent bullpens as their baseline. I believe there can be some exceptions, however. If there happens to be a guy that has submax data compared with data gathered from max intent pens, and he finds that pitch characteristics increase and decrease in somewhat linear fashion with his intent level, then that guy is probably okay. I would have to guess that he’s one of few, though. Maybe a guy finds that this happens with only one of his pitches. He would probably find value in a fine-tune day in front of his Rapsodo. Someone that has baseline data for both submax and max intent pens could even figure out exactly what pitch characteristics will change as RPE decreases. At that point, he could be able to decide what numbers to concern himself with or even disregard as they inevitably vary. But for most of us, including me, there is just too much we don’t know about our pitches unless we have the data on it.

If you’re wanting baseline data or retest data that you can take with you on the mound during a game, why wouldn’t you use data from throws at game-speed?