Slashdot Mirror
The Physics of the Knuckleball
snoop.daub writes "R.A. Dickey, pitcher for the New York Mets, has been in the news this week after two dominant pitching performances in a row, holding opponents to one hit in each of the games for the first time since Dave Stieb did it in 1988. He has taken over as the league's only knuckleball pitcher after Tim Wakefield retired last season. But just what is it about the knuckleball that makes it hard to hit? Conventional wisdom has it that the lack of spin on the knuckleball causes it to move in completely unpredictable ways, even changing directions in mid-flight. In the last few years, there has been a lot of good science done to understand baseball pitch trajectories, and a few months ago Prof. Alan M. Nathan showed that knuckleballs aren't really so different from other pitches. It turns out that the same 9-parameter equation that can be used to describe other pitch trajectories applies just as well to the knuckleball. The difference appears to be that, like in a chaotic system, knuckleballs depend sensitively on the initial conditions, so that small changes can cause randomly different forces at the start of the pitch which determine the resultant trajectory. Much of this and similar work depends on the Pitchf/x tool, which has recorded the complete trajectory, spin angle and spin rate of every MLB pitch since 2007! Baseball really does have the best sports stats geeks."
inb4 ew sports.
Guy 1: We have amazing technology that allows us to know EXACTLY what happened.
Guy 2: Awesome, so we don't have to rely on humans in those really close calls.
Guy 1: Well...not really...
Guy 2: ?
Guy 1: We're only going to use it to record pitches...
WhatMeWorry
Back in the late 80's/early 90's the Mets had a very successful pitcher named David Cone, and his fans were known as Coneheads... and sometimes dressed the part.
Now for R.A. Dickey.... hm.... I think we'd better focus on his pitching style instead of his name. Let's go with Knuckleheads.
I'm a hockey fan and it's not unusual to see goalies get beat by what seem like simple shots. Someone skates over the blue line into the offensive zone and shoots an average wrist shot towards the goal. It's a routine save for the goalie under normal conditions... a really low percentage shot. But if the shot gets tipped, even ever so slightly and even a long ways away from the goalie, the goalie can have trouble with it.
It's because the goalie reads the shot not by plotting the course of the puck but by seeing so many shots that by the motion of the shooter's stick and body language, he already knows where the shot is going and reacts accordingly. A tip, even a foot away from the shooter's release, turning a 20 foot shot into a 19 foot one, throws it all to the wind. You'd think it would give the goalie enough time to make the save but he's already moving to the top right corner before he realizes is going bottom left.
I'm sure it's the same in baseball. Batters don't have time to judge the ball's trajectory itself so they rely on the pitcher's delivery to tell them where the pitch is going. When a knuckleball comes their way, there's nothing to read because even the pitcher doesn't know where it's going.
What doesn't kill you only delays the inevitable
Baseball really does have the best sports stats geeks.
Meh. You've clearly never met any cricket fans.
Just ask one to describe the Duckworth-Lewis method of calculating scores for a rain-interrupted match.
That's because if you took the stats out of baseball, there'd be nothing left.
The article claims that knuckelballs are really no different than normal pitches, mathematically speaking. My guess is that the author has never tried to hit a knuckleball pitch.
Six years ago, from a professor at my alma mater: http://farside.ph.utexas.edu/teaching/329/lectures/node45.html This being slashdot, I didn't RTFA but the author seems to come to the same conclusion that Fitzpatrick did. Incidentally, if you ever need to know something about physics, chances are this fellow has excellent lecture notes posted on his website covering the topic (in hyperlinked html, pdf, and even a git repository for the latex code!).
Nathan has a conference paper with a lot more detail about the physics. PDF here: http://webusers.npl.illinois.edu/~a-nathan/pob/ProcediaEngineering34KBall.pdf
-Tom Duff
I know this is being picky, but if A "causes" B, that is not a "random" chain of events. A chaotic system may be unpredictable by an observer such as a Major League batter, but it is not in any sense of the word, "random".
"We receive as friendly that which agrees with, we resist with dislike that which opposes us" - Faraday
It's nice to see the pracitioner of a fine skill be successful where traditionally the best pitcher is the one who can throw the fastest (under control of course).
