Comment by Doris Kothe.

The statistics do not prove that the market in Wuhan was the early epicenter of the COVID-19 pandemic. The main statistical test used is nonsensical, according to Stoyan, and there are other flaws in the statistical reasoning. Stoyan and Chiu used the same geostatistical data as Worobey and colleagues. These are the residential addresses of the first 155 coronavirus cases. These addresses were entered as points on a map. In the Science article, this dot pattern is compared with dot patterns generated by simulation, finding significant differences. However, since these artificial dot patterns are incorrectly chosen, the test used must always reject the hypothesis that a location other than the market is the epicenter. In other words, Worobey and colleagues have excluded locations other than the Huanan Seafood Wholesale Market from the outset. However, other locations in the vicinity of the market could also be possible candidates: A large railway station and a huge shopping complex with hotels and restaurants.