C. Rowe, J. M. Harris, S. C. Roberts
It has been drawn to our attention by David Matsumoto and Stephanie Hata that our first-round analysis of the Athens 2004 judo competition was biased by non-random allocation of blue judogis to a small number of seeded competitors. Although only 25 of 301 contests analysed are affected, the assumptions upon which those analyses were based were therefore incorrect. However, their reanalysis (D. M. and S. H., unpublished results), which excludes first-round matches, supports our conclusion that winning biases exist for male judo competitors wearing blue. Our interpretation of visibility-influenced winning biases therefore remains a valid and plausible alternative to that of Hill and Barton (R. A. Hill and R. A. Barton Nature 435, 293; 2005).