In your travels through technical papers and wind-tunnel research studies you may run into one or more of the following people: Dr. Joel Hollenberg, Robert K. Adair, Robert G. Watts & Eric Sawyer, Kai Tang, Adam Kleinbaum, and Dane Shellhouse. For information about coefficients of drag, fluid dynamics, and Reynolds numbers, these are the people to hunt down. They have the numbers, the published technical papers, and the aerodynamical jargon all down cold, and they CAN explain the action of the knuckleball in techie terms.

Most of us experience MEGO when we wade into that, however. (MEGO: My Eyes Glaze Over.) And the whopping majority of us find this useless when it comes to actually throwing one or trying to get that strikeout.

And so, I'll boil all that down to language we all can understand.

Objects moving through air produce drag. Air doesn't want to slide along solid surfaces very well, but air readily slides over pockets of air moving in other directions. This is the key to the knuckleball.

The stitches on a baseball act like the air dam underneath the front of your car. The air dam pushes the air aside and forms a swirl ("vortex") of air that moves in all kinds of directions around all the car parts hanging under there. Without the air dam, the air would want to drag across all those parts and add drag.

The stitches on a baseball act to push the air flow away from the leather surface just enough to form a tiny swirl of air just behind them, which most of the air happily slides past. Where the airflow moves across smooth surface, it can't move so fast. Knowing your high school science, you know that air moving across a surface produces a lower pressure than air moving not so fast on the same surface. An airplane wing works that way: Air pushed aside by the curve on top has to move fast to meet up with the air moving along the bottom, so pressure is lower on top and the wing has lift. Low pressure draws the object towards it, so wherever the lowest pressure is on that object from moment to moment, that's where it will want to drift towards. This is known as the Magnus effect. However, since the stitches are so small and acting on such a heavy object, this effect is minimal on a baseball. The airflow across smooth surface at one point and stitches at another is uneven in speed and strength, so it trips and triggers shifts in the formation of the turbulence behind that ball. What most affects the movement of a baseball is the size, shape, and location of the wake behind it, which causes enough drag to determine the path of the ball. Rapid rotation allows this wake to fill, keeping it relatively small, but being established in a particular location behind the ball, it may develop some lift, as on a four-seam fastball, slightly less as on a two-seamer (which doesn't present as many seams to keep this wake as small), or some drag to one side as with a curve ball. This is greatly apparent on a rapidly rotating ball, as you see here:

This ball is rotating rapidly from right to left across the top, and it appears that this is pumping air across the top, driving the wake downward, and filling the space behind the ball with smooth-flowing air. That's an easy way to understand it.

The action of the knuckleball, however, takes into account the fact that some stitches are moving towards the flow of air in front, and others are moving away, at a very slow speed. The fact that the stitches move around the ball in quite a complex curve on a knuckleball and the ball may rotate at different rates in different ways causes these swirls behind the ball to change size and direction, form and disappear, and move location on the ball, producing changing locations and strengths of low pressure that really can't be predicted. The wake behind a knuckleball at various points in flight may look like these:

It's fast rotation that can partially counteract gravity. A hard-thrown fastball rotating front to back over the top produced lift just behind and slightly above the center of the ball, tending to hold it up so gravity doesn't drop it so fast. A ball with little, if any rotation doesn't generate that lift, and it produces a larger wake, so it naturally falls away, maybe a foot or more. This explains the drop of a knuckleball (and similarly, the forkball or split-finger fastball).

The air pressure around the outside of the ball is greatest, so that's where the airflow is tripped and shifted to deform the large wake behind the ball. It's thought by some that this is enough to kick the ball this way and that, but the little amount of pressure there does not explain some balls suddenly "hitting a wall" and diving hard. Wind tunnel views as you see above show that it's the deformation of the low-pressure turbulence behind the ball that knocks it around. Because of the low rotation, the ball already wants to fall away, but if this vortex suddenly swells, it can cause the ball to suddenly brake and cut hard off-line. There is no other way to explain the unpredictable and drastic darting of a five-ounce sphere propelled at speeds of 60 mph or better.

It's also said that the ideal rotation of a knuckleball is one half turn. Others have said one quarter turn. The problem with either of those is this does not take into account the end points that this measurement is taken: release to front edge of plate, or to a point first within reach of the bat, or the catcher's mitt? Also, considering the complexity of the curve of the stitches around the ball, any shift of the angle of rotation will produce an entirely different presentation of the stitches to the airflow, producing an entirely different action. Furthermore, what if you want to just get the ball to sink a few inches at the last moment, which maximizes the effectiveness of any pitch? How do you orient the ball and how much do you rotate it and in what direction? On top of that, if you were so good at throwing this that you could exactly reproduce some so-called "ideal" rotation, you now have a predictable and therefore eminently hittable pitch.

