Knuckleball Hangovers

WAR might tell you otherwise, and won-lost record certainly does, but so far this season, the ace of the Red Sox’s staff has been knuckleballer Steven Wright. Last night, Wright allowed just three hits while striking out seven batters in seven innings. Sure, it was against the Braves, so that might not count, but even before that start, Wright had a 1.40 ERA and had faced the Blue Jays twice and the Astros once in Houston. With all of the problems the Red Sox have elsewhere in their rotation, Wright seems to be earning himself a job for the rest of this season.

This post isn’t about Wright. It’s about the batters who will face Wright. Earlier this week, ESPN’s Eric Karabell wondered on his podcast whether hitters experience a hangover effect in the days after they face a knuckleballer, a question I thought would be fun to try to answer.

Since 2002, there have been seven starters who threw knuckleballs at least 50 percent of their pitches in a season(s):

There have been 885 batters who have faced one of those knuckleballers at least once in a season with 200 or more plate appearances. And if you count each instance of a batter (in a season in which he had 200 or more plate appearances) facing a knuckleballer separately, then there have been 5,307 such matchups.

To test whether or not hitters suffer from a knuckleball hangover in the days following a start against a knuckleballer, I compared hitters’ batting averages and wOBAs for the season (excluding their games against the knuckleballers themselves) to their lines in the days following their games against the knuckleballers. Then, as a sanity check, I also checked the hitters’ performances in the days leading up to the knuckleballer game. Here is what I found:

It’s not unusual for the combined average difference in batting average and wOBA to fluctuate 10 or so points above or below hitters’ seasonal numbers on any given day, so I made the chart using the cumulative performance of hitters in the days leading up to and the days since facing a knuckleballer. That means for the day after facing a knuckleballer, the chart shows hitter stats for that one day. For the second day after facing a knuckleballer, the chart shows hitter stats for both the first and second days. And so on. That approach causes both sides of the days when hitters face knuckleballers to converge toward an average difference of zero. However, notice that the two sides do not converge at an equal pace.

The day before facing knuckleballers, hitters happen—I am assuming this is because of randomness—to perform slightly worse (seven points of AVG and eight points of wOBA) than their seasonal lines, but as soon as you reach two days before, the convergence has already decreased the difference to within two points of AVG and three points of wOBA. That difference does not get bigger as you move backward toward seven days prior to the knuckleballer games.

The day after the knuckleballer games, hitters show a similar decrease (six points of AVG and 12 points of wOBA) that they did the day before, but notice that the difference generally remains for a few days before more slowly converging toward zero.

Perhaps this is evidence of a hangover effect for hitters after they’ve faced knuckleballers, but the effects are too small to make any difference in fantasy. So all you fantasy players with a lot of Braves on your team can rest easy knowing that they shouldn’t be too impacted by the Wright start when they face off against Clay Buchholz tomorrow. Of course, if you have a lineup full of Braves, you undoubtedly have much bigger problems to worry about.