Concept/Game Theory/No. 1102
Zero-Sum vs. Non-Zero-Sum
Zero-sum vs. non-zero-sum is a game theory distinction developed by John von Neumann and Oskar Morgenstern. In zero-sum games, players’ gains exactly offset their losses; in non-zero-sum games, total payoffs can change, so choices may create or destroy value for all players.
- Evidence
- Well established
- Read
- 6 min
- Links
- 10 connections
01You've seen this when…
- in life
You and your roommate both dread cleaning. You don’t mind laundry; they don’t mind vacuuming. Trading chores makes the same apartment less burdensome for both.
- at work
The finance director sets a fixed training budget. Every dollar sales gets is a dollar support cannot spend.
- out in the world
Drivers enter an intersection even though there’s no space to exit. Each tries to get ahead, but traffic locks up and everyone waits longer.
02The idea
Two departments can fight over a fixed budget without changing its size. A buyer and a supplier can fight over price while overlooking a delivery change that saves money. Both situations involve competing interests, but they have different structures.
In a zero-sum game, the players’ payoffs add to zero in every possible outcome. One person’s gain is exactly balanced by someone else’s loss. There is no choice that makes everyone better off together.
A fixed pot of money is technically constant-sum: the amounts received always add to the same total, which can differ from zero. Measure each person’s gain or loss against a starting allocation, and those changes add to zero. For strategic purposes, constant-sum games have the same strictly competitive structure.
In the usual practical contrast, a non-zero-sum game is one where the combined payoff can change. Choices can create or destroy value instead of merely transferring it. Mutual gains or mutual losses may be possible, though neither is guaranteed.
The useful question is what counts as a payoff. Dollars, profits, time and satisfaction are not interchangeable. Dividing a fixed cash prize is constant-sum in dollars. Dividing household tasks need not be constant-sum in effort or enjoyment.
Many real interactions contain both elements: first, choices about creating value; then, conflict over who receives it.
03Why it matters
- It changes what you should search for. If the only decision is how to split a fixed payment, concessions transfer money. That’s distributive bargaining. If timing, workload, quality or risk can change, look for arrangements that improve both sides’ position. That’s integrative bargaining.
- It separates competition from destruction. Hurting a rival can also hurt you. Two businesses can spend heavily to take customers from each other while reducing both firms’ profits. A policy dispute can leave every participant worse off if it prevents any workable agreement.
- It explains why cooperation can be difficult even when it helps. In the prisoner’s dilemma, each player has an incentive to defect even though mutual cooperation would leave both better off than mutual defection.
- Finding shared value and arranging incentives to capture it are different jobs. Individual incentives in a non-zero-sum interaction can favor uncooperative behavior.
This distinction also explains gains from trade: exchanging things can benefit both parties when they value those things differently. The exchange can increase value while leaving the number of objects unchanged.
04A worked example
Consider an invented negotiation between a manufacturer and a supplier. All figures are measured against making no deal, which gives each side a payoff of zero.
The manufacturer values an order delivered on Monday at $60,000. Producing it for Monday costs the supplier $40,000. At a price of $50,000, each side gains $10,000.
What it looks like A fight over price. Reducing the price by $1,000 gives the manufacturer another $1,000 and takes exactly $1,000 from the supplier. With everything else fixed, their combined gain stays at $20,000.
What’s actually going on Monday delivery requires overtime. Delivering on Thursday reduces the supplier’s cost to $35,000. The delay costs the manufacturer $1,000, so its value falls to $59,000. The available combined gain becomes $24,000: a $5,000 saving minus a $1,000 inconvenience.
At a new price of $48,000, the manufacturer gains $11,000 and the supplier gains $13,000. Both do better than under the original arrangement. The delivery decision creates value; the price decision divides it.
What would have helped Identifying which terms were fixed before negotiating price. The supplier needed to reveal the overtime cost, and the manufacturer needed to reveal how little the delay mattered. Neither fact would surface if both treated the entire negotiation as a contest over one number.
