Grudger wins instead of the copycat

[rafales]

in this step copycat was supposed to win but it didn't:

[jkoelling]

The description is a bit inaccurate there: It's unlikely that the Grudgers win, but it is possible. As Nicky's description says, the two tie when all the others are out of the game. In that scenario they behave exactly the same and always get the same scores.
The setup here always leads to 25 Cooperative and 1 Grudger peep and everything after that is randomly decided. Since it starts out 25:1 it should rarely happen that the Grudgers actually win.

@ncase For me it actually didn't seem to be possible to reproduce this (using the playground + code tweaks to get the setup with 26 peeps). The single Grudger got killed off almost instantly every time, which did not feel random at all. Its chance of survival for the first round should be ~80% (if I'm not totally confused about this). It looks like your current implementation (using the comparator) for randomizing peeps with the same score is not very random. I'd say this is completely fine for the narrative part, but depending on how intensively people use the playground it might lead to unexpected behavior here and there. Interested in a pull request to fix this?

For anyone who's interested, the problem of array shuffling has been nicely visualized by Mike Bostock: https://bost.ocks.org/mike/shuffle/compare.html

[jkoelling]

Fixed by @audreyt #28.

[nerdius]

I've noticed similar things happening all over the place when the randomness is involved. E.g., when you tweak the level of miscommunication, between 8% and 11%, it's possible for either copycats, copykitten, or cheaters to win.