Skip to main content

A Critical Assessment of the Evolution of Standard Game Theory

  • Chapter
  • First Online:
On Coordination in Non-Cooperative Game Theory

Part of the book series: Springer Studies in the History of Economic Thought ((SSHET))

  • 111 Accesses

Abstract

In this chapter, we adopt a history of economic thought approach to highlight the methodological problems encountered by standard game theory in its treatment of strategic reasoning and the consequences this has on the nature of players’ beliefs. We show that whatever contributions in the history of standard game theory; of von Neumann and Morgenstern of Nash of the refinement program in the 1980s and of epistemic game theory; by focusing only on the search for a solution concept, these contributions reduce strategic rationality to the existence of a solution. They reduce a game to a simple “black box,” regardless of how the game is played. We show that such focus on solution concepts is related to the norms of research of the mathematical community. We show in particular that this leads to serious difficulties in the epistemic program of game theory, which aims to define the knowledge and belief structure of players compatible with the concepts of solution. Epistemic game theory is grounded on the idea that to guarantee the existence of a solution, players’ beliefs must be a priori coherent and correct. From that prospect, we show that players’ beliefs are not individual and subjective mental states; regardless of the nature of the players’ beliefs a priori, they are by construction the result of their reasoning and not an element intervening in this reasoning. This means that players’ reasoning about other players’ choices or beliefs is excluded from the analysis of epistemic games, as well as strategic rationality.

This is a preview of subscription content, log in via an institution to check access.

Access this chapter

Chapter
USD 29.95
Price excludes VAT (USA)
  • Available as PDF
  • Read on any device
  • Instant download
  • Own it forever
eBook
USD 99.00
Price excludes VAT (USA)
  • Available as EPUB and PDF
  • Read on any device
  • Instant download
  • Own it forever
Hardcover Book
USD 129.99
Price excludes VAT (USA)
  • Durable hardcover edition
  • Dispatched in 3 to 5 business days
  • Free shipping worldwide - see info

Tax calculation will be finalised at checkout

Purchases are for personal use only

Institutional subscriptions

Notes

  1. 1.

    The way vNM sees modeling in the TGEB is that setting a new theory requires an heuristic stage allowing a move from nonanalytical and “commonsense considerations” (Giocoli, 2003, p. 248) to formal and mathematical methods. The next step is to provide a general and formal theory which requires conforming to axiomatic methods in order to be rigorous. Then, one should build a rigorous and conceptually general formal theory. Such a theory should be applied to the resolution of some elementary problems to test the conformity of the theory and then to more complicated problems to again test its conformity. Finally, such theory must be employed for predictions (von Neumann & Morgenstern, 1944, pp. 7–8).

  2. 2.

    Léonard (2010, p. 64), however, claims “[a]lthough, in general, maxxminyg(x, y) ≤ minymaxxg(x, y), it is not generally true that the equality holds.”

  3. 3.

    For more details on the indirect proof method see Giocoli (2003, pp. 263–76).

  4. 4.

    The misapprehension of the proof of the minimax theorem of von Neumann in 1928 is according to Léonard (2010, pp. 66–67), explained by von Neumann himself because in his equilibrium growth model of 1937 he makes the formal link between the fixed-point argument of the growth model and the minimax theorem.

  5. 5.

    This is linked to the maxims of behaviors. As explained by Giocoli (2003, p. 235), people must decide to follow a maxim of behavior or not: “there are two types of maxims: the unrestricted maxims, which can be followed regardless of the actions of the other agents, and the restricted maxims, which are followed on the basis of whether the others do the same or not… In the case of unrestricted maxims the choice poses no problems, because the agent’s evaluation is not disturbed by the behavior of other individuals. In general, however, the decision depends upon the agent’s forecasting ability which, in the case of restricted maxims, must also embrace the other agents’ behavior.”

  6. 6.

    Such interpretation in terms of BR reasoning is the most common one but Nash actually presents the NE in terms of countering (see Giocoli, 2003, p. 301 for more details).

  7. 7.

    More specifically, consistency formally entails that the set of players’ beliefs are “the limit of the conditional probabilities induced by players’ strategies in some perturbed game.” (ibid., p. 6).

  8. 8.

    The formal justification for the encoding of players’ hierarchy of beliefs according to the definition of the players’ type is later provided among others by Armbruster and Böge (1979), Böge and Eisele (1979), or Mertens and Zamir (1985), (see Brandenburger 2010). Mertens and Zamir (1985) in particular proved that the type of the player allow encoding of the player’s infinite hierarchy of beliefs.

  9. 9.

    For an epistemological reflection on the concept of “types” in Harsanyi’s work, see Hargreaves Heap and Varoufakis (2004).

  10. 10.

    See Perea (2012, 2014) and Brandenburger (2010) for an historical perspective on the transition from classical to EGT. See Brandenburger (1992, 2007), Geanakoplos (1992), Dekel and Gul (1997), Battigali and Bonnano (1999), Dekel and Siniscalchi (2015) for surveys on the epistemic program on game theory and its main assumptions and results from a formal perspective.

  11. 11.

    The formal proof of the equivalence is given by Brandenburger and Dekel (1993).

  12. 12.

    Friedell (1969, p. 31) more precisely defines the concept of “common opinion” and states the real conditions in which common opinion can rise. Like Lewis, and contrary to Aumann (1976) Friedell is “interested in real life situations” (Perea, 2014, p. 12). Aumann only states that “at a given state ω there is common knowledge of an event E among two persons A and B, if at ω both A and B only deem possible states in E, if at ω both A and B only deem possible states at which A and B only deem possible states in E, and so on, ad infinitum.” (Perea, 2014, p. 13).

  13. 13.

    Giocoli (2003, p. 392) specifies that “[t]he tradition in SEUT is to circumvent the problem either by adding an extra postulate (the sure-thing principle: ibid, pp. 39–40) or by having recourse to an imaginary experiment, — a lottery — able to turn the preferences’ domain into a probabilistic space. The former is the approach originally followed by Savage but it is quite involved, so it is the latter method, first proposed in the early 1960s by Anscombe and Aumann, that has gained acceptance.”

References

Download references

Author information

Authors and Affiliations

Authors

Rights and permissions

Reprints and permissions

Copyright information

© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this chapter

Check for updates. Verify currency and authenticity via CrossMark

Cite this chapter

Larrouy, L. (2023). A Critical Assessment of the Evolution of Standard Game Theory. In: On Coordination in Non-Cooperative Game Theory. Springer Studies in the History of Economic Thought. Springer, Cham. https://doi.org/10.1007/978-3-031-36171-5_2

Download citation

Publish with us

Policies and ethics