Theres an interesting component to this near the end of the game. You know the ants will be trying to go for the best restaurant to maximise their expected return. If you're behind you can't beat them by following the same action. Since it's just win or lose, you might as well gamble on getting good scores from the bad restaurant even though you have a lower expected return. Similarly, if you are ahead you can delay switching if you see the best restaurant change if you think the ants will take time to realise, because if you just make the same actions as the ants you will win.
I wonder what are good examples of such competitive binary win/lose situations in natural environments. To be evolutionarily successful it's often enough to do better than your competitors, regardless of how well either of you do absolutely, e.g.: http://youtu.be/_Eqc6M6RUFQ?t=1m15s
Richard Mann said, Yes, me again...
Theres an interesting component to this near the end of the game. You know the ants will be trying to go for the best restaurant to maximise their expected return. If you're behind you can't beat them by following the same action. Since it's just win or lose, you might as well gamble on getting good scores from the bad restaurant even though you have a lower expected return. Similarly, if you are ahead you can delay switching if you see the best restaurant change if you think the ants will take time to realise, because if you just make the same actions as the ants you will win.
I wonder what are good examples of such competitive binary win/lose situations in natural environments. To be evolutionarily successful it's often enough to do better than your competitors, regardless of how well either of you do absolutely, e.g.: http://youtu.be/_Eqc6M6RUFQ?t=1m15s