War Is A Video Game, And We’re Losing
We are just a short while away from the launch of our new website, and we are holding a sweepstakes to celebrate. You can win your own home bar kit or collections of books on strategy, Asia, and the Middle East. All you have to do is sign up!
It is often said that the rise of military robotics and cyber warfare is turning war into a “videogame.” But this thesis—which blames technology for a supposed loss of moral seriousness about war—gets the causation wrong. It isn’t bloodless technology that really makes war videogame-like. Rather, videogames are simple and deterministic in that they mirror the ways a cross-section of national security experts think about war. It seems that as hard as we try to be treat war as “tragic, inefficient, and uncertain,” we end up getting our military analysis from the same mental place that’s engaged by a shopping trip to GameStop. We might as well use this to our advantage by diversifying our unconscious war(games) rather than playing the same titles over and over again.
One of the most common tropes in both military analysis and popular culture is the danger of war becoming a “videogame.” From Matthew Broderick’s “game” with a military supercomputer in WarGames to Robert Gates’ recent criticism of drone warfare, there is a strong tendency to equate technology with both dehumanization as well as an overly stilted and abstract view of conflict. While this sort of rhetoric is primarily deployed to critique drones and other standoff technologies, it also is used to bash mathematical or computational methods of analysis. Quoth Gates:
For too many people—including defense “experts,” members of Congress, executive branch officials and ordinary citizens—war has become a kind of videogame or action movie: bloodless, painless and odorless. But my years at the Pentagon left me even more skeptical of systems analysis, computer models, game theories or doctrines that suggest that war is anything other than tragic, inefficient and uncertain.
In evaluating the videogame-ness of particular understandings of war or methods of war, Gates and others make one true—but highly banal—observation about how war becomes videogame-like. In games, the player controls virtual agents in a synthetic environment. If the agent dies, it doesn’t lead to the real world death of the player. But Gates himself, though heavily troubled by the deaths and injuries of soldiers, ultimately was not directly affected by warfare. If an IED went off in Iraq, Gates didn’t get caught up in the blast radius.
While gamers do not obviously bear the kind of emotional burden Gates did, fans have reacted with substantial empathy to the fates of game agents as well. When the villain Sephiroth murdered Aeris in Final Fantasy 7, an entire generation of gamers’ innocence was stolen. Aeris, despite being a bundle of pixels, was someone who many players very much wanted to see have a happy ending. Similarly, games like Spec Ops: The Line and The Last of Us force gamers to make difficult moral choices about characters they care about. Lastly, the Silent Hill and Resident Evil series at their best—like any good horror film—are calculated attempts to manipulate gamers’ emotions and physiological responses to terror, fear, and uncertainty.
Gates is, however, unintentionally correct that certain doctrines and theories lead to a videogame-like understanding of war, but not even remotely in the way that he has argued about drones, game theory, or computer models. Videogames, in the mainstream, are very deterministic. Even open world games, such as World of Warcraft, rely on regularities and rules that constrain the universe of player choices. I will focus, however, on two areas of determinism seen in military-themed videogames such as first person shooters and strategy games. Both of these qualities are reflected in the way that many analysts unconsciously think about war and warfare.
First, many game agents are coded as finite state machines (FSM). In computer science, a FSM is a kind of abstract machine that has a limited number of possible state transitions. Vending machines, for example, will output snacks and drinks once the correct combination of money is entered. All FSMs have particular conditions that trigger transitions to differing states. But just as you wouldn’t say a vending machine is “intelligent,” neither is a videogame agent that is coded as a FSM. As one prominent videogame programmer wrote, game agents are all about creating the “illusion” of intelligence. Why? Besides ease of programming, players also want “dumb” enemies. If Battlefield 4 enemies could fire and maneuver as well as Deep Blue plays chess, then who would want to play? For many casual players stymied by online (and often preteen) opponents’ boundless amounts of free time and blindingly fast reflexes, single player is at least a safe option for fighting and beating a computerized opponent.