A similar scenario happens in cricker where a great spin bowler can dismantle a team. Until the 1990's bowling in cricket was dominated by extreme speed where the best bowlers could bowl at over 150 Km/Hr. Along comes Shane Warne, considered the 2nd most influential cricketer in the 20th century who bowls at less than 100 Km/Hr but with a wicked spin and fantastic control.
Check out the "Gatting ball" video below for a delivery of pure beauty.
http://www.youtube.com/watch?v=rOVei8iTyM8
It was Warne's first Test Match delivery in England!
Comment removed based on user account deletion
Batters don't have time to judge the ball's trajectory itself so they rely on the pitcher's delivery to tell them where the pitch is going.
Yeesh. You're just plain wrong. At least you were nice about it. But you're just wrong.
Hitters watch the pitcher's release point and try to "pick up" the ball as soon as it leaves his hand. If a hitter doesn't visually lock onto the ball as soon as it leaves the pitcher's hand, he probably won't hit it--at best, he'll foul it off.
Pitchers generally try to maintain a consistent release point; it makes it easier to develop fine control and helps prevent injury. The release point can be the same, yet the pitch location can be all over the place.
I haven't even mentioned spin yet. A four-seam fastball and a two-seam fastball behave very differently, yet they begin the same, coming straight out of the pitcher's hand with backspin. If the hitter doesn't read the spin, he probably won't hit it well. A changeup or a splitter are even more different than those pitches, yet they also come straight out of the pitcher's hand with backspin. The hitter must see the seams of the ball as it's in flight in order to recognize the pitch type and be able to estimate its trajectory. And I haven't even mentioned curveballs and sliders yet.
I'll never forget the first time I recognized a slider while hitting. I remember seeing the dot right after the ball left the pitcher's hand. I had seen enough of them on TV replays while watching games that my mind recognized it quickly, and I knew the pitch would be a ball, low and away. If I hadn't seen the dot, it would have looked like a fastball down the middle, and I would have swung and missed. And all of that visual and mental recognition and processing has to happen in a fraction of a second. It was exciting! (If only I had had my vision corrected years earlier! I didn't realize I was capable of seeing the spin on the ball.)
As for knuckleballs, it's an exaggeration to say, "even the pitcher doesn't know where it's going." Not that a pitcher has fine control over it, but if it were as wild as you suggest, it would be useless. If you can throw a baseball already with decent accuracy, you can try a knuckleball for yourself and see. It's not that hard to get it in the strike zone if you're a decent pitcher.
Yes, I used to play baseball, both pitching and hitting.
(As an aside, while you might know a lot about hockey, please don't speculate so authoritatively about something you don't actually know about. It's a shame to see a post that's just plain wrong modded +5 Insightful.)
"Those who consume the bulk of goods are those who make them. We must never forget this secret of our prosperity."
TFA states that if there is no initial spin on the ball, the curve of its trajectory will be constant. However, if a ball starts with no spin and a seam is oriented to make more air friction than the opposite side of the ball, that will impart a spin to the ball as it travels.
Furthermore, TFA states that the uncertainty of the ball's position is 0.3 inches, and the claim is that they can reliably declare that the knuckleball has an rms unpredictability of 0.45 inches versus other pitches' 0.4 inches (choosing the right side of the displayed graph). The raw measurements need to be much better before claims can be made about trajectories that are buried in noise.
Contribute to civilization: ari.aynrand.org/donate
Baseball really does have the best sports stats geeks.
Well, the author should meet some Cricket fans.
What makes Dickeys knuckleball so good is that he has 5 variations of it. Depending on the grip, the speed is different and one actually goes up. Add in that he has a fastball, while not that fast compared to other pitchers it is serviceable. He can throw the fastball to keep hitters honest and the knuckleball is that much tougher to hit consistently well because of it.
I remember when he was a dominant pitcher with the Blue Jays, back when they won two world series. He was the ace of the staff. Had a great career but had a reputation for being somewhat prickly with the press. Real good player though. It's the first time I've seen his name in a while.
There was a great exchange (usually attributed to Hack Wilson) between a batter and umpire that eloquently describes the umpire's role in calling balls and strikes.
Wilson stepped to the plate and waited on the first pitch. He didn't swing, and the umpire called "strike." Wilson stepped back and said "That's a strike? Boy, you sure missed that one." The umpire didn't miss a beat and replied "I wouldn't have if I had your bat, Hack."
