Link: http://www.twitch.tv/twitchplayspokemon An interesting social experiment is taking place right now, where users all over the internet are invited to simultaneously command the movements of what seems to be a game of Pokemon. Amidst the wildest chaos, the users's efforts have managed to get them half-way through the game in just five days. With 8 different commands available and an average of 70,000 users online, the probability of the command issued being the correct one is 12.5%. However, I suppose getting through some places in the game require a certain amount of consecutive correct inputs. How bad does it get for our hero if this amount gets noticeably high? How long can the wisdom of the crowd prevail? What interesting statistical models can we apply to experiments like this?