Pitch Speed and Quality of Contact

I don’t spend a lot of time thinking about the Marlins, but of the time I do spend thinking about the Marlins, the bulk of it is spent thinking about that one home run Giancarlo Stanton blasted off Jamie Moyer. You know the one — the one that broke the scoreboard, the one that set an ESPN Home Run Tracker speed-off-bat record. The home run, just fair, was also a grand slam, and it was the hardest-hit home run baseball’s seen in at least seven years. Stanton didn’t catch up to a blazing heater. This was Jamie Moyer, after all. Stanton ripped a full-count changeup at 72 miles per hour.

There exists some sentiment that harder throwers will give up harder contact than softer throwers. Less contact, of course, but also harder contact. The idea is that Aroldis Chapman‘s balls in play will be struck harder than Livan Hernandez‘s balls in play. It’s based in very simple physics: we care about the velocity of the ball, the mass of the ball, the velocity of the bat, and the mass of the bat. Hold everything else constant and, as you increase the ball’s velocity, you end up with a greater speed off the bat. The reality is more complicated than this, but this gets to the core.

Yet this is a tricky thing to observe. We don’t see harder throwers posting unusually high home run/fly ball rates. We don’t see faster pitches leading to elevated BABIPs. I got curious about something that I’m sure others have already investigated, but I wanted to look at the slowest pitches and the fastest pitches hit for home runs during the PITCHf/x Era. President Appelman gifted me with a log of every home run hit over the last five years. Included was the corresponding PITCHf/x pitch speed, and I isolated the top and bottom 150. 150 was basically an arbitrary cutoff. Here’s a graph showing the difference between the two groups of dingers:

hrpitchspeed

Group 1 includes homers hit off pitches with an average speed of 66.0 miles per hour, +/- 2.7. Group 2 includes homers hit off pitches with an average speed of 97.7 miles per hour, +/- 0.8. Obviously, these two groups are significantly different — the average pitch in Group 2 was 48% faster than the average pitch in Group 1. Many of the pitches we see in Group 1 happen to have been thrown by Tim Wakefield. As a matter of fact, just under half of the pitches we see in Group 1 were thrown by Tim Wakefield.

Dingers identified and isolated, I then tracked them all down on the ESPN Home Run Tracker and made note of their listed speeds off the bat. This is more or less why I used a 150 cutoff — a bigger sample meant more manual labor. I didn’t go in with any expectations, and I was curious what my resulting graph would show. Here is the resulting graph, showing speed off bat, broken down into the two groups:

hrspeedoffbat

There’s nothing there to be found. The average speed off bat for a homer in Group 1 is 103.2 miles per hour, +/- 4.7. The average speed off bat for a homer in Group 2 is 103.3 miles per hour, +/- 4.3. Last season, the league-average speed off bat for a homer was 103.4 miles per hour. That hasn’t really changed since 2008.

The results are virtually identical. The slowest pitches hit for home runs came off the bat the same as the fastest pitches hit for home runs, at least in terms of speed. I didn’t check distance or launch angle, because I was most interested in the one number I noted. The fastest pitch hit for a home run was hit by Tyler Greene off Andrew Cashner in May 2012, and it left the bat at a recorded 101.2 miles per hour. The slowest pitch hit for a home run was hit by Ike Davis off Vicente Padilla in July 2010, and it left the bat at a recorded 102.7 miles per hour. The Cashner pitch was 47.4 miles per hour faster than the Padilla pitch.

Because I’m not a physicist, I can’t talk a whole lot about these results. It’s long been clear that bat speed makes a substantially bigger difference than pitch speed, simply because a bat is a lot more massive than a baseball. Yet pitch speed should matter, if you hold everything else constant. So one wonders if everything is truly being held constant. The faster the pitch, the less time the hitter has to react, so the less time the hitter has to get his bat to peak speed. What we might be seeing is reduced bat speed offsetting increased pitch speed. We know it’s more difficult to make contact with a faster pitch; we might infer that it’s also more difficult to make good contact with a faster pitch, perhaps because bat speed is lower. It seems intuitive to me, but I haven’t been a scientist for years. These days I’m just a guy who’s writing this blog post.