The strike zone is generally described as being the belt to the knees. When Hall of Famer Rickey Henderson broke into the majors, he brought an unorthodox batting stance that minimized the size of that space. (Picture a man standing in the batters box trying to touch his nipples to his knees, while simultaneously trying to scratch his butt by holding the bat over his shoulder.) Henderson's batting stance was so unorthodox and frustrating to some pitches that one major league pitcher threw his first pitch behind Henderson and barked at the umpire to "tell him to stand in there like a man."
The rules of baseball define the strike zone reasonably well. The practical interpretation is another matter...
There are some people that if they don't know, you can't tell 'em.
Since there is only one current pitcher that throws it, I wonder if school bans on this pitch have led largely to its death. I was banned from throwing the knuckleball in Jr High and High School because it was "too hard on a pitcher's arm," even though I used it in sandlot games fairly effectively. Ultimately lack of arm strength did me in as a pitcher (I could throw harder with a submarine pitch, which I had started using in sandlot games, but the school didn't allow that either), and I wasn't much of a batter once pitches hit 60-70MPH (much less 80 like some of our high school pitchers were throwing).
Knuckel-balls are not as simple as "a tube .... that follows a smooth curve." That is a wobbling knuckle-ball, and is generally what people think of. This is usually the first pitch knuckle-ball, does not really move much but you can control it better to get that first strike. But if that is all you have got, then you are going to hit around a bit. That is not the strike-out knuckle-ball. By making minor changes in the grip, you can produce more movement that can cause it to dive, cut-in, and break-away. One of the guys that taught me had a knuckle-ball that he could snake in almost at will, much more movement than the smooth curved tube. If first broke slightly left to appear to be wide of the strike zone but then broke hard right to fall back in for a strike. The problem is that good hitters can anticipate it after seeing it a few times.
A good Knuckle-ball has a slight rotation. Somewhere between half a turn and a turn and a half on it's way to the plate. This slow rotation slightly changes how the seam are presented to the high pressure area in the front thereby changing the disruption of the airflow around the ball. Just like an airfoil will cause low pressure on the top of a wing creating lift and moving the plane up, these changing disruptions cause temporary low pressure areas on the ball and cause a small amount of "lift" in a vectored direction from the center of the mass of the baseball. If these happen rapidly and evenly over the front surface, you get the wobbling knuckle-ball like he describes. If they appear mostly on one side, it will move in that direction. With practice, you can begin to throw the wobbler when you want a strike and a hard breaking knuckler when you want to get them to chase a pitch out of the zone by slightly changing your grip and orientation.
As an aside, and interesting read is The Physics of Baseball by Robert Adair
http://www.amazon.com/The-Physics-Baseball-3rd-Edition/dp/0060084367
(Though, I have to disagree with his opinion on the effect of ball rotation on a batted ball. If I remember correctly, he states that the effect is negligible. IMO and experience, I believe it is noticeable and sometimes determines fair/foul as some batted balls hook much more than others.)
Hint for curious, it involves taking a hair from each player on the field and burning them so that the oracle can divine a prophesy of how the match *would* have gone, and deciding the score accordingly.
(It actually uses detailed Databases of previous performances of teams and players, but for all effective purposes, you might as well burn the hair)
I am a Pakistani fan, and I remember one of our Coaches, Bob Woolmer (may he rest in peace!) who was know as Mr. Laptop, since he was always on the laptop, always calculating, to the ball, what the D/L score would be if it rained at that precise moment. If rain was expected, he would direct play so that D/L would be in their favour, which sometimes involved *slowing* down play. Yeah, D/L is *that* weird.
So yeah, Baseball, you got *nothing* on cricket when it comes to stats. Baseball uses it (AFAIK) for generic comparison purposes as to how the play *might* go, whilst cricket uses it to actually decide how the play *will* go.
Wiki Link: http://en.wikipedia.org/wiki/Duckworth–Lewis_method
I am an ACCA student. Got a query on Accountancy/Finance? Maybe I can help!
As if there wasn't already one statisticians' wet dream on the form of D/L, here comes another stats-gasm inducing system, which called the VJD.
India tried to enforce it internationally, but it was turned down in the recent meeting.
http://en.wikipedia.org/wiki/VJD_System
I am an ACCA student. Got a query on Accountancy/Finance? Maybe I can help!