NSU ECO 475 - Fall 2016 (NMh)

Games with Sequential Moves

 Sequential Games with perfect information: This happens when we model strategic interactions
with a strict order of play (like chess). Perfect information means players know everything that
has happened prior to making the decision.

 Draw Entry Game: (Refer to class notes or pdf file called sequential)

 Sequential move games are represented in extensive form using a game tree, which consists of
decision nodes, terminal nodes and branches. The game starts at the root of the game tree with
the initial decision node. At each decision node, a player chooses an action. Actions are shown as
branches emerging from a node and leading to another decision or terminal node. Each terminal
node (where no player has any action) is associated with an outcome of the game and payoffs for
the players.

 Player (or Agent): Someone who makes a choice within the game where that choice will affect
others.

 Action: An action is a choice a player can make in a game.

 Strategy: A complete plan of action. A complete strategy can be handed over to a representative
since every contingency is specified (even if the representative makes a mistake by accident).

 Outcome: An 'outcome' is defined in game theory as different from another if it has a different
'history'." This means that each distinct "play" (path from the initial point to an end point in the
tree representing the game in extensive form) leads, by definition, to a different "outcome." A
game has as many outcomes as terminal nodes.

 Payoff: Numbers associated with outcomes and they represent the motivation of players. It
captures everything a player cares about, including potential altruism and risk preferences.

 Equilibrium: A set of strategies such that each player is using her best strategy given the
strategies of all other players.

 Assumption of Rationality: Game theory assumes that all players behave rationally. Each player
has a consistent set of rankings (payoffs) over all logically possible outcomes and calculates her
best strategy. Two components of rationality are complete knowledge of personal interests and
flawless calculations (but not selfishness or identical value systems). In real world, are players
 truly rational?

 Backward Induction: We find the outcome of the game using backward induction. We find the
optimal choice at last decision node, and eliminate non-optimal actions. Subsequently, we find
optimal choice at previous decision node (actions now lead directly to terminal node). We
continue doing this until we reach the initial node of the game.

 Subgame: With perfect information, a subgame consists of a subset of the nodes and branches of
the original game that constitute a game themselves. Since the subgame consists of the subroot
and all its successors, it has a unique initial node called subroot. The subgame consist of the
subroot and all its successors. Every game is a (trivial) subgame of itself. How many subroots
does the draw-entry game have?

 Subgame Perfect Equilibrium (SPE): A set of strategies (one for each player) is a subgame
perfect equilibrium of a game, if this set of strategies is optimal in every subgame of the game.
This leads to the question: are there equilibria in a game which are not subgame perfect? There
are, but we can't find them using backward induction. In games with perfect information, the
backward induction process provides us the subgame perfect equilibrium only.

 If you are asked to find the subgame perfect equilibrium in a sequential game with perfect
information, carry out backward induction to find the outcome, and then write the equilibrium
strategy.

 There can be multiple subgame perfect equilibria in a game, even in the draw-entry game shown
above. You just have to play around with the payoffs at the terminal nodes.

 We can have more than two players. Let's do a three player sequential game with perfect
information. (Refer to class notes or pdf file named sequential)

 We can have nature as a special player to introduce random decisions. It is used to model
uncertainties in the game that do not arise from behavior of players or their lack of knowledge
about each other. Nature chooses between actions based on a fixed probability distribution, and
does not receive any payoff unlike real players. The players choose their action based on
expected payoff if they move before Nature.

 We need to learn how to make game trees (extensive form) from stories. Let's try.
The Park Board decides whether to open a large or small concession stand at the beach selling
ice cream. Nature decides on the weather, sunny or rainy, with equal probability. After observing
the weather, Giorgio decides whether to bring his ice cream cart to the beach or stay home. If the
weather is sunny, there are ice cream sales of 10 to be made. There are no sales if it is raining. If
 only one ice cream seller is open, it attracts all the sales. If both are open, the concession stand
gets 70% of the sales if it is large, and 40% if it is small. Giorgio has operating costs of 2 for his
ice cream cart, the small concession stand has operating costs of 2, the large stand has costs of 4.
(Refer to class notes or pdf file named sequential)

 Backward induction in reality: It is useful to explain actual outcomes in simple games with few
players and moves. The concept is less useful in complex situation move games such as chess,
where the game tree is too large to draw (Checkers is the largest game that has been solved
through backward induction - it took two decades and 1014 computations). People usually don't
look far ahead enough or fail to draw correct conclusion from looking far ahead.