The Knuckleball Hangover: Part 2

In part one of the "Knuckleball Hangover" series we saw that Major League batters demonstrated a decrease of approximately six points of batting average and twelve points of wOBA the day after facing a knuckleballer. We suggested that if a hangover is real, then managers might use it by placing high-strikeout pitchers after knuckleballers in the rotation -€– taking advantage of batters whose pitch identification routine was recently jarred by trying to follow the atypical, fluttering movement of knuckleballs.

Although there seems to be statistical validity to the decrease in average and wOBA, the variables potentially responsible are innumerable. To attribute causation, that is, to demonstrate a knuckleball hangover is more than correlation or coincidence, we should be able to observe the hangover effects elsewhere in the data. To this end we settled on launch angle and plate discipline as two logical statistics that would have the fingerprints of a knuckleballer on them, if indeed a hangover exists.

In part 2 we'll look at launch angle. Every kid who grew up watching Ken Griffey Jr. has practiced swinging like him. Although most professional batters swing in relatively similar ways, each batter has a routine, a "normal" or base swing. We are examining launch angles to see if there is any evidence of batters producing atypical angular results following a match with a knuckleballer. That is: does attempting to adapt their swing to the strange movement of a knuckleball move them off of their base swing motion for more than the knuckleball game?

Let's begin with a quick primer on what different launch angles actually mean for balls put in play.

Andrew Perpetua gives us a chart that demonstrates with clarity that Expected Batting Average on Contact (xBACON) falls off pretty quickly after its peak of 14 degrees whereas Expected wOBA (xOBA) peaks around 27-degrees before fading. In sum, this tells us that across all parks and pitch types, balls hit between -3 and 36 degrees have average or better xOBA, and balls hit between -3 and 31 degrees have average or better xBACON.

If we define -3 to 36 degrees as the range in which the average launch angle falls, then a knuckleball hangover, might increase the number of launch angles found above and below that range, or, at least, push the angles toward the edges of the range more frequently.

To determine this we need to define the common launch angles for knuckleballers and for league average fastballs. (Assumption from Part 1: the league average fastball is the most common pitch faced so batters' swings are optimized to hit it.)

First, the Major League's two primary knuckleballers –€” Steven Wright of the Red Sox and Toronto's R.A. Dickey.

Dickey's average launch angle this season is 11.0 degrees while Steven Wright's is 8.9 degrees.

Next, according to PitchF/X Pitch Value/100, this year's league-average fastball representative is Danny Duffy of the Kansas City Royals.

The first takeaway here is that more balls in play launch at negative angles for knuckleballers than the league average fastball pitcher. Since we know any launch below -3 degrees drops below league average xOBA and xBACON, it is fair to say that producing a greater number of less optimal negative launch angles is part of the value produced by a knuckleball. There is, however, no evidence that knuckleballs produce sub-optimally high launch angles.

These observations can be validated by digging into our three representatives batted ball profiles. The two knuckleballers have a significantly higher groundball rate (44.25) than Duffy (33.6). The range of line drive rate is less than 1.5 percent, and Duffy's fastball produces a 45.8 percent fly ball rate versus the mid-thirties produced by the knuckleballers. (This makes Dickey's 17.4% HR/FB compared to Duffy's 10.7% and Wright's 5.2% especially interesting.)

Now that our ingredients are prepped (knuckleballs produce more negative launch angles/ground balls but not more fly balls), we know what a knuckleball hangover might taste like (besides regret), and that's a statistically meaningful increase in negative launch angles/ground balls. Anecdotally this jives with the old knuckleball aphorism "if it's low let it go, if it's high let it fly." When a batter believes a knuckleball is coming in up in the zone they swing, and, as it turns out, more often than not, they swing slightly above the optimal location causing a negative angle/ground ball. But does this persist?

In short, no. Over the past two seasons when the next pitcher in Dickey or Wright's rotation faced the same team groundball rates were elevated above their season average roughly 52.5% of the time. Well within the realm of standard baseball variance. It should also be noted that we only examined the knuckleball versus the fastball. Their negative launch angle and groundball percentage (44.3%) fall almost exactly on the league average GB% of 44.9%.

If there is a knuckleball hangover, then, it must be in plate discipline because there is no evidence of it in launch angles.

Until next time.

***

Steven Demmler is a contributing writer at Beyond the Box Score. He can be found on Twitter at @sedemmler.