Same one spot if you always have the same foot back.
To determine the number of potential "skid patches", you start by factoring each gear. Find the least common multiple, which means multiplying all of the unique terms in each gear. So if I had a 48/16 ratio, my terms are 2*2*2*2*3 = 48, 16 = 2*2*2*2--The set for 48 actually contains all of the unique terms, so multiplying them = 48 = the LCM(48,16). On the other hand for 46/18, the terms are 46 = 2*23, 16 = 2*2*2*2, so the LCM(46,16) = 2*23*2*2*2 = 368. So whenever the number of teeth the chain passes over counts a multiple of 368, that means both the cranks and wheel have returned to the same position that they started in.
That comes out to 8 rotations of the crank, so on any given skid (assuming the same leg position), you have a 1-in-8 chance of hitting any given skid patch on the wheel.
If your ring and cog have relatively prime numbers of teeth (that is, they share no terms in common), then you have a best case scenario where the number of skid patches is equavalent to the number of teeth in the cog. My case, with a 47/16 ratio: LCM(47,16) = 47*2*2*2*2 = 752. 752/47 = 16 (big surprise) => 16 unique skid patches.
Last edited by bostontrevor; 11-21-04 at 10:32 AM.