What is observed in the lab as a theoretical ideal may not translate well to real-world situations.

Now, there's one effect that nobody talks about that I've observed that explains why a non-rotating knuckleball may still swoop all over the place. I call this the Ferris Wheel Effect.

Ride one and notice that although you always face forward, the air comes at you from above as you rise up, then it shifts to the front as you reach the top, then from below as you ride down the front.

A non-rotating knuckleball, thrown slow in a big arc, "sees" the wind from slightly above the front-center, then directly in front, then slightly below front-center. This movement of the "relative wind" (as skydivers call it) along the front of the ball will naturally produce shifts in where and how those stitch-produced swirls ("vortices") are produced and therefore how the wake is shaped and sized. It's known by wind tunnel tests that only a small rotation of a knuckleball can produce a huge change in this wake, so practiced knuckleball pitchers who can keep the spin off experiment with different orientations of the ball in their hands to produce the ideal action for them personally. Some settle on a "horseshoe" facing front, others prefer the point where the seams come closest together, and others try variations on those.

I have no hard figures, but it seems that a thrown ball begins to react to the wind at speeds around 50 mph or so. Non-rotating balls thrown at, say 90, have much less time to move much. The tradeoff seems to be around 70 mph, fast enough for some Little Leaguers to try their hand at one at major-league distances. You can see how helpful it may be to intentionally change speeds on your knuckleballs. Throwing some in the high 70s and low 80s will cause them to move little, but they'll still likely shudder and shake enough to knock that very small sweet spot off-line enough to make a clean hit all the harder. However, if it doesn't move at all, it's essentially a BP fastball,which may make it all-too easy to hit deep. If  you have an insufficient number of them sinking the way you want (sinking being the best way to miss the bat) your option's there to change velocity of each pitch on your own, further adding to the batter's confusion. There is something to be said for having the command of the knuckleball so as to produce a lot of strikes and produce enough difference from one to another to have a good day on the mound, whether the ball does it all on its own, or if you have some say in what it does.

It has been found, though, that if the ball rotates slightly clockwise or counter-clockwise and the stitches are aligned properly from the start, there may be some lift produced to follow, producing a ball that actually corkscrews! I've thrown them-- tough to do all the time, to say the least. This may be what Hoyt Wilhelm called it his "spinner". It rotates about an off-center axis pointing generally towards the plate, gradually sinking on the way in, making it relatively easy to catch and to throw for strikes but completely unhiuttable except with blind luck.

Also throwing into the wind produces more air speed. Over the same distance, the ball may move a LOT more, compressing its action into a much shorter distance. A REAL catcher's nightmare!

UPDATE  3/26/07

It's been some time since I authored this page, and I now have further information in-hand to make this complete.

The Knucklebook has been on the market for a year, and I wrote it with all the how-to information I had at the time, knowing that publication would bring out people complaining that I missed important notes. And I did. Here they are:

Through the relentless experimentation of Charlie Hough, it's learned that a knuckleball will produce it's wildest and most repetitive unpredicatable action when one of the U-shaped seam areas is faced forward, and the ball slowly rotates forward not much more that 1/2 rotation. I'm sure that extensive experimentation in a wind tunnel will confirm this in hard numbers, but as a practical matter, it's tough to beat the efficiency of a curious legendary major-league pitcher.

One may also note that too much forward spin will produce more of a predictable path. If you want this or not when on the mound, that's your decision.

Understanding all this and looking at an ideal split-finger fastball should make you realize they're essentially the same pitch, except the grip to start them off is different, and the splitter is designed to look a lot like a fastball, only it will sink hard at the last moment. More predictable than a knuckleball, yes, but it's your decision what's the more interesting, or useful.

Simply understanding that a repetitive seam orientation and axis and amount of rotation, all found to be ideal through examination, understanding of the shifting turbulence, and long experimentation, can turn a lab curiosity into a major league pitching career. For most people, keeping it that simple will bring about the desired results the soonest. As to how one throws the ball to make it do those things, well, that's for another web page. Or you can try the book, which explains many other views of the knuckleball.

Dave Clark,