Conversation

The 5 mother sauces of game theory 1. Prisoner’s dilemma 2. Ultimatum 3. Dictator 4. Rock Paper Scissors 5. ??
49
160
Umm... people are going all over the place in the replies. I’m talking about the 5 most interesting games that fit formal game theory in von Neumann/Morgenstern sense. Games that can be described by payoff matrices for finite sets of simultaneous moves.
2
16
This doesn’t fit classic game theory since it’s an alternating move game, but a great almost-fit candidate. Alternating move games are too vast a space to admit a list of 5 eigen-games I think.
Quote Tweet
Replying to @vgr
5. You cut the cake, I choose the piece.
3
16
Replying to
I think to save this thread I’ll allow for 2-step alternate move like cut cake/choose piece, but not open ended ones like chess. So mother sauces = highly fertile games, describable by payoff matrices for simultaneous (1 step or iterated) move or alternate move (2 moves only).
1
3
Trying to identify a decently limited set of models to “cook” with. Point is models for situations that admit equilibrium conditions to emerge. Nash, ESS, and whatever we call things like cut cake/choose piece leading to exact half being the optimum.
1
5
They are classes of phenomenology that may or may not be amenable to modeling as games. For example, the scene in A Beautiful Mind where Nash figures out how to get partners for a group by ignoring the prettiest girl is a model of a status dynamics situation.
Quote Tweet
Replying to @vgr
Social status games/signalling (or is that not considered game theory?)
1
5
Yeah this is a fun bunnytrail. I once did such a model for PD using kinship as a parameter to create a spectrum from pure cooperation to pure adversarial. If PD payoff matrix is P, consider the game defined by P+ λP’... you get fun moving-equilibrium behavior as you vary λ
Quote Tweet
Replying to @christophewllms and @vgr
If you’re trying to map out foundational games, it would be interesting to lay out which hyperparameters tie them together. Small changes in incentives near a boundary on the game map can drive large structural differences.
1
3
🤔
Quote Tweet
Replying to @vgr
Sorry if this was already in the replies or doesn't meet exactly what you're looking for but: -Stag Hunt (en.wikipedia.org/wiki/Stag_hunt) is an interesting variant of Prisoner's Dilemma -not sure if "mother sauce", but one of my faves is El Farol Bar problem (en.wikipedia.org/wiki/El_Farol_)
2
3
There’s so many derivatives and interesting degeneracies this really needs a sort of class hierarchy tree with 5-9 phylla
1
4
There is good raw material in this thread for someone who wants to write a modern intro essay to game theory from a kind of ethnographic salience perspective rather than a textbook perspective.
1
5
It feels like game theory is used backwards, as a class of math problems that have applications. I prefer to think of it as a schema for organizing social phenomenology with a minimalist set of triage models. If you can’t find a model that fits, you have to actually think.
1
17
When I was younger the distinction between zero-sum and non-zero-sum seemed profound to me. Now it strikes me as a high-modernist red herring that gas been overloaded into the philosophy equivalent of factory farmed junk food.
1
5
To snowclone a Michael Pollan phrase, game theory used as analysis results in “processed metaphysics-like substances” Game theory used to organize field observations otoh is interesting. But to do that you have to appreciate variety rather than Nash everything together with PD.
2
5
Game theory may have jumped the shark when that guy wrote that mega thread “time for some game theory” in 2016. It might be impossible for a formalized subject to recover from that kind of muddied thinking abuse.
1
8