Open this lesson in your favourite AI. It'll walk you through the why, explain the demo, and quiz you on the try-it list.
Before any payoff matrix, a game is just three things: players, the moves each can make, and the outcomes those moves produce. Once you can name all three for a situation in front of you, you've already done the hard part — most real-world 'games' go wrong because someone left out a player, forgot a move, or mis-stated what success even looks like. This lesson is about slowing down at that step.
A formal game in normal form is a triple . Players, strategies, payoffs. Every game you'll meet in the course is some flavor of this:
Use these three in order. Each builds on the one before.
In one paragraph, explain what counts as a 'game' in game theory like I'm a smart non-mathematician. Use an everyday example.
Walk me step-by-step through how I'd convert a real-world situation (say, two neighbors deciding whether to share a fence cost) into the tuple (N, S, u). Where is it easy to make mistakes?
I have a situation with three parties and one of them has hidden information. Does the (N, S, u) formalism still apply, or do I need a different object? If different — what's the smallest extension?