Simulated grunts throw off non-tennis-players in a “where’s that ball” task

Do y'all remember, from earlier this summer, the outrage that was arising in the pro tennis world over grunting? Pro-grunting players argue that producing such a sound when hitting the ball is a completely reasonable technique to get as much power as possible, while anti-grunting players argue that "excessive" noises are cheating, by preventing the opponent from hearing the sound of the ball hitting the racket, and are distracting to boot.

A PLoS ONE paper that just came out looked at precisely this question. Well, sorta this question. A laboratory approximation of this question, in which undergraduate students (none with "more than recreational tennis experience," they reassure us) watched clips of a tennis ball being hit and had to say where it landed.

They had "a professional tennis player" (no discussion of the gender or skill level of this player) hit forehand or backhand shots down a court towards a video camera centered on the far baseline, and collected clips of shots that landed in a 2 x 2 meter square at the corner of the sideline and the baseline. They counterbalanced their collection so that they had equivalent numbers of forehand and backhand shots going to each side, and then they showed each clip four times: twice with a simulated grunt (500 ms of white noise) and twice without; of these, each was shown once with the clip ending immediately upon the ball's landing, and once with the clip playing for 100 ms longer. (The instant-cutoff made the judgment task harder; the longer time to process the visual was easier. Maybe grunts only matter for hard-to-perceive things.) They asked the students who watched these clips to determine if the ball landed to the left or the right of the camera.

And what did they find?

Tennis fig 1
Here we see reaction time (so, lower is faster/better); dark bars are when the video was shown with a grunt, white bars are clips shown in silence. The hard judgements (when the clip ended exactly as the ball landed) are clearly made more slowly than the easy judgements – that's not surprising. More impressive is that the simulated grunt clearly slows people down (their ANOVA found no interaction between decision difficulty and sound; that is, the grunt had the same effect on easy-decision trials and hard-decision trials). People are 20 – 30 ms slower when they hear the noise at the same time as the player hits the ball.

Tennis fig 2
Same pattern for accuracy – people make fewer errors on easy trials than on hard ones, and fewer errors when the grunt is played than without. It's a difference of 3-4 %, but that's non-trivial in professional tennis!

This is a very pretty set of data – they've got a clean-cut effect, that's significant when measured in both error/accuracy and in reaction time – I'm jealous. My data should be this pretty.

My questions for these guys are about the validity of their task. In particular, they're showing people monocular stimuli -  a video on a computer screen isn't going to give you the three-dimensionality that you get when viewing stimuli with two eyes in the real world.

They also used a 500 ms burst of white noise for the simulated grunt – how does the length of that compare to the actual tennis grunts that are the source of such controversy?

Finally, how precise was the timing of that burst of noise? They only say "during the shot" – we know from work on the bouncing/streaming illusion that the timing of sounds played during visual perception can have huge effects on how we perceive objects to be moving.


Sinnett S., & Kingstone A. (2010). A preliminary investigation regarding the effect of tennis grunting: Does white noise during a tennis shot have a negative impact on shot perception? PLoS ONE 5(10): e13148. doi:10.1371/journal.pone.0013148

Advertisements