The sensation of Deep Blue vs. Kasparov aside, playing a skilled human opponent is still preferable to an untiring, unyielding machine that massively exceeds human cognition in a few narrow (but important) areas of play. Yes, “he” might be a bit stiff compared to online human opponents. But knowing that he’s predictable won’t matter if he still routinely owns you. And what’s the real reward for beating him? If the reward is just the challenge, most people prefer to play against creative and interesting flesh and blood enemies online, even if they’re foul-mouthed 12-year-olds.
So what’s the connection? Human behavior is complex and full of variation. We have agency, can make decisions, and especially make decisions based on what we think others will do. But you wouldn’t know that from both videogames and certain modes of security analysis.
It isn’t exactly apparent that the military wants to test itself against “smart” enemies either, as the Millennium Challenge 2002 wargame demonstrated. As many recountings of the exercise argue, the wargame organizers were distinctly unhappy when Marine Corps commander and strategic theorist Paul van Riper used creative tactics to show how a clever enemy could defeat the most powerful military in the world. Army soldier Ben Zwiebelsen has also argued that military training, like videogame agents, aims for the illusion of intelligence. Instead of simulating opponents that have a truly dynamic range of behavior befitting their cultural, political, and strategic differences from the U.S., soldiers train against copies of themselves that have superficial resemblances to the enemy.
Finally, until the 1980s and 1990s, the military used force planning models that assumed opponents fought like bad videogame enemies—enemies that would rigidly fight until they were all or mostly killed. This stands at variance to both qualitative military history and combat data compiled by famed military mathematicians like Robert McQuie. But wargaming and combat simulations aren’t the only places where analysts treat “real life” as if it were a videogame.
When Middle East analysts assume that all of the region’s problems can be blamed on Sykes-Picot and other Western great power agreements, they assume that Sykes-Picot—like a programmer’s design for a single player level of Battlefield 4—constrains the “agents” that populate the Middle East to a limited and deterministic set of behaviors. Play the Middle East game under Sykes-Picot mode, and you only have to worry about a given range of game agents and a simple range of their behaviors in addition to the features of the game environment. But if the Middle East was that deterministic, we would have already beat that game handily. Unfortunately, there is no BradyGames strategy guide for our adventures in the Middle East, which like the Final Fantasy series seem to grow worse with time. And Robert Pape’s simplistic theory that foreign occupations trigger suicide bombings is so deterministic that it could be likely implemented as an extensional feature of a Monopoly-like board game. Occupy certain real estate on the board and you trigger the suicide bombing in-game event. You are prompted to draw a card with a picture of a car bomb on it. You roll the dice, and depending on whether or not the attack is successful, you may lose soldiers, diplomats, and NGO personnel. It may sound ridiculous, but many of Pape’s fans believe it nonetheless.
Second, most strategy games, role playing games, and first person shooters incorporate a rigid and deterministic technology tree. Players must advance to the bottom of the tree in order to gain crucial technologies and capabilities. In many strategy games, it is difficult for players who are at the top of the tech tree to compete effectively with players who have reached the bottom. The latter players can exploit superior technology, tactics, and economies of scale in weapon and soldier production. The idea of a rag-tag band of technologically unsophisticated guerrillas being a serious problem for a player in, say, Starcraft or Warcraft IIt is ridiculous. Asymmetric counters are possible, but players have to either keep pace technologically or kill their opponents when they are both at the beginning of the tech tree, as Starcraft‘s Zergling rush tactics demonstrates.
There is also none of the ambiguity in the military tech tree seen in real life. Military-industrial production is so simplistic that mathematical algorithms can find the most optimal way to invest in-game resources, something that would be computationally intractable for real life military-technical problems. No matter how mathematical the operations research models, global strategic planning in real life cannot be completely automated with the algorithms used to find optimal ways of moving through the Starcraft tech tree and optimizing the best military formations for each playable faction.
