The Knuckleball and Home Runs

There’s a sense that the knuckleball is more prone to the home run. The guys on the MLB Network, while watching R.A. Dickey give up three home runs to the Reds, thought out loud that the lack of spin meant that it would go further upon contact and lead to more bombs. The home run seems to be the source of some of the ‘risky’ label attached to pitchers that use the knuckler. Even physics professor Porter Johnson said in a recent interview that if a knuckleball “doesn’t move, it’s basically a home run.”

As with all conventional wisdom, this link is worth unpacking.

First of all, does the population of knuckleball pitchers give up any more home runs than the regular population? On first glance, it’s tempting to say yes — after all, Dennis Springer, Steve Sparks, Tom Candiotti, Tim Wakefield and Phil Niekro have all had seasons in which they gave up more than a homer and a half per game. That’s extreme. Then again, some of them pitched in the homer-happiest of times: the late 90s. In 1996, just to pick a year out of the hat, pitchers gave up 1.64 home runs per nine. In 1917, pitchers game up .04 homers per nine.

If we want to compare Springer’s work with that of, say, Eddie Cicotte, who pitched in the late teens, we’ll have to index each knuckleballer’s seasonal home run rate to the league rate. The resulting number, a HR/9+ if you will, can help us compare knuckleballers to the rest of the league. Using this custom list paired with the league stats on this site, that’s easy enough. We’ll use 70 innings as the cutoff, so that we get the Hoyt Wilhelm closer seasons in there alongside the starters.

The average indexed home run rate for all seasons by knuckleball pitchers is 99.06 (100 is league average). Since 2002, the average home runs per fly ball for knuckleball pitchers is 9.85%, and the league-wide number is between 9 and 10% annually. Your garden variety knuckleball pitcher does not give up more home runs than the league average pitcher.

But are there knuckleballs that are more likely to be hit for home runs? The truism that comes to mind is “If it’s high, let it fly.”

Here’s a graph of R.A. Dickey‘s home runs since the 2010 season began, thanks to Joe Lefkowitz’s PITCHf/x site:

You could read this graph a couple ways. In the binary high/low way, it seems to suggest that high knucklers are no more likely to be hit for homers than low ones. After all, of the 48 dots represented on the graph, 24 are at 2.5 feet or higher. But if you look at the strike zone in thirds, you do see that the bottom third has fewer home runs in it than the other two-thirds.

Of course it does. Dave Allen famously graphed the representation of ground balls versus swinging strikes based on placement in the zone:

As you can see, that bottom third of the strikezone (represented by the leftmost space between the two red lines) is where the ground balls live. So, yes, if the knuckleball is high, let it fly, but only because all high pitches are more susceptible to fly balls and therefore home runs.

So what about this spin idea? Maybe high knuckleballs also lack spin — after all, we know spin and movement are intertwined — and are therefore even more likely than a high fastball to end up deposited in the seats?

One thing we probably know for sure is that it’s not the spin itself that’s leading to the home runs. Physics professor Alan Nathan showed in an excellent piece of research that the “spin of a batted baseball is less dependent on the spin of the pitched baseball than previously thought.” In other words, the spin of the baseball coming into the bat does not determine much of the spin on the way out.

But there’s still the possibility that the knuckler with no spin gets hit hard. After all, Dickey has said that he wants a quarter-turn from his knuckleball, so less would be bad. Less spin should mean less movement, which should mean a straighter, possibly higher pitch, and definitely means an easier pitch to hit for a home run. “When it was there, it was filthy,” Dickey said to reporters according to Andrew McCullough at the Star-Ledger following that loss to the Reds. “And when it wasn’t, it kind of tumbled up there.”

So, in other words, a bad knuckleball gets hit for home runs. That’s true of a lot of different types of pitches.

Print This Post

Graphs: Baseball, Roto, Beer, brats (OK, no graphs for that...yet), repeat. Follow him on Twitter @enosarris.

16 Responses to “The Knuckleball and Home Runs”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Bryce says:

    I thought he wanted a quarter turn because no spin moves too much. As he told Letterman, knuckleballs with a full turn get hit out of the park.

    Vote -1 Vote +1

    • Eno Sarris says:

      Is true that he said that now that I think about it. Even linked to that clip I think. I think too much spin on the knuckler just means more predictable movement… maybe I applied regular pitch philosophy to the knuckler. Either way, not much of a link between these things and home runs.

      Vote -1 Vote +1

  2. Alan Nathan says:

    As you may know, John Walsh wrote a very nice article for Hardball Times in November 2007 ( in which he looked at how various batting statistics depends on the *movement* of the k-ball (as opposed to the location). The analysis was based on 1 yr of Wakefield data. Here are John’s results:

    Break | NP | Hit% | HR% | BABIP | OPS |
    | Small | 47 | 0.383 | 0.021 | 0.370 | 0.979 |
    | Medium | 71 | 0.338 | 0.028 | 0.319 | 0.873 |
    | Large | 79 | 0.253 | 0.025 | 0.234 | 0.684 |

    Note that HR% is not strongly correlated with movement (“Break”) but the other quantities are. With nearly 5 more years of Wake and RA data available, this analysis is ripe for a new investigation.

    Vote -1 Vote +1

    • garik16 says:

      That article does need to be relooked at in terms of Dickey – for example, iirc when I looked at it 2 years ago, Dickey doesn’t get in the “Large” area that often, mostly in Medium. When Dickey does get in the large area, he often suffers because the ball is often out of the zone.