The slowest pitches hit for home runs have been hit just as well as the fastest pitches hit for home runs. There are various scientific concerns, and sample-size concerns, and selectivity concerns as well, since I’m selecting for home runs instead of looking at all balls in play. But Aroldis Chapman shouldn’t be considered a threat to allow a home run because he throws 100 miles per hour. Aroldis Chapman should be considered a threat to allow a home run because he’s a pitcher, and sometimes pitchers allow home runs.



Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


Sort by:   newest | oldest | most voted
cass
Guest
cass
3 years 3 months ago

This isn’t going to be too helpful, but I did read an article in the last couple of years examining the physics of this question and it came to the conclusion that the pitch velocity has virtually no effect on the velocity of the ball as it comess off the bat. Tt’s nice to see the physics angle backed up with empirical data, though.

Kevin
Guest
Kevin
3 years 3 months ago

Can you post a link?

cass
Guest
cass
3 years 3 months ago

Unfortunately, I wasn’t able to find the article. I was hoping someone else know what I was referring to. And, indeed, there is at least one article about the physics linked in the comments below.

AL East Martyr
Guest
AL East Martyr
3 years 3 months ago

The Jeff Sullivan Offseason Spectacular continues… Bravo, sir

Baltar
Guest
Baltar
3 years 3 months ago

Agreed. Another great article.

BostonBrahmin
Guest
BostonBrahmin
3 years 3 months ago

I’d have to think the difference is in the quality of contact. A hitter is probably more likely to hit a Jamie Moyer pitch on the screws than they are an Aroldis heater.

Isolating speed off the bat for just HRs is a bit misleading as well, as the fact that its a HR means it was well-hit. All fly balls would be much more conclusive. How many fly balls hit off Moyer die at the track when a few extra MPH might have made the difference in putting it over the fence?

Art
Guest
Art
3 years 3 months ago

I wonder if the fact that so many of group 1’s pitches are knuckleballs affects the study. Perhaps you should look at the results if you examined pitches which had an average speed of 75 mph or so. That way you’d be looking at more common breaking balls and would thus diversify the number of pitchers who were included.

Roto Wizard
Member
Roto Wizard
3 years 3 months ago

Probably not. A slow, straight pitch at 65 is the same as one at 75. Which is essentially what a knuckle-ball that doesn’t break is. Psychics is psychics my friend. The ball has speed, the bat has speed. Those two relate to produce a distance traveled.

Psychic
Guest
Psychic
3 years 3 months ago

I see 10th grade English class in your future

Kevin C
Guest
Kevin C
3 years 3 months ago

Spin of ball should also have an effect on distance traveled. While speed off of the bat probably has nothing to do with ball speed after contact, it surely must play some part in how far the ball travels (aka will it fly further, and therefore be a HR if it’s a knuckle as compared to if it’s a curveball hit at the exact same speed and spot on the bat?).

While maybe it means nothing, I think it’s worth looking into.

Choo
Member
3 years 3 months ago

I also wonder how collision/compression plays a role – the ball meeting the bat. There may be a near-maximum compression rate for the ball and bat that, once reached, doesn’t really change much in terms of measurable velocity.

Neil
Guest
Neil
3 years 3 months ago

Never change, Jeff.

Don Baylor
Guest
Don Baylor
3 years 3 months ago

That homerun is a perfect example of how hitters may be able, in certain circumstances, to generate greater bat speed on pitches that are slower. Bat speed depends on the level of separation between the hips and the shoulders. The greater the separation at the start of the swing, the greater potential for quicker torso rotation and bat speed.

Hitters are taught to keep their hands back to give them a chance against off-speed stuff. If a hitter is out-front, their hips may start to turn while they are still delaying their shoulders from rotating and increasing the separation. Then, the hitter has greater separation when their shoulders start to rotate and generate greater bat speed than on faster pitches where they may not have generated as much separation.

I’m obviously not using any physics here, but the idea of separation is a big deal to all high level hitters and does effect bat speed.

Choo
Member
3 years 3 months ago

This topic always reminds me of Jeromy Burnitz. Huge separation. He also struck out a lot because of it, but watching (and hearing) him connect was always phenomenal to witness.

prs130
Guest
prs130
3 years 3 months ago

Agree with the focus of this comment. Yes, what the batter is doing matters. When a guy is throwing hard, coaches will tell you to shorten your swing and “let the pitcher do the work.” Is there a chance that hitter consciously shortens their swing (and therefore lose bat speed) when they are facing a guy with a 97.7 mph fastball? Yes. Is there a chance that a guy will lengthen his swing against a guy throwing 72? Again yes. This could potentially skew the results, unless there is another more dominant factor.

