Why a Prisoner’s Dilemma Matters
In a strategic game, each decision-maker chooses an action while the payoff from that action depends partly on what others choose. Sometimes each person has a clear individual incentive to take an action that leads to a disappointing result for everyone. The prisoner’s dilemma is the standard model of this conflict: it shows how individually rational choices can lead to an outcome that both players rank below mutual cooperation.
Recognizing the dilemma requires more than noticing that the outcome seems poor. The key is to compare each player’s payoffs across the other player’s possible actions. A player’s preferred action must be the same whether the other player cooperates or defects. The payoffs must also show that both players would prefer mutual cooperation to the outcome produced by those incentives.
A payoff represents the player’s ranking or utility in the game; a larger payoff means that player prefers the outcome. It need not be a dollar amount. In an economic application it may represent profit, but it may also summarize several consequences. Payoffs should be compared player by player, not by adding both players’ payoffs.
Checking the Incentives
A strategy is strictly dominant for a player if it gives that player a strictly higher payoff than every alternative, for every action the other player could choose. In a standard prisoner’s dilemma, \(D\) is strictly dominant for both players. The payoff rankings provide the required comparisons: if the other player chooses \(C\), defecting yields \(T\) rather than \(R\); if the other chooses \(D\), defecting yields \(P\) rather than \(S\).
When both players have the same strictly dominant action, the outcome in which both choose that action is called the dominant-strategy equilibrium. For a strict prisoner’s dilemma, that outcome is \((D,D)\). Yet \(R>P\), so each player would rather receive the mutual-cooperation payoff than the mutual-defection payoff. The tension is not a failure to understand the payoffs: each player has an individual incentive to defect even though both would prefer the result of mutual cooperation.
Why the Outcome Follows
The logic can be derived directly from the payoff comparisons. Suppose the row player chooses first in the matrix layout and the column player’s action determines which column applies. If defecting gives the row player more than cooperating in each column, then the row player’s choice is \(D\) regardless of the column player’s decision. The same reasoning applies to the column player across the two rows. Neither needs to predict the other’s action in order to identify their own strictly dominant action.
In the symmetric payoff notation, the two relevant comparisons for either player are \(T>R\) when the other chooses \(C\), and \(P>S\) when the other chooses \(D\). Therefore each chooses \(D\), producing \((D,D)\). The separate inequality \(R>P\) establishes the cooperation conflict: both players’ payoffs at \((C,C)\) exceed their payoffs at \((D,D)\). This conclusion uses only payoff comparisons; it does not require the players’ payoffs to be measured in money.
In every payoff pair, the first number belongs to the row player and the second to the column player.
For each player, compare \(C\) with \(D\) once for each action the opponent might choose.
If \(D\) yields a strictly higher payoff in both comparisons for both players, both choose \(D\).
Compare each player’s payoff at \((C,C)\) with that same player’s payoff at \((D,D)\). The dilemma requires both to prefer \((C,C)\).
Worked Examples
Worked Example: A Shared Walking Path
Two neighborhood associations can either contribute to maintaining a shared walking path (\(C\)) or withhold their contributions (\(D\)). A hypothetical payoff assessment, in utility points, gives the following matrix. The first payoff in each cell belongs to Association A.
| Association B: C | Association B: D | |
|---|---|---|
| Association A: C | (6, 6) | (2, 9) |
| Association A: D | (9, 2) | (3, 3) |
Method: compare each association’s own payoff across the relevant alternatives. If B contributes, A receives 6 by contributing and 9 by withholding; A prefers \(D\), since \(9>6\). If B withholds, A receives 2 by contributing and 3 by withholding; A again prefers \(D\), since \(3>2\). Thus \(D\) is strictly dominant for A.
For B, if A contributes, B receives 6 by contributing and 9 by withholding, so B prefers \(D\), since \(9>6\). If A withholds, B receives 2 by contributing and 3 by withholding, so B again prefers \(D\), since \(3>2\). The dominant-strategy outcome is therefore \((D,D)\), with payoffs \((3,3)\).
At mutual contribution, the payoff is \((6,6)\), and \(6>3\) for each association. The outcome is a strict prisoner’s dilemma: each association has an individual incentive to withhold in either case, even though both would receive higher payoffs if both contributed. Common mistake avoided: the decision is not made by asking which action gives the larger combined payoff. It is made by comparing each association’s own payoff for each possible action of the other.
Worked Example: Different Payoffs for Two Firms
Two hypothetical firms decide whether to adopt a shared inventory-reporting standard (\(C\)) or to skip adoption (\(D\)). Their payoffs are annual net benefits in thousands of dollars. The game is not symmetric: the payoff amounts differ between the firms.
| Firm B: C | Firm B: D | |
|---|---|---|
| Firm A: C | (7, 8) | (2, 11) |
| Firm A: D | (10, 3) | (4, 5) |
Method: test dominance for each firm without assuming equal payoffs. If B adopts, A compares 7 from adoption with 10 from skipping, so A chooses \(D\) because \(10>7\). If B skips, A compares 2 from adoption with 4 from skipping, so A still chooses \(D\) because \(4>2\).
If A adopts, B compares 8 from adoption with 11 from skipping, so B chooses \(D\) because \(11>8\). If A skips, B compares 3 from adoption with 5 from skipping, so B again chooses \(D\) because \(5>3\). Both firms therefore have a strictly dominant action \(D\), and the resulting payoffs are \((4,5)\).
