An inmate in a Colombian penitentiary first told me about the Billiards War. I’ll call him Carlos. Lean, muscular, in his late twenties, he had run a drug corner for a Medellín street gang before his arrest.

If we take our eyes off the roots of war and focus on misleading causes, we are unlikely to make the choices that will promote peace.

A powerful criminal group called Pachelly ruled Carlos’s cellblock, selling drugs and charging inmates for beds and phones. Revenues from these activities made the prison hallways profitable territory, just like the streets Pachelly controlled on the outside.

A rival gang, El Mesa, lived in the same prison block. Their power was growing outside of the prison’s walls, and they had begun to resent Pachelly’s rule. Tensions between the two groups came to a head one afternoon as men from both gangs were playing billiards. Though Carlos cannot recall why they started arguing, he does remember that the fight got out of hand quickly. Members of El Mesa pulled out their guns and fired. Miraculously no one was killed, but twenty-three were injured in the shootout.

Over the next few days, the prison saw a series of retaliatory attacks. Soon anger and recrimination spilled outside of the penitentiary. Pachelly and El Mesa activated their alliances. Hundreds of gangs lined up on either side. Medellín’s underworld geared for war.

Over the following weeks, however, the city never spiraled into bloodshed. The Billiards War never happened. Instead, tense negotiation ensued. Pachelly ceded some territory—a prison hallway here, a street corner there. None of these territories were worth a battle.

The reason was simple: violence seemed too costly. Fighting kills little brothers and friends, and no one wants to pay protection money or buy drugs in the middle of a gunfight. Most of all, though, a war between the gangs would have brought police attention to the crime bosses and risked their arrest. These leaders could not care less about civilian casualties. Mass violence would have compromised leaders’ bottom line and their freedom.

So Pachelly and El Mesa compromised, negotiating an agreement to avoid violence: El Pacto del Fusil, or the Pact of the Machine Gun. Some things have not changed, of course. The gangs still despise one another. They maneuver for drug corners and prison hallways. They occasionally have skirmishes (where things can get pretty rough). But they steer clear of war, knowing that it would result in vastly more damage. Today their pact has been held for a decade and Medellín’s homicide rate is nearly half that of U.S. cities such as Chicago.


Medellín’s hostile peace is not unusual. Indeed, its gangs are an allegory for our wider world. The globe is a patchwork of rival territories. Controlling them brings wealth and status. Different groups covet their neighbors’ resources and prey on the weak, but most do their best not to wage war.

Now, when I say war, I don’t just mean countries duking it out. I mean any kind of prolonged, violent struggle between groups. This includes villages, clans, gangs, ethnic groups, religious sects, political factions, and nations. At every one of these levels, compromise typically wins for the same reason it does in Medellín: war is ruinous.

Nothing destroys progress like conflict. Fighting massacres soldiers, ravages civilians, starves cities, plunders stores, disrupts trade, demolishes industry, and bankrupts governments. It undermines economic growth in indirect ways too. Most people and business won’t do the basic activities that lead to development when they expect bombings, ethnic cleansing, or arbitrary justice; they won’t specialize in tasks, trade, invest their wealth, or develop new ideas. These costs of war incentivize rivals to steer clear from prolonged and intense violence.

Of course, it seldom feels like peace is our natural state. “The story of the human race is war,” said Winston Churchill, “except for brief and precarious interludes, there has never been peace in the world.” Certainly it often seems so, especially today as a major conflict rages in Ukraine and the number of civil wars in the world climbs to levels not seen since the 1990s.

But that sentiment is misleading and comes from ignoring the quieter moments of compromise. One example of peace came two weeks after Russia’s invasion of Ukraine, when India accidentally launched a cruise missile at Pakistan. Pakistan was outraged, but violence did not break out. Calm ensued, as it has for decades. War would have been so costly that both sides strove to avoid it.

Likewise, for two decades, President Vladimir Putin tried every means possible to co-opt Ukraine without invasion: dark money, propaganda, political stooges, assassinations, and support for separatists. War was his last resort. He succeeded in subjugating neighbors such as Belarus and Kazakhstan without a major fight, and he hoped to repeat the trick in Ukraine.