Yet, the technological determinism seen in strategy game tech trees mirrors the causal determinism seen in national security discourse. As War on the Rocks contributor Chris Mewett argued, technological determinism characterizes a lot of thinking about how war, society, and politics evolve. Many international relations thinkers have argued that discrete technological regimes that favor “offense” and “defense” exist, with WWI-style technologies leading to “defensive” warfare and missile-era technologies giving the advantage to “offensive” technologies. Warnings of a “cyber Pearl Harbor” assume that cyber weapons—like wizard-esque units that can summon powerful and devastating spells in Warcraft and Starcraft—are going to be inevitably used and a country without offensive cyber capabilities is like a Terran player in Starcraft without his own supply of deadly Ghosts.
Theories like the so-called “modern system” of military tactics argue that a kind of videogame-esque technical and doctrinal tree exists in which militaries that can fire, maneuver, and exploit cover in autonomous groupings of maneuver units dominate those who cannot. Armies that can’t—for technological or cultural reasons—get to the bottom of the tech/doctrine tree will fail against those that have. Many arguments about the diffusion of military technology and innovation visualize geopolitics as a giant strategy game in which actors race to reach the terminal nodes of the tech tree. Just like a game played long enough will lead to the diffusion of a unit type or tech to the entire group of players, many theorists believe that guided munitions, drones, and cyber weapons will eventually diffuse to the entire set of great powers and sub-state actors. Again, the idea that a tech/doctrine tree might not be linear, or that, as Colin Gray said, “weapons don’t make war” is utterly foreign to this perspective. In this mode of theorizing, weapons and doctrine do make war to the point where cyberwar hypotheticals lead to changes in the “nature” of war itself.
So yes, Secretary Gates, plenty of people who couldn’t solve a game theory problem to save their own lives and/or likely have a variety of opinions about the efficacy and morals of drone warfare think about war, warfare, and strategy in ways that would not be out of place in a top-selling videogame. This isn’t because they’ve been bewitched by technology or abstract models. It’s because all of us create models of the world in our own minds, whether we are trying to analyze theory, build virtual worlds, or find simple explanations of complex things in day-to-day life.
Videogames (and games more broadly), however, offer opportunities for creative exploration that other models don’t. We can immerse ourselves in alternative identities and possibilities and play through them, even if game rules are simple and deterministic. The fact that we can re-spawn when our characters die allows us to experiment and try different strategies, much in the way that Bill Murray always has a second chance in Groundhog Day. We can play against other people in dynamic multiplayer games that allow us to explore the parameters of a given map or play mode via free play. And strategy games, like Starcraft, that force you to play each game campaign set of missions as a different military faction, give the player a general command of abstract ideas about strategy and tactics that are independent of a particular context.
In this context, it is depressing that the same leaders who dragged us into a decade-plus of indecisive war and are poised to muddle through yet another “strategic” planning cycle could not and cannot treat war as a videogame. A skilled Team Deathmatch or role playing game player—unlike many defense analysts and military men—intuitively understands that combined arms is the only way to win. An imbalanced World of Warcraft party that has too many close combat specialists and not enough ranged weapons will lose to one that better balances capabilities. Playing as Zerg, Terran, and Protoss in Starcraft allows the gamer to truly grasp the choices available to anyone who plays as any one of the three factions. Sadly, in Afghanistan we did not listen to those who produced detailed analysis of our own allies, and further ignored those who produced detailed assessment of the enemy.
If we had challenged our internal cognitive models by gaming, perhaps we might have been able to think more deeply and creatively. Gaming matters precisely because war is “tragic, inefficient and uncertain.” The stakes are too high for us not to try to creatively probe our assumptions and experiment with new possibilities. If we already treat war as a videogame internally, at least we can add a diversity of game titles to our mental PlayStation 4s instead of playing the same game over and over again.
Adam Elkus is a PhD student in Computational Social Science at George Mason University and a columnist at War on the Rocks. He has published articles on defense, international security, and technology at CTOVision, The Atlantic, the West Point Combating Terrorism Center’s Sentinel, and Foreign Policy.