At \((C,C)\), Firm A receives 7 rather than 4, and Firm B receives 8 rather than 5. Each firm would prefer mutual adoption to the dominant-strategy outcome. The example has the key prisoner’s-dilemma incentives even though the payoffs are not identical: the crucial comparisons concern each player’s own payoffs, not equality between the players’ numbers. Common mistake avoided: assuming a prisoner’s dilemma must be numerically symmetric. Symmetry is part of the standard textbook version, but the incentive conflict can also be checked with player-specific payoff comparisons.
Worked Example: Funding a Shared Flood Barrier
Two neighboring municipalities decide whether to contribute to a shared flood barrier. Each municipality receives a benefit of 10 utility units if both contribute, at a contribution cost of 6 units per municipality. If only one contributes, the barrier produces a benefit of 5 units for each municipality; the contributor pays the cost of 6, while the non-contributor pays nothing. If neither contributes, neither receives a benefit. Calculate the payoffs and determine whether this is a prisoner’s dilemma.
Method: calculate each payoff as benefit minus that municipality’s own cost. If both contribute, each receives \(10-6=4\), giving \((C,C)=(4,4)\). If A contributes and B does not, A receives \(5-6=-1\), while B receives \(5-0=5\), giving \((C,D)=(-1,5)\). If B contributes and A does not, A receives \(5-0=5\) and B receives \(5-6=-1\), giving \((D,C)=(5,-1)\). If neither contributes, each receives 0, so \((D,D)=(0,0)\).
| Municipality B: C | Municipality B: D | |
|---|---|---|
| Municipality A: C | (4, 4) | (-1, 5) |
| Municipality A: D | (5, -1) | (0, 0) |
If B contributes, A compares 4 from contributing with 5 from not contributing, and chooses \(D\) because \(5>4\). If B does not contribute, A compares \(-1\) from contributing with 0 from not contributing, and again chooses \(D\) because \(0>-1\). The same comparisons hold for B by symmetry. Thus both municipalities have \(D\) as a strictly dominant action, producing \((D,D)=(0,0)\).
Mutual contribution gives each municipality 4, which exceeds 0. The payoff ranking is \(T=5>R=4>P=0>S=-1\), so this is a strict prisoner’s dilemma. The cost calculation matters: using the benefit of the barrier as the contributor’s payoff while omitting the contribution cost would change the matrix and could lead to an incorrect classification. The calculations check: a sole contributor receives \(5-6=-1\), while a non-contributor facing that contribution receives 5.
Common Exam Traps
A frequent error is to add the two payoffs and choose the action profile with the largest total. For the walking-path matrix, mutual contribution has total payoff \(6+6=12\), while mutual non-contribution has total \(3+3=6\). That comparison shows that the total is higher under mutual contribution, but it does not establish that either association has an incentive to contribute. Against contribution by the other association, each gets 9 from withholding rather than 6 from contributing. Against withholding, each gets 3 from withholding rather than 2 from contributing. The correct individual comparisons show that both choose \(D\).
A second trap is to check only whether one player can gain by defecting when the other cooperates. That comparison alone is not enough to establish dominance: the player must also prefer defecting when the other defects. A third is to call any game with mutual defection a prisoner’s dilemma. The mutual-cooperation outcome must also be preferred by both players to mutual defection, and the specified actions must have the required incentive rankings.
Check Your Understanding
For each problem, compare the payoffs belonging to each player rather than adding payoffs across players.
- A symmetric game has \(T=14\), \(R=9\), \(P=6\), and \(S=1\). Is it a strict prisoner’s dilemma, and what is its dominant-strategy outcome? (Answer: Yes; \(14>9>6>1\), so \(D\) is strictly dominant for both players and the outcome is \((D,D)\).)
- A symmetric game has \((C,C)=(5,5)\), \((D,C)=(8,2)\), \((C,D)=(2,8)\), and \((D,D)=(4,4)\). Identify the ranking and classify the game. (Answer: \(T=8>R=5>P=4>S=2\); it is a strict prisoner’s dilemma, with dominant-strategy outcome \((D,D)\).)
- A symmetric game has \((C,C)=(6,6)\), \((D,C)=(9,1)\), \((C,D)=(1,9)\), and \((D,D)=(7,7)\). Does it have dominant defection, and is it a prisoner’s dilemma? (Answer: Yes, \(D\) is strictly dominant for both, but no, it is not a prisoner’s dilemma because each player gets 6 at \((C,C)\), less than 7 at \((D,D)\).)
- An asymmetric game has \((C,C)=(3,6)\), \((D,C)=(7,2)\), \((C,D)=(1,9)\), and \((D,D)=(4,5)\). Do both players have dominant defection, and does mutual cooperation give both a higher payoff than mutual defection? (Answer: Yes, both have strictly dominant \(D\); no, because Player A gets 3 at \((C,C)\) and 4 at \((D,D)\), so it is not a prisoner’s dilemma.)
- For a symmetric game, a sole defector receives 12, mutual cooperation pays 10 to each player, mutual defection pays 5 to each, and a sole cooperator receives 4. State the ranking and the outcome. (Answer: \(T=12>R=10>P=5>S=4\); this is a strict prisoner’s dilemma, and the dominant-strategy outcome is \((D,D)\).)