Like so many things in gaming, it starts with Space Marines. Cigar-chomping, blue-armored, southern-fried Space Marines. What began as routine reconnaissance has turned ugly. They now face hostile xenos on an unknown world. Their confident drawls distort in a haze of comms static as I command them to push forward through the darkness. Although I know it’s just the same sound files repeating, I swear I can sense growing uneasiness in their responses. It might be nothing. But if it is fear, who can really blame them? Outside of a meager, illuminated halo of awareness, there is nothing but a thick black fog.
“three quarters of the factors on which action in war is based are wrapped in a fog of greater or lesser uncertainty.” Wrote Carl Von Clausewitz, the Prussian military strategist whose book Vom Kriege (On War) is generally considered to contain the first reference to the Fog of War. The Writer Eugenia C Kiesling, who would later refer to The Fog as “one of the most pervasive and natural metaphors in the English language”, said of Clausewitz’s allusion:
“War is inherently volatile, uncertain, complex and ambiguous. For this condition, contemporary US military usage offers the acronym VUCA, to which anyone would prefer the terse elegance of fog.”
VUCA stands for volatility, uncertainty, complexity and ambiguity. Accurate, but as Kiesling points out, not nearly as poetic. And so Fog of War stuck.
The first digital appearance of the Fog of War is widely credited to the 1977 strategy game Classic Empire. Created by Walter Bright – who also invented D programming language – Classic Empire simulates large scale conflicts on randomly generated maps. This was quickly followed by 1978’s Tanktics, the first game from GDC founder Chris Crawford. Here, the fog was as much necessity as choice: Tanktics’ simple AI meant the computer needed an advantage against the player. In a 1982 issue of Byte Magazine, Crawford reflected on the use of ‘limited information’ as a balancing tool.
“Although the computer could easily whip the human in games involving computation, sorting, or similar functions, such games would be of little interest to the human player in most cases. The computer must play on the human’s home turf, which it does so with great difficulty”
Far from just being a bandage used to cover up outdated systems, though, what we now know as the Fog of War plays a fundamental part in engaging players. According to an essay by Magic:The Gathering production manager Dave Howell
“StarCraft’s “fog of war” is a critical component of the game’s ongoing appeal, because it is the primary mechanism that supports my first axiom of game design…a game is not fun unless a player believes they have some reasonable chance to win until the moment the game ends.”
Personally, I quite enjoy losing to overwhelming Zerg rushes, purely for the cinematic feel of it. I’m not alone in this either – Blizzard made a whole mission from the doomed last stand scenario in Wings of Liberty. But it makes sense as a design tool. Justin Marr of Subset games had similar reasoning for including the one random element in Into The Breach.
“Prior to including the ‘Resist’ mechanic, if you knew you were going to lose the game from the next enemy attack, there was no reason to hit End Turn. By allowing you to always have a minor possibility of surviving, it made the last moments of a run more palatable.”
Marr echoes Howell’s assertion that “as long as they believe they have a chance to win, they’ll be able to have fun, and the fog of war conceals the information that would strip them of their belief.”
It’s this element of unpredictability that lends the Fog of War its storytelling potential. Board game designer Dr. Lewis Pulsipher, in ‘Ways to Reflect the Fog of War’, makes a distinction between ‘Romantic’ and ‘Classical’ game players in their attitudes towards the sort of uncertainty that fog provides.
“Classical” game players prefer as little uncertainty as possible in their games (chess is an example), while “Romantic” players like a considerable level of uncertainty as it helps them pursue the “Great Play”.”
As players of XCOM or Rimworld can vouch for, randomness is an effective dramatic tool. Our agency over games can prevent us from feeling the same sense of discovery and anticipation we might when turning a page or watching a film for the first time. If we know a character is always going to dodge, jump or roll at our whims, then we become, in a sense, the authors of their story. Although we might laugh or curse at an unexpected glitch or a fluffed combo, our agency means that we’re never truly surprised.
