Tags

, , ,

I did a brief Game Theory camp with Leo and one of his friends, Justin. It was just a half-day camp (although we may do another half-day in a a couple weeks). We started out by talking about the three central concepts in game theory: Strategy, Payoffs, and Signaling. We divided signaling into the 2×2: Intentional vs. Accidental x Truthful vs. Deceitful:

After discussing these a bit, Leo and Justin played several Minecraft-based games (Bed Wars, Bow Spleef [don’t ask me!]Bow Spleef [don’t ask me!], and TNT Wizards) and afterward we discussed how the above concepts arose in each game.

We only had three hours altogether, and unsurprisingly, the kids were more into game playing than the game theory, however, we did discover a bunch of interesting phenomena in just this brief period using these few games.

It was hard to make payoff matrices for these games. The costs in the payoff matrix are complex, mainly wasted time, and the payoffs, because there’s no real money, are hard to measure (you get emerald, or something like that). Later we also played a simpler game called “Kuhn Poker” in order to explore payoffs in a bit more detail, although I have to say that that was too simple, I think.

The signaling in this setting turned out to be really interesting. Because the game is both cooperative and competitive, and the in-game chat channel is public, it’s hard to do back-channel communication, so everything you do to cooperate with your team-mates is signaled to your competitors. HOWEVER, because Leo and Justin are sitting right next to one another (see above pic!), they have a private back-channel, sort of like the way that teams collude in online poker (called “ghosting”).

I might update this if we take it up again next week.