Game Theory II: Axelrod's Tournaments - Gateway, Lecture 4

0.0(0)
studied byStudied by 0 people
GameKnowt Play
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
Card Sorting

1/30

flashcard set

Earn XP

Description and Tags

Vocabulary flashcards covering key terms, strategies, tournaments, and concepts from the Axelrod lecture notes.

Study Analytics
Name
Mastery
Learn
Test
Matching
Spaced

No study sessions yet.

31 Terms

1
New cards

ALL D

Defects every round, regardless of the opponent.

2
New cards

ALL C

Cooperates every round, regardless of the opponent.

3
New cards

RANDOM

A strategy that defects or cooperates with equal probability (50% each) each round.

4
New cards

Best response to ALL C

ALL D.

5
New cards

Best response to ALL D

ALL D again.

6
New cards

Best response to RANDOM

ALL D.

7
New cards

Robust strategy

A strategy that performs well across a range of opponents and conditions, not tied to a single scenario.

8
New cards

Axelrod's Tournament

A competition to find robust strategies in a 200-round Prisoner’s Dilemma by pitting submitted strategies against themselves, RANDOM, and all others.

9
New cards

Tournament format (1st tournament)

Experts submit strategies; each faces itself, RANDOM, and all others in a 200-round PD; highest final score wins.

10
New cards

1st Tournament Goal

Find a strategy that’s robustly successful.

11
New cards

DOWNING

Defects on unresponsive players; starts with a double defection.

12
New cards

FRIEDMAN

Starts by cooperating; if ever defected on, defects back forever.

13
New cards

TIT FOR TAT (TFT)

Cooperates initially and then copies the opponent’s last move; known for being nice and retaliatory.

14
New cards

JOSS

Like TFT, but with a 10% chance of defecting on a cooperator; performed worse than TFT.

15
New cards

TIT FOR TWO TATS

Defects back only after two defections by the opponent.

16
New cards

LOOK AHEAD

Strategy inspired by chess AIs; won the preliminary tournament described in advertisements.

17
New cards

REVISED DOWNING

Like DOWNING, but with added niceness (more forgiving).

18
New cards

ECOLOGY

Process where after every round, lowest scorers are replaced by copies of the highest scorers; iterated until stable.

19
New cards

The Rachet

It’s easier to increase cooperation than to decrease it.

20
New cards

w (continuation probability)

Probability of the game continuing to the next round; higher w emphasizes the future and the shadow of future payoffs.

21
New cards

Stability (solo)

ALL D is collectively stable; no single other strategy can invade; TFT cannot outperform ALL D in isolation.

22
New cards

Stability (clusters)

A cluster of TIT FOR TATs can take over because mutual cooperation within the cluster outweighs losses to natives (threshold around 5%).

23
New cards

The great asymmetry

Stability varies by population structure: ALL D can be stable against individuals but not against clusters; TFT is stable against clusters.

24
New cards

ALL D collectively stable?

Yes.

25
New cards

ALL D against clusters?

No.

26
New cards

TIT FOR TAT collectively stable?

Yes.

27
New cards

TIT FOR TAT against clusters?

Yes.

28
New cards

The PD (Prisoner’s Dilemma)

A two-player game where defection dominates, but cooperation can yield better joint outcomes; used extensively in Axelrod’s studies.

29
New cards

Equilibria

Strategic outcomes where no player benefits from unilaterally deviating; in PD, mutual defection is a classic equilibrium; repeated PD can yield other equilibria.

30
New cards

Pareto efficiency

An outcome where no one can be made better off without making someone else worse off.

31
New cards

State of nature (Hobbes)

A hypothetical pre-social condition described by Hobbes; life without government is often characterized as harsh; exit from it via covenants and civil society.