Does CDT imply that they can’t have a member named Schelling who figures out that it’s preferable to freeze out the 9-bots while taking the 5s from cooperators. Pre-commit to 5, problem solved. In any starting mix (with at least 2 of each) of 9-bots and 5-bots, the 5-bots will eventually completely take over, as they’re the only ones who ever get any payoff.
If nobody ever says 1, the 9s die pretty quickly.
I don’t actually understand whether CDT is so weak as to be a useless strawman that never described any actual agent, or whether the analogies and descriptions we’re using are just stupidly ignoring all of game theory and strategy research.