05Where people trip up
- Zero-sum is a property of every possible outcome. A single observed outcome with a winner and a loser leaves open whether the game offered opportunities for mutual gain.
- A scarce resource can be part of a non-zero-sum interaction. There are only so many hours in a day, but better scheduling can reduce waiting and duplicated work. Scarcity constrains what is possible; the payoff structure determines whether every gain requires an equal loss.
- Non-zero-sum interactions can still involve conflict. After creating another $4,000 of value, the manufacturer and supplier can still disagree fiercely about its division. Cooperation and competition can occur in the same deal.
- A shared benefit can hide someone else’s loss. A factory and its customer may both benefit from cheaper production that pollutes a river. Include the people bearing that externality before calling the arrangement a general improvement.
- Zero-sum games and zero-sum bias are distinct. Zero-sum bias is the mistaken assumption that someone’s gain must be someone else’s loss. Zero-sum games exist. Correctly recognizing one is sound reasoning.
06Where it doesn’t settle the deal
Creating more total value is only part of the case for an agreement. The people giving something up may go uncompensated. Some losses cannot be compensated with money, and some benefits cannot be transferred between participants.
A Pareto improvement makes at least one person better off without making anyone worse off. An increase in the total is a weaker claim: it can leave particular people substantially worse off. Keep those claims separate.
Even mutual gains depend on the comparison. An agreement may beat the current offer but lose to one party’s best alternative. Assess the deal against what each participant could actually do instead. An artificially bad fallback distorts the comparison.
07Roots
John von Neumann’s 1928 paper tackled a problem ordinary optimization leaves out: your opponent gets to choose too. In a two-player zero-sum contest, a move that works against one response may fail against another. His minimax theorem showed that, when players can randomize among their moves, there is a game value one player can guarantee and the other can prevent them from exceeding. Mixed strategies made that guarantee possible.
The economist Oskar Morgenstern wanted a theory that took this mutual dependence seriously. A firm’s best decision depends on what other firms do; it cannot always treat their behavior like fixed background conditions. Working with von Neumann at Princeton, he helped turn strategic interaction into the subject of their 1944 book, Theory of Games and Economic Behavior. It made zero-sum analysis foundational while also examining economic exchange and coalitions.
John Nash then developed tools for a wider range of interactions. His 1950 bargaining paper addressed how parties might select an agreement from possibilities that could benefit both. His work on noncooperative equilibrium extended analysis beyond two-player zero-sum contests. The enduring lesson was that interactions have different payoff structures, and those structures determine which methods can explain them.
08How solid is this?
This is a formal distinction in game theory, supported by mathematical results rather than a psychological effect. Applying it requires specifying the players, possible outcomes and payoffs; mutual gains in a model do not guarantee cooperation in practice.
09Connections
- Often confused withZero-Sum Bias
- IncludesDistributive Bargaining, Prisoner’s Dilemma, Gains from Trade
- See alsoIntegrative Bargaining, BATNA, Pareto Efficiency, Externality, Mixed Strategy, Nash Equilibrium
10Origin and sources
John von Neumann established the two-player zero-sum minimax theorem in 1928. His work with Oskar Morgenstern in Theory of Games and Economic Behavior (1944) made the distinction foundational to modern game theory.
- [1]von Neumann, J. (1928). Zur Theorie der Gesellschaftsspiele. Mathematische Annalen, 100, 295–320.
- [2]von Neumann, J., & Morgenstern, O. (1944). Theory of Games and Economic Behavior. Princeton University Press.
- [3]Nash, J. F. (1950). The Bargaining Problem. Econometrica, 18(2), 155–162.
- [4]Nash, J. (1951). Non-Cooperative Games. Annals of Mathematics, 54(2), 286–295.
Suggest an edit· Updated 2026-10-02