Prisoner dilemma is a game (serious one) where two prisoners are separately given two choices: confess or remain silent. If A confesses, while the other remains silent, A goes free. However, if A remains silent, trusting that his partner won't betray him by confessing, and his partner confesses, he'd 4 years sentence. On the top of that, if both confess, both are sentenced for 8 years! But, if they trust each other, without ever communicating, they get 1 year sentence.
Game Theory assumes that humans are selfish, hence the dominant incentive for both prisoners, who're locked in separate cells with informational asymmetry, is to confess, because none of them would risk trusting a betrayer. The dilemma is: executing the incentive or dominant strategy leads to worse pay offs!
This game is well-known and has various versions. I happened to know a similar game my friend used to play unconsciously.
On a rainy or unusual day, school teachers expect a very low turnout. My friend sees this as an opportunity to stand out from the rest by coming to class to impress his teachers. When he turns up in class, he's always doomed to see more than enough students turning up with the same intention (at least a considerable size of the students). Had he communicated with all and assured them that he won't come, lest others don't come, what would have been the outcome? I presume others would come too, unless the group bonding is very strong. In fact, when i was in 11th class, my seniors were united as a wall. And every other Saturday, literally none would come: girls and boys - without co-education and least outward interaction. Only 2 or 3 students out of 500+ would attend the class. After all, cooperation may be a dominant strategy (perhaps weakly) which requires development and cohesion.