Look around and tense instances of peace appear everywhere, whether it’s the gloomy impasse between North and South Korea or the constant deadlock over Taiwan. We see it at lower levels too. Take ethnic and sectarian groups, who are mostly covered in the news when they purge or pillage their neighbors. When social scientists tallied the number of rivalries in the world and counted how many turned violent, however, they found that fewer than one in a thousand actually fight.

We forget this because few people write about the conflicts avoided. Instead, our gaze is often pulled to the horrific, violent struggles that do happen. That’s natural. We have to pay attention to wars, but we shouldn’t focus solely on the hostilities that do occur.

Solely examining the times that peace failed is a kind of selection bias—a slanted accounting of the evidence that distorts our view of the world. It is a mistake that has two important consequences. One is that we exaggerate how much we fight. Yes, our natural condition is conflict and competition, but most of the time this jostling is not violent. Enemies prefer to loathe in peace.

The second and greater harm of selection bias is that we mistake the roots of war and follow fruitless paths to peace. When we trace the steps leading up to a fight, we find familiar features: ancient hatreds, poverty, historical injustices, and an abundance of arms. But when we look at the times rivals did not fight, we often see many of the same preceding conditions. We ought to be skeptical, then, that these are the causes of the violence. To discern why we fight wars, we should begin with why we do not. 


Most nations, political factions, ethnic groups, and gangs behave strategically. Like players of poker or chess, they try to think ahead, discern their opponents’ strength and plans, and choose their actions based on what they expect them to do. Even when they’re poorly informed, mistaken, biased, or impassioned, they still want to attain the best outcome. Strategic behavior is thus a good starting point for thinking about war.

Solely examining the times that peace failed is a kind of selection bias—a slanted accounting of the evidence that distorts our view of the world.

The science of strategy is game theory, which works out how one side will behave based on what it believes its opponent will do (knowing that the opponent is thinking in the same way). Consider the decision facing Pachelly and El Mesa in the aftermath of the billiards shootout. The foes were evenly matched, and so each gang had a roughly fifty–fifty chance of winning a winner-takes-all war. The two rivals also knew that war would have dire consequences, regardless of the winner.

Let’s suppose that the territory they contested was worth $100, and that both gangs expected fighting to cost one-fifth of that total, $20. War’s destruction means that both sides are better off finding a peaceful split than going to war. The $20 is a peace bonus they get to divide. It creates a whole range of territorial splits both sides prefer to fighting it out. Consider Pachelly. Fighting gives them even odds for gaining a shrunken $80 prize. That means that any peaceful split that leaves Pachelly more than 40 percent of the $100 territory is a better deal. The same is true of El Mesa, and both sides know it. The $20 that will be saved by not fighting gives the rivals an array of acceptable compromises—what political scientists and economists call a bargaining range.

This strategic insight is decades old. The early applications were not to military conflict, but to commerce. Take negotiations between firms and unions, for instance. Just replace “war” with “labor strike,” and the result is the same. Both groups want the best deal for their group—shareholders and management on one side, workers on the other. Strikes and stoppages are costly to both sides. Thus, most firms and unions try to avoid them. Or consider court battles. Lawsuits are expensive and inefficient. It’s better to settle, and most litigants do. Long, messy court battles happen only when something hijacks the normal incentives for settlement.

Now we should not use this game theory blindly. Some employ these models to paint a picture of humans as unreasonably rational—Homo economicus—but this species still manages to commit an awful lot of violence. Groups and their leaders are not always logical or all-seeing, and  people rarely hold coherent beliefs that the body politic faithfully represents. Still, game theory serves as a base frame of reference—not because people are necessarily peace-loving, but simply because they are self-interested. The losses from violent conflict usually outweigh the gains.

The five reasons for war

This allows a new perspective on violence. If societies avoid war because it is ruinous, then every explanation for fighting is the same: a society or its leaders either ignore the costs of war or choose to pay them. Whether it is Russia’s attack on Ukraine, the U.S. invasions of Afghanistan and Iraq, or the gang wars that have (on occasion) turned Medellín into the most dangerous place on the planet, something overcame the usual strategic incentives for peace, pushing opponents away from nonviolent politics and toward bloodshed.

There are five main reasons incentives for peace break down: unchecked interests, intangible incentives, misperceptions, uncertainty, and commitment problems.