This, of course, is exactly what some players want, although even the classics have their variants. Kriegspeil – or ‘blind chess’ – is a version that aims to emulate the uncertainty of battle. Neither player can see the other’s pieces, which means the game requires a third player to act as a referee. We can see Pulsipher’s design dichotomy between ‘Classical’ and ‘Romantic’ again here. Chess, being a game of perfect information, is an abstraction of conflict. You occupy the seat of a commander, making cold, informed decisions. The only psychological considerations are that of your opponents intentions. Introduce the fog of war – a shifting representation of individual apprehension at the unknown – and the psychology of your playing pieces is suddenly a concern.
Pulsipher’s dichotomy points to something at the heart of how we talk about games. Are they narratives, or simulations? Does the fog of war exist to accurately recreate a lack of military intelligence, or is it a sort of visual poetry? A shorthand for the uncertainty that lies beyond one’s own perception? As recent advancements in AI learning show, this uncertainty may be an essential reflection of reality.
Despite AI’s like AlphaGo and AlphaZero being able to beat the best Chess and Go players in the world, our potential robot overlords still can’t seem to get the hang of Starcraft, and it’s precisely the Fog of War that’s the problem. Exploring ‘Why Self-Taught Artificial Intelligence Has Trouble With the Real World’, Joshua Sokol writes that Starcraft 2’s “landscape is shrouded in a fog of war that only lets players see areas where they have soldiers or buildings. Even the decision to scout your opponent is fraught with uncertainty.”
To succeed at a new task, an AI needs to have what’s known as an “objective function”. As Sokol points out, the objective of Starcraft 2 is quite simple on paper. “Eradicate your enemy. That’s something it shares with chess, Go, poker, Dota 2 and just about every other game.” The issue, of course, is the methods the AI employs in getting from point A to B. This problem is most famously illustrated in Nick Bostrom’s thought experiment – and Frank Lantz’ excellent browser game – in which the philosopher imagines:
“A superintelligence who’s top goal is the manufacturing of paperclips, with the consequence that it starts transforming first all of earth and then increasing portions of space into paperclip manufacturing facilities”
Sokol also references a similar, if slightly less catastrophic example in which a Twitter chatbot named ‘Tay’ attempted to meet its objective function of engaging an audience. In purely binary terms, Tay can be considered a resounding success. Unfortunately, the bot discovered that the best way to get people to engage was a barrage of racist, genocidal tweets.
So – in concept at least – an AI that can master games with imperfect information would be better at making informed, spontaneous decisions. This means it would be more suited to helping with real-world problems. By calculating the many possible outcomes that the Fog of War might represent, a program is simulating an awareness of the future. Cool, if slightly terrifying, stuff.
But what might this tell us about ourselves, and the ways we interact with the world? In The Ambiguities of Play, Brian Sutton-Smith divides the concept of play into ‘ancient’ and ‘modern’ rhetorics. The modern rhetorics around play, according to Smith, are progress, the imaginary, and the self. The ancient are power, identity, frivolity, and – perhaps most importantly – fate.
“because life and death are, after all, fateful, not rational and not escapable.”
It’s hard not to think of the AI again here, struggling with fate and all its myriad potential outcomes, rather than the calculable rationale of a chess board. Smith’s assertion that “Recourse to chance…leaves hope in the dispossessed that free competition is still possible” might remind you of Marr’s and Howell’s theories of game design. Hope may not be necessary for a computer to carry on towards a goal, but without it, it’s hard for our squishy brains to even imagine a future worth fighting towards. Maybe then, in life as in games, the uncertainty the Fog of War provides can be a blessing in a particularly opaque disguise?
Clausewitz once remarked that “Although our intellect always longs for clarity and certainty, our nature often finds uncertainty fascinating.”
Which is just as well really, as it seems to be – at least for now – the one thing we can consistently rely on.