i see that i elided from my reasoning the immediate match of the original marble. 39/51% of the time, the second marble will not be identical to the first. the third marble has a 24/50 chance to match either the 1st or the 2nd marble. therefore, 1-39/51*24/50 of the time, there will be a match.

sorry, shoddy reasoning on my part.

what a relief, i was thinking i'd have to quit playing poker!