>>That's true, but the algorithm should work for the worst case scenario too.
>
>I was just thinking that if I were given 1 go with 2 balls, would I hedge my bets that I have a 55% chance of being better or at least no worse than another person using the more complicated method.

The probability of the ball to break on any given floor is 1/100. We can calculate probability to find this floor for n steps...

However, if we would not involve probability theory, Steve's solution is the most compact, though a little bit more complex, than Sergey's.