Main

If, as Hill and Barton claim, there is something special about the colour red, then contests using other colour pairings should not be biased. We tested whether this could be the case by analysing data (http://www.athens2004.com) from contests in which red was not used for competitors' outfits: in judo matches, one player wears blue and the other wears white. We followed the methodology originally used by Hill and Barton3 for boxing, tae kwon do, and Greco–Roman and freestyle wrestling.

After ensuring that outfit (judogi) colour in judo contests was allocated at random (http://www.ijf.org), we found a significant winning bias for players wearing blue compared with those wearing white (χ2=7.34, d.f.=1, P<0.01), and a similar effect of contest symmetry on winning bias to that reported by Hill and Barton (Fig. 1). We also found the winning bias for players wearing blue when considering only contests in the first round of competition (χ2=4.85, d.f.=1, P<0.05). This result excludes the possibility that the observed bias might arise through skilled contestants being placed, by chance, in draw positions where they wear blue more often as they progress through the competition.

Figure 1: Influence of judogi colour on the outcome of judo matches in the Athens 2004 Olympics.
figure 1

The black line at 0.5 indicates the expected proportion of wins by blue or white under the null hypothesis that colour has no effect on contest outcome. Blue bars, proportion won by players in blue; white bars, proportion won by players in white. There were significant differences between the number of blue and white wins for all contests combined (χ2=7.34, d.f.=1, P=0.007). This difference is most evident at low degrees of asymmetry in relative ability of the two competitors in each bout (subdivided using quartiles of points difference, after Hill and Barton3). Differences between the number of white and blue wins were as follows: no asymmetry (χ2=1.53, d.f.=1, P=0.22), small asymmetry (χ2=6.31, d.f.=1, P=0.012), medium asymmetry (χ2=3.19, d.f.=1, P=0.07) and large asymmetry (χ2=0.57, d.f.=1, P=0.81). Number of contests recorded are 301, 79, 99, 53 and 70, respectively.

Our results indicate that there is nothing inherently special about red in terms of colour-associated winning biases.

We can think of no plausible evolutionary explanation based on animal behaviour or evolutionary psychology that might account for a winning bias for blue contestants. We propose instead that outfit colour affects opponent visibility, which is crucial for avoidance and interception, and for anticipating behaviour. Visual abilities that could influence sporting performance include being able to follow rapidly moving objects and perform fast visual searches4,5,6. And the hue, saturation, brightness and contrast of an object (or opponent) could enable it to be picked out against its background7,8. These factors are critical for combat sports and for detecting teammates on the field of play (http://www.liv.ac.uk/researchintelligence/issue1/manunit.html; 1999).

In judo, the white judogi is likely to be perceived as brighter than the blue and may have higher contrast against the background. Men wearing blue may therefore have a visual advantage in being able to anticipate their (white) opponents' moves. We do not know the reflectance spectra, lighting arrangements or other visual factors that might have affected the visual salience of the red and blue outfits worn in the sports studied by Hill and Barton. Although our hypothesis is untested, visibility differences could also explain the biases they found. The visual attributes of sporting wear should therefore be considered in this wider perceptual context.