And while there’s a reason for every war and a war for every reason, there are only so many logical ways that the incentives for peace can break down. There are five main reasons it happens: unchecked interests, intangible incentives, misperceptions, uncertainty, and commitment problems. Each eliminates the incentives for compromise in a distinct way. To explain them, I’ll step away from the gangs of Medellín and consider the great conflict of this moment: Russia’s invasion of Ukraine.

1. Unchecked interests

Like gang leaders, autocrats and oligarchs can ignore many of the costs of war because the soldiers and citizens who pay them cannot hold their leaders to account. Think of Putin. Dictators like him must only heed the damage and risks to themselves and their cabal—a fraction of the total costs. They are quick to use violence because others pay much of the price.

Unchecked leaders are also free to pursue their private interests, and this can lead them to go to war against their society’s interests. Again, consider Putin. Ukrainians had tossed out two Russian-facing leaders in the previous two decades. But while an increasingly restive and democratic Ukraine was hardly a danger to ordinary Russians, from Putin’s point of view, Ukrainian democracy was a dangerous example to dissidents at home, potentially threatening his system of control. Co-opting the regime and installing a puppet government could eliminate that threat. War was not in the interest of Russia at large; it was in the interest of Putin’s narrow regime.

Democratic societies are vulnerable to unchecked interests as well, such as a president who hopes to boost popularity ahead of an election, military leaders who see opportunity in a conflict, or extremist splinter factions who want to spoil the peace. Each of these examples involves decision-makers who are imperfectly accountable to the populace. Dictators are the most extreme and dangerous of the lot, because they are accountable to the fewest people and bear the least costs.

2. Intangible incentives

Sometimes violence can deliver value, such as glory, renown, or justice. Any ideological rewards might offset the costs of war, spurring sides to fight instead of bargain.

Consider status. A desire for glory and dominance has driven kings of the past and tyrants of the present to warfare. This is probably the most common explanation for Putin’s invasion: nationalist pride and a desire to see Russia restored to its imperial glory. If it adds to Putin’s personal renown and place in history, so much the better.

Our natural condition is conflict and competition, but most of the time this jostling is not violent. Enemies prefer to loathe in peace.

This sounds like a case of unchecked interests, and it may be true that Putin prizes glory most of all. But he is not alone; at least some Russians support him. Unaccountable leaders are not the problem if the populace shares nationalistic and ideological goals. When they do, the leader is faithfully representing the group’s ideals by going to war. Most commonly, however, rulers and their cabals seek personal and national glory.

Another example of intangible incentives is the value societies place on certain pieces of land. When rival street gangs split territory in the example above, their interests were mostly mercenary. The neighborhoods had no sacred value. The soil wasn’t tied to their identity. But there are other times when a territory is hallowed. In those cases many fight because the idea of compromising over a sacred space or ethnic homeland is unfathomable.

Similarly, sometimes fighting is a society’s only path to achieving other righteous ends. Throughout history colonized and repressed peoples have decided to fight for their freedom because compromise was abhorrent. One example is the American Revolution. The colonists rebelled against a tyrannical superpower because they did not want the semi-sovereignty on offer (even if that is all their military weakness could reasonably earn them).

This is possibly the most important and least talked about explanation for the conflict in Ukraine. Putin demanded Ukraine sacrifice its sovereignty, a price other neighbors have paid. But like the U.S. revolutionaries over two centuries before, Ukrainians refused the bargain. In both cases they were willing to bear some costs of war because compromise on freedom and sovereignty was simply repugnant.

3. Misperceptions

A third way rivals get the costs and benefits of war wrong is when they form and hold mistaken views, even when evidence piles up to the contrary. In this scenario decision-makers do not stop acting strategically, but rather strategize from a set of delusional and biased beliefs.

We often misperceive others. Humans are prone to demonizing their enemies and attributing to them the worst intentions. We often treat opponents too harshly—repressing protesters with deadly force, striking insurgent leaders with imprecise drones—and then wonder why they angrily reject the miserable deal on offer. People are especially prone to these errors when the conflict follows ethnic and religious cleavages: the troubles in Northern Ireland, the civil war in Syria, or the U.S. war on fundamentalist Islamic armed groups.