Antonio bananas
Guest
Antonio bananas
3 years 3 months ago

I agree with the guy that said only using HR is skewing this. There is a minimum speed a ball must come off the bat to be a home run, but there is also a human maximum, so the results have an artificially narrow band. Maybe just try to look at all contact if that’s possible.

Bob
Guest
Bob
3 years 3 months ago

What if you determined the difference in batted ball speed for individual hitters versus fast and slow pitchers? If more homers off of slow pitchers are being hit by players with higher bat speed, this would make the ball speed seem higher than if it was based on an average bat speed group. So, maybe the faster swingers make up a greater portion of the slow pitching home runs and the slower swingers are a larger part of the faster pitching group and you’d expect the average ball speed to even out.

MGL
Guest
MGL
3 years 3 months ago

I echo the sentiments of Bananas and Brahmin. My first thought was that this was similar to posing the question, “Let’s look at the difference in speed between balls that were hit with a speed of at least 100 mph and balls that were hit with a speed of at least 100 mph.”

A ball that is a HR is already somewhat of a proxy for batted ball speed.

You (Jeff) might want to look at the difference in the pools of batters for the two groups. It could be that the average batter in group I is bigger and stronger than the average batter in Group II, and thus has a higher average bat speed and a higher average bat speed in those HR.

It might also help to look at the locations of the HR’s in Groups I and II. It could be (and likely is) that more of the HR in Group I (slower pitches) are pulled. Pulled batted balls tend to be hit harder for various reasons, including the fact that they are hit later in the swing when the bat speed is greater.

Baltar
Guest
Baltar
3 years 3 months ago

Somebody somewhere who has time to burn might make one of these extensive studies. Jeff’s little study at least discourages making effort in that particular direction. There are other more promising studies to do.

Nevin Brown
Guest
Nevin Brown
3 years 3 months ago

One issue I have with this is that curve balls, because of the way it spins have increased backspin when hit, causing the speed off the bat to be greater. And since many curveballs would fall into the slow category it skews the data. A better way to look at it would be seeing the difference of distance among one pitch so their is less variance based upon the rotation of the pitch. Also I did a physics project in high school and I’m basing it off what I learn then, so I might be totally wrong.

Røark
Guest
Røark
3 years 3 months ago

To second these sentiments, there (I believe) is evidence to suggest that the backspin a ball generates when hit at the angle that would induce a homerun, runs in concert with the spin already provided with a curveball. It could be the case that the change of spin the ball endurs to move at a home run speed and angle negatively affects the way a ball would fly off a fastball versus a curveball. Would a pitch of the same speed, squared up the same way, with bats of equal speed and mass, under the same conditions, fly the same speed if it were from a fastball or curveball? Unlikely.

It would be lovely to see this study done with the same type of pitches. And further, it would fantastic if there were sample sizes large enough to show fastballs/curveballs/knuckleballs/etc. all hit for homeruns at the same speed.

Ryan Gosling
Guest
Ryan Gosling
3 years 3 months ago

I could be wrong, but I believe there is evidence to show that spin of the pitch doesn’t change the spin of the ball off the bat…

Alan Nathan
Guest
3 years 3 months ago

Nice article. Some comments on the article and on some of the other comments.
First, I recommend taking a quick look at an article I wrote a few years ago for Baseball Analysts: http://baseballanalysts.com/archives/2010/01/comparing_the_p.php. You will find a formula there relating batted ball speed to pitch speed and bat speed: BBS=0.2*pitch_speed+1.2*bat_speed, for a typical wood bat at the sweet spot. Note the big difference between the 1.2 and the 0.2 factors, showing that indeed bat speed matters much more than pitch speed. However, the reason *is not* because the bat is so much heavier than the ball (contrary to Jeff’s assertion). The reason is primarily because the bat is already moving in the same direction as the outgoing ball. That factor alone accounts for the extra “1” (i.e., 1.2 rather than 0.2). Now, given these numbers, I would expect a difference in BBS of 0.2*(98-66)=6 mph for the two groups, all other things equal. That you do not see such a difference suggests that all other things are not equal, as some have speculated in the comments. Roark poses this question: “Would a pitch of the same speed, squared up the same way, with bats of equal speed and mass, under the same conditions, fly the same speed if it were from a fastball or curveball?” The answer to this question is almost surely yes, unlike Roark’s answer. The difference in spin between a fastball (backspin) and curveball (topspin) has little or no affect on the speed of the batted ball. I tend to think that a batter is out in front a bit more on a slower pitch, so that his bat has had more time to accelerate and is therefore moving faster (as someone upthread has already proposed). But that is just speculation on my part.

