Talk:Elo rating system/@comment-188.178.187.14-20120131094316

LOL is comparable to a flip the coin game. Only the outcome is relevant, and it can be either win or loss.

If you take 1000 equally good players, or 1000 monkeys, or 1000 identifiable bots, and play 10 games, the average win/loss ratio will be 50:50. However 0,1% or 1 of the test subjects are gonna win all 10 matches, and end up with a very high ELO. The outcome of the experiment is likely to follow a normal distribution.

In this statistical environment a single player has to try and make a difference. That is quite dificult, as you only account for 1/5th of the power in the experiment.