Talk:Succinct game

Shouldn't the representation of anonymous games have size $$O(sn\tbinom{n+s-2}{s-1})$$? For each player, there are $$(n-1)$$ other players to distribute over the $$s$$ actions; Thus we have $$O(sn\tbinom{(n-1)+s-1}{s-1}) = O(sn\tbinom{n+s-2}{s-1})$$ outcomes. 129.187.211.170 (talk) 19:12, 15 June 2012 (UTC)
 * This has already been changed and is supported by Algorithmic Game Theory on top of that. HypesII (talk)

I question the validity of the complexity results for circuit games (which are also sometimes called boolean games - Harrenstein et al. 2001). I've already changed the complexity of finding a pure strategy Nash equilibrium from NP-complete to $$\Sigma_2^{\rm P}$$-complete which is supported both by the source incorrectly cited for NP-completeness and Dunne and van der Hoek 2004 concerning boolean games. HypesII (talk) 23:39, 18 February 2013 (UTC)