Managing Software for Growth: Without Fear, Control, and the Manufacturing Mindset

Muddling Through

We can't predict very far because the range of possible outcomes is dizzyingly large. Consider chess and its 10¹²⁰ possible moves. Game theory says there is an optimal solution, but remember that people are playing the game, and great googley-moogley that's a lot of potential moves.^[2] It's impossible for a human player to analyze all those moves and pick the best one at each step. Even computerized chess programs don't do that. They test small subsets of possible moves based on strategies derived from the current state of the game. Human players do the same thing. We use “rules of thumb” based on experience. Which rule we use depends on the patterns we see at the current point in the game.^[3]

^[2] Waldrop, 150.

^[3] John Holland, ...