ender516 wrote:It sure can. If I do the math on my own statistics as supplied by Map Rank, which states that I have at least a 40% win loss ratio playing 1v1, that means 60% of the time I will lose my first game and have to play a second, then I will lose 60% of those games and have to play a third, which I will lose 60% of the time. That sounds bad until you multiply it out: 0.6 * 0.6 * 0.6 = 0.216. So, since 1 - 0.216 = 0.784, based on my record, I have a 78.4% chance of making it to the next round. I can probably improve my odds by a judicious choice of map each time, since that 40% includes a lot of maps I have only played once.
To put it another way, anyone with a win-loss ratio of better than one minus the cube root of one-half, or roughly 20.6%, has an even chance of advancing.
...Against statistically random opponents.
Now think about meeting people who've been through the same winnowing process. Suppose EVERYONE in a round has a nominal 1v1 % of 60%? That throws this analysis out the window eventually. Also, any given 3-game sample is easily small enough for the RNG to play a major factor. I'm not saying your calculations are useless... just that they don't adequately express the probabilities anymore after, say, half a dozen rounds.
It won't last forever, but it will take QUITE some time to play out. HA was estimating about a year... I'm thinking 1 and a half to 2. Later rounds will be a BIT faster, but there's a practical limit on round speed, especially for the early, huge rounds.