I have some data on batted balls that might possible shed some light on this. Namely, the April 2009 HITf/x data (which were made publicly available) and some “bootlegged” TrackMan data from 2009 and 2010. It has the advantage that all batted balls are in the data set. I will take a look over the weekend and post something here if I learn anything interesting.

stathead
Guest
stathead
3 years 3 months ago

Hey look! An error bar!

LK
Guest
LK
3 years 3 months ago

I think it would be more interesting if the study were limited to just fastballs to isolate the effect of speed.

Bip
Member
Member
Bip
3 years 3 months ago

The high-velocity group in this study would remain unchanged and the low velocity group would be relabeled to read “home runs hit off of Jamie Moyer’s fastballs.”

Brent
Guest
Brent
3 years 3 months ago

I’d be curious to compare the distance of home runs on fast v. slow pitches, not just the speed off the bat. There won’t be any PITCHf/x data for it, but I still remember the one McGwire hit off Randy Johnson back in the day. It was like “here’s the heater, right down the pipe, see if you can hit it.” It went all the way to the back wall of the Kingdome in left-center.

Bip
Member
Member
Bip
3 years 3 months ago

I really don’t get the point of comparing home runs to home runs. There could be a significant effect of ball speed on speed off the bat, but I would still expect this study to show the same results as it did.

Let’s assume that a faster pitch actually does come off the bat faster. Let’s go even farther and assume that bat speed does not diminish on a faster pitch, it just causes the bat to miss the mark more often. If this is true, then we would expect a squared-up home-run-power swing on a fast pitch to go farther than the same swing on a slow pitch. This would suggest that homers off fast pitches should be faster. However, what we would also see is that some not-quite-squared-up home run swings and some squared-up not-quite-as-powerful swings that are right on the border, such that those swings produce barely-out homers on fast pitches and warning track fly balls on slow pitches. This would bring down the average speed of the fast pitch sample and not affect the slow pitch sample.

In summary, the author acknowledges the selection bias of his sample, but in this case the selection bias basically determines the results and makes the study redundant. Perhaps we can instead look at the average distance or speed-off-bat of balls given constant trajectory and pitch speed, or look at what percentage of those balls are homers, controlling for those other factors. Maybe fast pitches result in harder hit balls, but they also are less likely to result in balls hit at the right trajectory to be home runs.

Alan Nathan
Guest
3 years 3 months ago

In thinking about this problem a bit more, I am becoming convinced that there is a selection bias in considering only home runs. Here’s why. Suppose you take seriously my estimate that the two groups should have a difference in BBS of about 6 mph. That is a really huge number, since it corresponds to about 30 ft in distance for a typical fly ball on a home run trajectory. Many of those fly balls will not get over the fence, except if the bat speed is higher. That tends to even out the differences in batted ball speed between the two groups.

I have started to look at the HITf/x data from April 2009. Still have more work to do but the initial look shows a definite correlation between BBS and pitch speed. More later.

Alan Nathan
Guest
3 years 3 months ago

Sorry for that last post, which I wrote before reading the one by Bip just preceding it and saying just about the same thing.

Christian
Guest
Christian
3 years 3 months ago

Somebody get me Robert Adair, stat!

Alan Nathan
Guest
3 years 3 months ago

For those of you who have Adair’s book, look at Fig. 2.5 and 5.9 to get the change in batted ball speed for each mph of pitch speed (3rd ed). You need to do a bit of algebra, but when you do so you find a value of 0.13, whereas I use a value of 0.20. But, just because Adair writes it doesn’t make it correct. My number is based on experimental data taken by impacting bats with baseballs projected from a high-speed air cannon. Adair did not have the benefit of those experiments when writing the book. For impacts that are “squared up” and at the sweet spot of the bat (about 5″ in from the tip of the barrel), the value of 0.20 is pretty representative of 34″/32 oz bats. Smaller (but not a lot smaller) for shorter/lighter bats.

adohaj
Guest
adohaj
3 years 3 months ago

These days I’m just a guy who’s writing this blog post.

respect +1

Alan Nathan
Guest
3 years 3 months ago

I have taken a look at the HITf/x data from April 2009. At the time, the data were publicly available. Here is a link to a plot showing some of the results (sorry, I don’t know how to post an image directly here):

.

