I think your thought experiment illustrates well that often the “Rationality is Winning” meme doesn’t quite carve the space too well. Here, rationality is using the right tactics or, if only one is available, spending the right amount of time on the right tasks proportional to how much they value their goals and how achievable they are.
If we resurrect Alice and Bob as hypothetical monovalue agents who only exclusively value deadlifting X, and who have only one method of attempting/training to deadlift X, then the game tree is skewed, Bob wins faster, Alice is screwed and wins slower. Both are fully rational if they spend all available resources on this goal (since it’s all these hypothetical agents value), even though Alice spends more resources for longer before achieving the goal.
For more game theory mumbo-jumbo: I view “rationality” more in terms of how you build and navigate the game tree, rather than a post-hoc analysis of who ended up in the best cell of the payoff matrices. Or, to put it differently, rationality is ending up at the best cell of your payoff matrix, regardless of whether someone else just has +5 on all cells of their matrix or has more options or whatever.
So if my understanding that you were making with this a critique of the “Rationality is winning” meme, I agree that it’s a bit misleading and simplistic, but it still is “taking the best course of action with the resources and options available to you, reflectively and recursively including how much you spend figuring out which courses of action are better”—Expected Winning Within Available Possible Futures
I think your thought experiment illustrates well that often the “Rationality is Winning” meme doesn’t quite carve the space too well. Here, rationality is using the right tactics or, if only one is available, spending the right amount of time on the right tasks proportional to how much they value their goals and how achievable they are.
If we resurrect Alice and Bob as hypothetical monovalue agents who only exclusively value deadlifting X, and who have only one method of attempting/training to deadlift X, then the game tree is skewed, Bob wins faster, Alice is screwed and wins slower. Both are fully rational if they spend all available resources on this goal (since it’s all these hypothetical agents value), even though Alice spends more resources for longer before achieving the goal.
For more game theory mumbo-jumbo: I view “rationality” more in terms of how you build and navigate the game tree, rather than a post-hoc analysis of who ended up in the best cell of the payoff matrices. Or, to put it differently, rationality is ending up at the best cell of your payoff matrix, regardless of whether someone else just has +5 on all cells of their matrix or has more options or whatever.
So if my understanding that you were making with this a critique of the “Rationality is winning” meme, I agree that it’s a bit misleading and simplistic, but it still is “taking the best course of action with the resources and options available to you, reflectively and recursively including how much you spend figuring out which courses of action are better”—Expected Winning Within Available Possible Futures