The only way this seems plausible to me is if you’re interpreting “expected net value difference, aggregated over the entire system, of using that tactic, as evaluated by agent A” as being roughly equivalent to “expected net value difference to A of that tactic being used, as evaluated by A.”
That is, if you are assuming an evaluation something like “There’s three of us, we all want the same girl, this tactic works, so either of you two losers using this tactic is BAD, but me using this tactic is GOOD.” Which, I agree, if everyone’s evaluating it that way, 1 and 2 are highly unlikely.
OTOH, if everyone is actually calculating the expected net value difference, aggregated over the entire system, of using that tactic, I would expect a very different result in most cases.… something more like “There’s three of us, we all want the same girl, this tactic works, so one of us using this tactic is better than none of us doing so, and probably better than all three of us wasting resources using it at once if we can agree on some other way of deciding which of us gets to use it, like drawing straws, or letting Sam have her because past experience shows he’s better at this than we are and will win anyway, or some other method.”
But I would agree that most people don’t actually do that in practice.
(blink)
The only way this seems plausible to me is if you’re interpreting “expected net value difference, aggregated over the entire system, of using that tactic, as evaluated by agent A” as being roughly equivalent to “expected net value difference to A of that tactic being used, as evaluated by A.”
That is, if you are assuming an evaluation something like “There’s three of us, we all want the same girl, this tactic works, so either of you two losers using this tactic is BAD, but me using this tactic is GOOD.” Which, I agree, if everyone’s evaluating it that way, 1 and 2 are highly unlikely.
OTOH, if everyone is actually calculating the expected net value difference, aggregated over the entire system, of using that tactic, I would expect a very different result in most cases.… something more like “There’s three of us, we all want the same girl, this tactic works, so one of us using this tactic is better than none of us doing so, and probably better than all three of us wasting resources using it at once if we can agree on some other way of deciding which of us gets to use it, like drawing straws, or letting Sam have her because past experience shows he’s better at this than we are and will win anyway, or some other method.”
But I would agree that most people don’t actually do that in practice.