Originally Posted by mander
The skid patch theorem says it's the # of chainring teeth whose oddness or eveness makes that difference. Is that wrong?
Looking at each column in the table I would say it is the cog not the chainring.
EDIT: I was thinking about this a bit more. The best way to maximize the number of skid patches is to use a chainring and a cog that are prime numbers this will give you a total number of skid patches equal to the number of teeth on the cog. Since the max number of skid patches one can have can not be greater than the number of teeth on the cog, I would say the cog dominates.
However, upon further inspection I also realized that one can have a cog that is not prime but does not divide the chainring and still obtain skid patches equal to the number of teeth on the cog. In other words the cog and the chainring will maximize the number of skid patches if they are relatively prime.
Also, I just found this.