Other times we misperceive ourselves, overestimating chances of victory or underestimating the costs of a fight. Sometimes we owe this overconfidence to psychological biases. Other times it arises from an institutional problem that insulates decision-makers from bad news.

When social scientists tallied the number of rivalries in the world and counted how many turned violent, they found that fewer than one in a thousand actually fight.

The best evidence of overconfidence comes not from politics—where the assurance of leaders is hard to pin down—but rather from business. Whether it is mutual fund managers, CEOs, or entrepreneurs, economists have shown that many are prone to repeating the same overconfident mistakes—repeatedly overestimating the success of a merger or investment strategy.

Political leaders could overestimate their chances of victory, underestimate the costs, and be too quick to take their society to war. This is one of the most common charges laid against the Bush administration’s invasions of Afghanistan and Iraq. It is also one of the most familiar narratives for Russia’s invasion of Ukraine: an isolated and insulated ruling cabal overconfidently expected to easily topple the Zelensky regime, so much so that they were unprepared for the possibility of failure.

4. Uncertainty

Misperceptions are important, but generally speaking, people attribute too much influence to mistakes, and too little to uncertainty. After all, Russian leaders weren’t the only people surprised by Ukraine’s military resistance. Most leaders simply do not know the strength or resolve of the other side. The fourth root of war is this uncertainty.

The first months of Russia’s invasion of Ukraine showed how hard it is to estimate the probability of victory or the costs of violent struggle. The competence of Russia’s forces, resolve of the Ukrainian people, and the west’s sweeping sanctions were all fundamentally hard to judge in advance. While it was always possible that Russia would get a bad draw on all three, almost no politician or military expert predicted it—least of all Putin. The problem was not simply overconfidence, it was also the tremendous unpredictability of the situation.

Granted, war is so ruinous that both sides ought to invest enormous time and energy in discerning their rival’s strength and resolve. And they do—this is the function of diplomacy and intelligence. Moreover, no nation wants to be mistaken as weak. To avoid attack, most rivals signal their true strength through military exercises, missile tests, and (if necessary) skirmishes. In many cases, however, an opponent’s strength is hard to determine. Arguably neither Ukraine nor Russia knew their own strength or resolve before testing themselves on the battlefield. That uncertainty makes war a gamble.

More importantly, it is hard to trust any signals from a rival. While no nation wants to be underestimated, everyone would like to be overestimated. If you have played a game of poker, you grasp the game theory already. You don’t know what cards your opponent holds, but you do know that the uncertainty gives them an incentive to bluff. Your best response is not to believe them and fold every time.

The economic costs of war incentivize rivals to steer clear from prolonged and intense violence.

Likewise, in war, enemies do not know the other’s strength or resolve. They may be bluffing. The optimal approach is to play what is called a mixed strategy: occasionally you ought to fold; occasionally you ought to call (risking war). Each time is a gamble. If uncertainty explains why wars break out, it also helps to explain why the average war is so brief—typically counted in weeks rather than years. Once the uncertainty dissipates, each side prefers to bargain rather than fight.

5. Commitment problems

Commitment problems comprise the fifth and final strategic dilemma—one that lies behind many of history’s longest wars. Commitment problems arise whenever one side believes its opponent has an incentive to renege on a peace deal—to use some future advantage to attack. Knowing this, a deal unravels before it can even begin.

A classic example is the preventative war, where one side expects its adversary to become more powerful in the future and renegotiate the deal in their favor. One side capitalizes on its ability to strike while still strong. When neither side can commit to future peace, a quintessential commitment problem ensues—both sides would prefer a deal that avoids the ruin of war, but that bargain is not always credible. Historians and political scientists use this dilemma to explain conflicts as varied as the Peloponnesian War, World War I, and the U.S. invasion of Iraq.

Commitment problems also arise when leaders are unable to make credible deals. Both democrats and autocrats face this challenge, though for different reasons. Elected presidents, such as Ukraine’s Zelensky, can agree to peace terms, but a year down the road, should circumstances change, a legislature could refuse to ratify the agreement, or citizens could elect a leader who rejects the previous terms. Again, a deal unravels before it begins. Meanwhile, dictators such as Putin have even greater difficulty making deals because nothing constrains them from changing their mind later. Why should Ukraine agree to peace terms if they worry that Russia might use any pause to regroup and renew its strength, only to attack again? Part of the problem here is an autocrat’s inability to credibly keep their word.


