Consider a poker tournament, if you have any familiarity with them... There is much variance in a poker tournament, and the largest ones can have thousands of entrants. The "best" poker players aren't guaranteed to win, but they have a much better chance of winning than the 97-year-old guy who is playing his first tournament. When they get down to the final players, the majority of the final-table players are the ones who were considered the "best" tournament poker players before it started.
It seems like you have this idea that if we could number the snakes from 1 to 10,000 in order of "fitness," we should see them finish in precisely that order in a mating contest, and anything else disproves evolution. That's just not how it works.
Consider this scenario (I'm not saying this is what happens, just making something up based on the article): Let's say you have a snake that is much stronger than the others; when he wraps himself around something, he can squeeze with more force than the other males. In these mating balls, the males wrap around the female and she rolls to get them off her. Maybe the rolling process is to test for the stronger males, since she won't be able to roll them off. Or in a related scenario, maybe the "fittest" males are the ones with more endurance. The weaker males get tired quickly after being rolled off the female a few times and they give up. But the fitter males have more endurance and they get to mate because they don't give up as quickly. In these scenarios, these fitter males, who may be in a mating ball with 100 other males, don't have a 1/100 chance of mating (which, as you say, would seem to negate his "fitness advantage")--they have more like a 1/10 chance or something like that.