      Vote -1 Vote +1

  3. Paul Clarke says:

    After all, Dickey has said that he wants a quarter-turn from his knuckleball, so less would be bad. Less spin should mean less movement

    How much difference in movement could a quarter-turn cause? If my mental arithmetic is working that’s about 1/50 of the spin on a typical fastball, which “rises” about 8 inches. Is movement proportional to spin rate? If so then the movement from the quarter-turn would be about 0.2 inches, much less than the movement from other sources.

    Vote -1 Vote +1

    • Alan Nathan says:

      For “ordinary” pitches, movement is more or less proportional to spin rate, due to the so-called Magnus force. But the mechanism that causes the knuckleball to move is very different and it has to do with the character of the air flow changing as the air passes over the seams. The resulting forces on the ball depend on the seam pattern, which is not very symmetric for a baseball. Wind tunnel experiments have measured those forces. If you use the wind tunnel results to simulate a pitched ball, you can get big differences in the trajectory with a small change in the orientation of the seams upon release or in the amount of rotation. The effects get very small if the ball is rotating too rapidly (meaning, less movement). Of course, if rapidly rotating, then the Magnus force kicks in.

      Vote -1 Vote +1

  4. JimNYC says:

    I’m fascinated to learn that Phil Neikro pitched in the late ’90’s. I had no idea that that was a thing that happened.

    Vote -1 Vote +1

  5. Fatbot says:

    Trying to use RA Dickey to make conclusions overall about the knuckleball pitch is flawed because Dickey is an enigma. He’s pitching in the biggest K% era in history. Some people believe that is because pitchers are better (e.g. relievers specialized), but a just as plausible explanation is because hitters of this era are the worst in history at making contact (they swing for the fence or nothing, especially seen with the universal removal of the “two-strike approach” from baseball).

    So basically Dickey’s knuckleball appears better than any knuckleball in history because the batters trying to hit it are awful at trying to make contact. The flawed hitters of today are really exposed against movement, they whiff at stuff outside the zone. That is why if you look at all starters over the years since pitchFx data has been available, the percent that they throw fastballs has fallen from 62% down to 55% — why throw a fastball to today’s hitter when they go fishing for junk? This is hugely magnified against Dickey.

    The Dickey fans try to just conclude he’s great, but any explanation of why Dickey should be better than any other knuckleballer in history doesn’t make sense. The argument he “throws it faster” makes no sense because faster means less movement which means it should be crushed, and other knuckleballers have thrown it as fast. The argument that he has better control with it doesn’t make sense either, as everything great about the knuckle comes from its unpredictability, so if he’s “controlling” it, it’s removing the unpredictability, and therefore it’s just a pitch with movement, not the special gimmick that defines a knuckle.

    Basically everyone should just throw knuckleballs now, because it works! Eri Yoshida has only given up 6 HR in 73 IP!

    Vote -1 Vote +1

    • Alan Nathan says:

      The less movement on the higher speed knuckleball is true. However, the batter also has less time to respond to the movement at higher speed. So, which effect “wins” as far as the batter is concerned? The answer is not obvious to me.

      Vote -1 Vote +1

    • garik16 says:

      “but a just as plausible explanation is because hitters of this era are the worst in history at making contact (they swing for the fence or nothing, especially seen with the universal removal of the “two-strike approach” from baseball).”

      No? This is a gross oversimplification of the current era. This is far from a deadball era. Yes, Ks have gone up, because hitters have realized that effective contact more than can make up for increased strike out rates. That doesn’t make hitters the “worst at making contact” – it means they’re trading pure contact ability for the ability to make BETTER contact.

      ” The argument he “throws it faster” makes no sense because faster means less movement which means it should be crushed, and other knuckleballers have thrown it as fast.”

      Again, an oversimplification. Less movement doesn’t necessarily mean the pitch should be crushed as long as the movement is unpredictable, and higher velocity more than makes up for this by reducing batter reaction time. In addition, higher velocity has been shown to increase GB rates on EVERY type of pitch, thus reducing “crushing”.

      Vote -1 Vote +1

    • BlackOps says:

      I’m seeing a lot of research as to why Dickey is having one of the most successful seasons by a knuckleballer in history and very little done on how hitters are the worst in history on making contact.

      There isn’t much data here. Yes, batters are swinging outside the zone more, but are making as much contact as they did ten years ago.,ss&rost=0&age=0&filter=&players=0

      And to add to garik’s point about higher velocity inducing more ground balls, fastball velocity has gone up or stayed the same every year but one since 2002 (89.9 in 2002, 91.5 in 2012.) Despite batters making the necessary adjustments, league-wide GB% is still at its highest rate ever, by almost a percentage point. Another statistic showing that batters aren’t simply worse at making contact but making more selective contact is the fact that HR/FB% is the highest it’s ever been.

      Again, only ten years of data, but it would be disingenuous to compare this era to any other. The information we have now is changing the way the game is played quickly.

      Vote -1 Vote +1

  6. Baltar says:

    Yeah! That’s what I’m talkin’ about.
    Great hard data post with great hard data comments!
    Another “common sense” notion kicked into the trash.
    Keep ‘em comin’, FanGraphs.

    Vote -1 Vote +1

  7. John C. says:

    A problem for the knuckleball is that a poorly executed knuckler is pretty defenseless because it (arguably other than R.A. Dickey’s “fast” knuckler) just floats across the plate. Or, as Jim Bouton observed in his book Ball Four about one such knuckler: “float like a watermelon, fly like a rocket.”

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>