OK, OK, there really wasn't a lecture 23*. But if I had given a lecture on evolutionary game theory I couldn't have done any better than the one delivered by Ted Bergstrom
here in 1995. Fine red wine is known by it's bottling date. You're in for a real treat here . This is an introduction to the
evolutionary game theorist's way of thinking with the added flavour of
Ted's razor sharp wit.

Continue on with analysis of repeated games: tit for tat strategies, the "grim" strategy. Threats and promises may need to be made credible. We start with a simple 2x2 sequential "threat game", identify the credibility problem, and then see how it is resolved in a clip from a classic cult movie
Dr Strangelove, available on a new special edition DVD. Fred Kaplan has an excellent review of this movie here.The lecture ends with a brief examination of the 10 methods for achieving credibile threats promises and comittments from Dixit and Skeath ch10.

Repeated games create possibilities for future rewards and punishments to be used to elicit "cooperative" behaviour. This lecture looks at a simple twice repeated game to get the basic idea across, then at the infintely repeated prisoner's dilemma.

This lecture continues our discussion and analysis of strategic interaction with asymmetric information and uncertainty. We examine the problem of adverse selection and the "lemons" problem (why is the used car you buy likely to be a "lemon"?) . Here, bad quality products drive out good quality products. Next we look at attempts to solve the lemons type problem: can uninformed buyers find tests or screens to help resolve their uncertainty? Can privately informed sellers provide signals like warranties or certification reports that buyers will believe?

I like this lecture. We start reviewing and re-explaining the natural frequency & truth table ideas about basic technological and diagnostic uncertainties in the venture capitalist game. Then we sketch the game tree here, concentrating on using information sets to help us set out what the players know and don't know in the game, and what they know and don't know about what other players know and don't know..etc....Then we try to figure out : what's it reasonable to believe in a game like this? Here we use our previous ideas about player's strategies to help use our beliefs about diagnostics and technologies to help figure out what should I believe here? Naturally the answer is complicated...but the intuition is worthwhile, coz we run smack into the problem of multiple reasonable things to believe...and "bad" equilibria...all becasue of private information. Two examples: why men in NZ are reluctant to enter into primary care education, why career oriented women may have a tough time finding good jobs.

We finish off our brief introduction to the element of surprise via mixed strategies by showing how mixed strategy Nash Equilibria can be found in in other 2x2 games. Then we start a new topic : games of imperfect and asymmetric information. Here, uncertainty remains a key feature, but different agent's have different uncertainties, or put another way, different agent's have different bits of private information that might be relevant to the game. We set up a venture captialist example from Kreps' Microeconomics for Managers in the latter half of this lecture (for this and the next lecture)

This lecture has two parts. The first part tests your understanding of inverse probability reasoning (we look again at the question of how much to believe a witnesses report in an accident case). The second part introduces the idea of a mixed strategy in a simple 2x2 simultaneous game that has no Nash Equilibria in pure strategies.

The second of two lectures on uncertain beliefs about two things . In our basic example one thing we're uncertain about is underlying ability that is difficult to observe directly and the second thing is a more readily observed, but also ambiguous, signal about that ability, like an aptitude test score. We use the truth table approach to get a handle on basic ideas, then explore how to think about uncertainties about our truth table uncertainties: uncertain beliefs about uncertainly held beliefs.

This is one of two lectures thinking deeply and critically about uncertainty in situations of strategic interaction. The first part of the lecture develops the operational subjective concept of probability, using the prices/odds set by the sports betting bookies at NZ's TAB to show how subjective probability, while personal, can be measured, by asking agents to "put their money where their mouth " is. The second part uses Gigerenzer's "natural frequency" method of communicating and thinking about uncertainty to explain inverse probability and Bayes theorem (no formulas please - only a "truth table" and, if you're a visual person, a graph, to develop "ballpark" boundedly rational assessments for inverse proabilities.)

Class cancelled - no lecture today

We start with discussion on uncertainty, probability, belief and expectation, then introduce the idea of contingent payoffs, a list of possibilities "if event A, the payoff B". Averaging out contingent payoffs with probability weighted averages. What does probability "mean"? - 3 interpretations of probability

We introduce the idea of a sub game perfect Nas Equilibrium analyzing the entry deterrence game as a simultaneous game. Looked at this way there turn out to be several Nash Equilibria...not all of which are "reasonable". The latter part of the class introduces a simple card game as we begin a new section on games of imperfect information.

In this lecture we start making connections: How do simultaneous game concepts relate to sequential game concepts? We analyze sequential games as simultaneous games, dig a bit deeper into the concept of Nash Equilibrium, then introduce information sets as a way of analyzing simultaneous games as sequential games

Simultaneous games, cont'd. After much confusion over the payoffs in the class example at the end of lecture 9, I introduce a slightly easier version of the minimum effort coordination game, and we spend most of the class walking through the analysis of this game and the problem of coordination ( eg in joint or team work) that it is designed to shed some (strategic) light on. Then a quickie, further extension of the Prisoner's dilemma to 3 players.

Simultaneous games, cont'd. We introduce "best response" type reasoning in simultaneous games, compare it with "dominance" reasoning, then use it to introduce the key idea of Nash Equilibrium. Then we use it to examine an important strategic tension in coordination games, starting with simple 2x2 coordination games: pure coordination, assurance, battle of the sexes, chicken. The class ends introducing the minimum effort coordination game.

Simultaneous games, cont'd. We analyze the idea of dominance reasoning and extend it from 2x2 to a more complicated voluntary contributions mechanism game (give all none or various levels in between) and to games where not every player has a dominant strategy, introducing a new idea: iterated elimination of suessively dominated strategies.

We change tack now and begin looking at simultaneous games. The lecture startes with a classroom game, the "voluntary contributions mechanism" [think supporting community projects via matching grants], then introduces a 2x2 payoff table to analyze this game and introduce dominance reasoning; finish with the classic prisoner's dilemma

Seq games cont'd (4th of 4 lectures): bargaining over a shrinking pie - the alternating offer bargaining game (classroom experiments & game tree analysis), the ultimatum game; using simple game tree's to analyze the strategic issues of lock-in and hold-up

In this lecture I discuss how to interpret, and how not to interpret, payoff numbers. The discussion uses the simple 2x2 game tree we constructed to analyze sequential games, but it applies equally well to payoffs in simultaneous games. Then we begin to change the games we look at, starting with changing the order of moves in a 2x2 game. A simpler type of game tree can be used to analyze games with your future self - especially whether or not to take mind and mood altering substances, from nicotine through to LSD and cocaine. Note, the Audio got corrupted (due to radio wave interference in a neighboring lecture theatre). It's 90% OK after editing but annoying in places.

Continuing on with our simple 2x2 sequential game we now examine in more detail the concept and identification of a "strategy", first for our simple game, then for more complex games..including the stop-go game from the first lecture.

This is the first of four lectures on sequential games. We start by talking a little bit about "theory", summed up in an acronym PDIP (who are the players, what can they do, what information/ignorance do they have, what are their payoffs). Then we look at how to analyze a simple 2x2 game using a neat graphical concept : the game tree.

More hands on games (and discussion of strategic reasoning) in class: the stop go game, the colour matching game , strategic games with yourself - well your future self.

First lecture of the term (starts in late Feb down under): I introduce you to some basic ideas of game theory, provide some illustrative examples, and play some simple strategic games (for chocolate bar payoffs).