Although I agree with their conclusion, their statistical analysis is sketchy. E.g.:
The sole anomaly arose around the most critical number: zero. Based on current tabulations, four states cluster around the zero, showing a victory margin of less than 1%. Four of four show Biden ahead: Georgia (0.2%), Arizona (0.5%), Pennsylvania (0.7%), and Wisconsin (0.7%). That heavy skew towards one side is statistically anomalous—occurring with probability less than 0.01 (one chance in 100). A quick check of the past four Presidential races (i.e., 2004-16) shows that the parties split close races evenly.With four very close elections, it's basically a coin flip for which way each will go. There's a 1 in 16 chance that Biden would win all four, much larger than 1 in 100.
But it’s not a coin flip; you’re in effect taking the average of averages, which is a no-no.