These five roots of war provide a diagnostic tool to discipline our thinking. Every time a conflict is explained by reference to a specific factor, we can use this lens: “How does that factor obscure or override the incentives for peace? How does it fit into the five?” It may not. There are many misleading ideas about war that arise from focusing on the failures and attributing the conflict to false causes.

Take poverty, which most people associate with a risk of conflict. War is expensive, though, and poor societies often have even more to lose. This is why sudden financial crises, price shocks, and droughts seldom lead to war.

Another common mistake is to blame inept and ideological leaders alone. This is the dominant account of war after war: World War I, the U.S. invasion of Iraq, or Putin’s attack on Ukraine. But in all these cases a set of subtler strategic factors—unchecked political systems, uncertainty, and commitment problems—narrowed the range of peaceful settlements to a sliver, to the point that glory-seeking, overconfident leaders and chance events could take their societies to war. We need both the psychological and strategic explanations if we are to understand why fighting happens.

For the same reason, we must resist the allure of chance events. History books on wars are filled with random human foibles, economic tumult, natural disasters, lucky coups, new technologies, and assassinations. These play a role, but perturbations matter for war and peace only when the five fundamentals have left rivals with little room to maneuver. Peace, in contrast, is what happens when societies are insulated from the five roots of war and can find ways to compete without violence.

The paths to peace

“Peace is not the absence of conflict,” Ronald Reagan once said, “it is the ability to handle conflict by peaceful means.” Indeed, successful gangs, ethnic groups, cities, and states all share the same strategy: they develop institutions and interventions that pay attention to the fundamentals and seek to minimize all five kinds of breakdown. They check the power of centralized leaders. They share information. They make commitments easier to hold. In short, they compel rivals to consider the costs of war and to make peace strategic.

Consider the gangs of Medellín. Long before the Billiards War, they built a governing board called La Oficina—The Office—to manage these kinds of disputes. After the shootout in Bellavista prison, the kingpins that dominated La Oficina sat El Mesa and Pachelly down to negotiate, to clear up any uncertainty and misperceptions. Both sides wanted revenge and spoils, but La Oficina countered those unchecked interests with powerful threats, promising to sanction anyone who continued to break the peace. The kingpins lorded over an unequal and illegitimate system, but to some degree it has worked. Medellín remains a dangerous place, but its murder rates are far lower than those of most U.S. cities.

When a society fails to contain violence, it is usually because it has allowed one or more of the five logics of war to flourish.

Once again, the gangs of Medellín are an analogy for our wider world. We too have kingpins and organizations that steer rivals from violence—the member states of the UN Security Council or NATO, for instance, and the institutions and tolls they have established, from peacekeepers to sanctions regimes.

Like La Oficina, these institutions are imperfect. The promise of sanctions was clearly not enough to deter Putin from his prize. And, also like La Oficina, our tools are unequal. International institutions work poorly against great powers, such as the United States or Russia. (For the same reasons, La Oficina is powerless to constrain a major kingpin. When one of the great powers of Medellín decides to fight, the city becomes the most violent place on the planet.) Still, the evidence suggests our international institutions and tools work, to a degree. The world is a more peaceful place with a general assembly, peacekeeping forces, sanctions regimes, and ranks of mediators than it would be without.

Nations and cities have been even more successful, as they have built social and legal institutions that counter the five roots of violence. Their states and social norms sanction violent offenders. Their political systems check the power of elites and compel them to consider the interest of wider groups. Their institutions share information and dispel fake news and other misperceptions. Everyday work and social life integrate one group with another, giving them shared interests and additional incentives not to fight. And their ideologies emphasize shared identity—as citizens, or as people possessing universal human rights—blurring the factional boundaries that separate competing groups.

Of course, many people worry about the erosion of these checks in their societies—and they should. When a society fails to contain violence, it is usually because it has neglected these fundamentals and allowed one or more of the five logics of war to flourish. If we take our eyes off the roots of conflict and focus on false or misleading causes, we are unlikely to make the choices that will promote peace.

Editors’ Note: Adapted from Why We Fight: The Roots of War and the Paths to Peace by Chris Blattman (Viking, 2022).