I looked only at line drives, which I defined to be hits with a vertical launch angle between 0 and 20 degrees. The plot shows two histograms. The blue one shows the distribution of batted ball speeds for pitch speeds between 85 and 90 mph, where I have used the pitch speed at home plate rather than the release speed (it is the speed at contact that matters for the ball-bat collision). The red one shows the same thing for pitch speed between 65 and 70 mph. I have normalized the two histograms so that they have the same number of events, for purposes of displaying them on the same plot. It is quite clear that the two look nearly identical with the blue one shifted about 4 mph higher than the red one. And 4 mph is the expected shift based on “q~0.2” (see my first post in this thread). So, it all seems to work out pretty well. The mean values are shifted by a bit more than 4 mph, but these are weighted by the low tail, which might be a bit different. The peak values are much closer to the expected ~4 mph shift. Based on this result, I conclude that there must be a selection bias (as discussed upthread) doing an analysis using only home runs.

Alan Nathan
Guest
3 years 3 months ago

Sorry…the label on the x axis should be “Batted Ball Speed (mph)”

mylrunner
Guest
mylrunner
3 years 3 months ago

Pitch speed has little to do with it. Think of extremes, you can hit a ball off a tee, hard and pretty far (that’s zero velocity for the ball sitting on the tee). Now consider a zero velocity bat and throw a ball at it (similar to a bunt), it won’t go very far. HR derby shows pitch velocity has very little to do with it either. It’s about the bat speed and mass.

Alan Nathan
Guest
Alan Nathan
3 years 3 months ago

I can’t tell whether mylrunner is agreeing with me or not. So let me parse what he posted and comment on each part:

“Pitch speed has little to do with it.” Agreed. Little, but definitely not zero.

“Think of extremes, you can hit a ball off a tee, hard and pretty far (that’s zero velocity for the ball sitting on the tee). Now consider a zero velocity bat and throw a ball at it (similar to a bunt), it won’t go very far.” Excellent example and one that I often use in talks that I give, most recently last Friday night!

“HR derby shows pitch velocity has very little to do with it either.” Not sure I agree with you on that one. HR Derby is all about getting a high percentage of HR per swing. That requires more than just raw bat speed and pitch speed. It also requires good contact. So, the pitch is thrown slower and in a more predictable manner than in a game situation, allowing the batter to “load up” on his swing without having to worry about misjudging the speed or movement of the pitch. I suspect therefore that batters in HR Derby get higher bat speed than in game situation. So, the lower pitch speed is more than compensated by higher bat speed, leading to lots of HR. In that regard, I once did an experiment using high-speed video to measure the bat speed of elite slow-pitch softball players. They achieved far higher bat speed than is typical in MLB. Once again, the pitch is slow and highly predictable, so the batter can get a higher bat speed.

“It’s about the bat speed and mass.” Yes, those things matter, as does the COR of the ball. But pitch speed matters too, albeit much less than bat speed.

Given that the HITf/x data do show that batted balls are hit from slower pitches at lower speed than off fast pitches (as expected, all other things the same), the real question is why the HR data do not show the same thing. Upthread, I gave a possible reason. There may be others.

Alan Nathan Fan
Guest
Alan Nathan Fan
3 years 3 months ago

Alan Nathan is my new favorite writer on fangraphs..
Thanks

Hurtlockertwo
Guest
Hurtlockertwo
3 years 3 months ago

I used to be able to hit the ball over the left field fence (maybe 320 feet)
just hitting flyballs to the outfielders for practice. (this used to be one of the extra pitchers jobs when you were not throwing) I’m not a real big guy, 5’11”, 185 lbs. Regulation sized bat.

jerry
Guest
jerry
3 years 21 days ago

i hate yall

jerry
Guest
3 years 21 days ago

i hate all yall

wpDiscuz