As I’ve written before, the random player isn’t exactly random. I refer to it in my code as SmartRandom, and it uses a very simple heuristic to determine how to play:
Look for winning moves and, if found, choose one of them at random
If none are found, look for the opponent’s winning moves, and block one of those
If none are found, then choose a move at random
It’s a very simple system, with no forethought or strategy, so I didn’t expect it to do very well. However, it has dominated the tournament so far, with only one loss in 300 games. This begs me to ask, how often does this player win?
Well, this is encouraging. Sure, the absolute numbers aren’t great (this is games won out of 100, so the numbers are effectively percentages), but the fact that they CAN win against the SR means I should see it start to fall any round now.
Although, if the numbers are this high, then I should have seen the SR lose more by now. I’ll be watching the next few rounds to see if it happens.