3

Apologies for not knowing the nomenclature; I'll try to describe what I mean.

The optimal strategies for some imperfect information games involve purposefully doing badly in some cases, so that bluffs are more effective in other cases. You purposefully lose more in some lower margin subtrees, to create confusion in some higher margin subtrees and win more there.

I don't know if poker has this counterfactual property or not. We're talking about something a bit more twisted than just bluffing or hussling. I wrote a blog post describing an example game with this counterfactual property, if that helps.

Personally, I find this counterfactual property counter-intuitive. Intuitively I would have guessed that you should always try to do your best in every case. This intuition gets stronger (and wronger) when I think of it in terms of things that matter a lot (imagine trying to explain to someone that you're purposefully losing a war so that you could counterfactually win it more often in alternate histories! They'd arrest you on the spot!).

Has this particular possible disconnect between intuition and optimal play been investigated? Have there been studies checking if humans have difficulty following optimal strategies that involve losing for counterfactual gain?

Craig Gidney
  • 151
  • 2
  • You may be referring to Game Theory (http://en.wikipedia.org/wiki/Game_theory), which is a very large field. The Prisoner's Dilemma (http://en.wikipedia.org/wiki/Prisoner%27s_dilemma) is a well-known example where players have to counter-intuitively take a hit to come out winners, and occasional strategic bluffing produces optimal results. This field is far too broad to cover in a SE format, so you'll need to narrow down the scope of your question a lot to get more useful feedback. – Arnon Weinberg Apr 04 '15 at 21:22
  • @ArnonWeinberg I'm not sure I understand your advice. The question is already about a narrowed down aspect of game theory. If I had asked "Are there any studies of bluffing?", which is a more general property, I expect that would have already been specific enough to get answers. – Craig Gidney Apr 04 '15 at 23:28
  • There's some research on this, but I would have to dig through books I haven't touched for years at home. From memory: yes, (American) people do tend to engage in hill climbing strategies (i.e., do your best in every case/move in a direct line from starting state to goal state), and have to either discover or be trained to use other strategies. – Christian Hummeluhr Apr 11 '15 at 07:12

0 Answers0