We can probably count on one hand (and certainly no more than two) all the well known knuckleballers in MLB history – the Niekros, Tim Wakefield, R.A. Dickey, Wilbur Wood, Hoyt Wilhelm, and Charlie Hough. Well, now we can add another to the list as the 30 year old Steven Wright has been called up for his third start in three years. Ian York uses his incredible talents and visual analysis to help us imagine the speed, spin, and break that a MLB hitter is subjected to when facing the unpredictable knuckleball.
The knuckleball is not a difficult pitch to throw. Many pitchers, from Little League to the Majors, can throw a knuckleball. The knuckleball is, however, a very difficult pitch to throw for a strike. Very few pitchers at any level can consistently hit, or get close to, the strike zone with a knuckleball. This is why of the several hundred pitchers in professional baseball, only a handful at any one time are primarily knuckleball pitchers. So far in 2015, two pitchers have thrown knuckleballs in anger: R.A. Dickey and Steven Wright. (Three, if you count veteran catcher David Ross. In one mop-up inning for the Cubs on May 9, Ross threw 11 pitches – eight of which were deemed by a baffled PITCHf/x to have been knuckleballs).
Steven Wright has been tapped to replace Justin Masterson in the Red Sox rotation, at least until the latter is off the Disabled List and, potentially longer, if Masterson can not regain his effectiveness. Wright started his first game of 2015 for the Red Sox on Sunday, May 17, lasting five innings against the Seattle Mariners, giving up 3 runs (two earned), and taking the loss in a 5-0 game. Even though it was a fairly unexciting start – neither impressively good, nor spectacularly bad – knuckleballers are unusual enough to take a closer look at his output.
This was actually Wright’s third major-league start – one per year since 2013 – to go with his ten relief appearances, so he isn’t completely new to the big leagues. He throws a fairly fast knuckleball, hovering around 75 mph, about the same as R.A. Dickey’s, but significantly faster than the mid-60s knuckler that Tim Wakefield featured. He mixes in rare “fast”balls (in the mid-80s) and slightly less rare curves that drift in at about 66 mph.
The most challenging thing about throwing a knuckleball is throwing it for strikes, but Wright has been fairly successful in this regard with about a 65% strike rate for his knuckler in 2015. On May 17, only eight of the 37 he threw to left-handed batters were called balls, for a 78.4% strike rate. He was less fortunate with RHB, with 14 of 35 being called balls (60% strike rate). Only three of his knucklers were hit successfully:
He threw just three of his anemic fastballs, all to RHB, and presumably due to the factor of surprise, they were extremely effective, drawing two swinging and one called strike:
His eight curves were his least effective pitches in this game, yielding five balls, one out, and two hits, including a home run:
The key to a knuckleball is its unpredictability. Knuckleballs are thrown with very low spin rates – Wright’s average about 750 RPM, Dickey’s about 800 RPM; standard curveballs average nearly double that, fastballs nearly triple — so that each pitch follows a different, unpredictable trajectory. (These spin rates are not accurate. While PITCHf/x gives a spin rate for pitches, they are not directly measured but are inferred from the pitch movement. Since knuckleball movement doesn’t come directly from the spin, the algorithm PITCHf/x uses to infer spin rate gives meaningless numbers. In reality, the spin on all these knuckleballs is very low, probably about 100 RPM. Thanks to Alan Nathan for noting this!) However, Alan Nathan has shown that while the trajectory is unpredictable, it is not erratic. As with other pitches, the trajectory of a knuckleball can be calculated from its initial velocity, direction, and accelerations, meaning that PITCHf/x can determine the path of a knuckleball. (Nathan adds that a knuckleball does have some erratic motion, but only on the order of a few tenths of an inch.) However, the batter has no way of telling how any particular knuckleball will move, or where (or if) it will cross the plate, making these pitches very different from the rest of the pitcher’s arsenal. Normally, because certain pitch types move in predictable ways, a batter can afford to be selective. With a knuckleball, which could end up anywhere in or around the strike zone, selectivity goes out the window.
Here are all the knuckleballs that Wright has thrown in 2015 in MLB (including his relief appearances as well as his start). Unlike most pitches, which form one or two tight clusters in a specific region, these form a diffuse cloud that evenly covers the whole strike zone:
The following charts show pitch velocity, break angle, and break length, which serve to cluster pitches into three groups: Fastballs, curves, and of course knucklers:
The knuckleballs form an uncertain smear in the middle of the speed axis, with a wide range of break angles and break lengths. It is interesting that Wright’s curves also form a rather loose cluster, spanning a variety of break angles and break lengths.
Here is what these unpredictable pitches are like. This animation shows Robinson Cano’s third at-bat against Wright. He had flied out in the first inning, grounded out in the third, and in the fifth he faced eight consecutive knuckleballs before striking out:
Baseball has its share of teenage prodigies, magnificent athletes who excel in every sport, and million-dollar arms that overwhelm the opposition through pure power. Steven Wright is a 30-year-old journeyman pitcher who has spent most of his nine seasons in the minors. He has a slight pot belly and throws about the same speed as a decent middle-school pitcher. But he throws a knuckleball for strikes, which has given him the chance to play a significant role on the Red Sox this year, and he struck out a man who is earning $240 million dollars to hit baseballs.
This is why baseball is fun to watch.