Stealth games can learn from human behaviour in public toilets

You sit there in a dark corner of an otherwise lit courtyard.  There are three armed men, vigilant in their task to find someone, anyone, that dares try and steal their secrets from the server room.  They’re searching for you, the super spy tasked with bringing an end to the cold war-esque rise of a wayward nation state, the only man who can put an end to the crisis.  You sit there, heart pounding, waiting for your opportunity to make a clean break for it – or die trying. 

There’s just one problem- You know you’ll never be found.

GarretI really like Stealth games and find myself in an enviable position as Thief currently occupies my Xbox One and Metal Gear Solid V: Ground Zeroes is about to sit in my Playstation 4.  Rest assured I will drain every ounce of gameplay from them as I sneak, shoot and strangle my way through carefully choreographed courtyards and corridors, dashing in and out of the shadows, and making opportunistic dashes across lit areas out of sight of the enemy.  There is nothing quite like ‘outsmarting’ enemy AI and getting to your objective unseen and unheard. I like procedural things, methodologies, processes – I like to think that the world is governed by a simple set of rules and that if you follow those rules you’ll more likely succeed than not.  Those rules aren’t necessarily ‘fair’, but they make navigating life a hell of a lot easier than it would otherwise be.

But within this order, I also like an element of randomness within the bounds of the ruleset.  I don’t want things to be entirely predictable, because predictability may as well be synonymous with monotony, but I want enough in the way of stochasticity to make things interesting. A while ago I wrote about how way back in 2002, Tom Clancy’s Splinter Cell blurred the line between playing a game and understanding it.  It required an understanding of how the game world would react to your input, and how to exploit the game’s rules to win the game.  And Stealth games by their nature require this type of implicit communication between player and game designer in order to succeed.   And that was amazing at the time as the game appeared to present almost unlimited freedom to approach the scenarios, the logic problems, presented to the player.   The artificial intelligence, the playfield and the rules of engagement were all fixed – the only variable is the player’s interaction therein.  Basically the game gives you a set of guidelines and tells you to “solve for X”.  SC_XBOX Of course with enemy artificial intelligence and almost binary measures of light and dark the game really became an exercise in memory and timing, as you remain hidden until such a time as you can exploit the enemy’s predictable movement around the level.  It is never terribly cerebral, and when it comes down to it it amounts to nothing more than keen observation.  There is nothing random about it as it becomes predictable – enemies follow the same routes, at the same time, looking the same way, seeing the same things. It is, in a word, monotonous.  Unfortunately with that certainty comes the reality that if you follow the clearly defined rules, you know you’ll never lose.

Of course all of this changes once you’re spotted by an enemy and its as if they’ve taken a proton pill that has increased their intelligence ten-fold.  It’s at this time, until (to use the Metal Gear terminology) Alert Phase ends, that they begin to act more like real life human beings in pursuit of a known target.  They pursue, take cover, flank and shoot in a bullish yet logical fashion.  Predictability is replaced by stochasticity and the player is forced to react to what they see and think, rather than what they know.  It’s paradoxical that the most exhilarating and primal aspects of stealth games are the ones that are the result of ‘playing them wrong’.

But how do you change such a well-worn but well-established genre?  Dishonored (which Lucius rather liked, enough to make it onto our best games of last generation list) gave the player greater freedom – better tools of movement and battle – to add an element of faux randomness to the player’s, and by virtue of that fact, enemy behaviour.  The result was a game that felt like a huge leap forward for the genre but really what  it did was hide the limitations of enemy intelligence and behaviour with a more stochastic and agile player. It was clever, and the closest thing we’ve had to a revolution of the stealth genre, but it still didn’t address the key underlying issues. Dishonored screenshot 2 Of course the other option is to change the way the artificial intelligence in-game behaves, forcing a sort of structural change in the way players approach each and every situation.  The thing is these games hide behind a very abstract depiction of human behaviour.  We are creatures of habit but not ones of predictability.  Humans are rational to an extent but behind every decision lays a complex thought process that can take into account any number of factors before we act, all in a split second.  And so while, on the surface at least, it may look as though we are acting in a predictable and rational manner, we are in fact making decisions based on all available information before deciding on our default action.  In economics we rely heavily on assuming consumers act in a rational manner, but we bloody well know that is not the case and that our models, our predictions and forecasts, are compromised by the highly volatile, or at least complex and calculating, nature of the human mind.

Case in point: think about every time you walk into a public toilet.  I don’t know about you, but upon walking into a bathroom, numerous stalls lining the wall, about one thousand different factors run through my mind in making the tough decision as to which to choose.  The obvious first point to consider is state of the bowl itself, which I’m sure any human in the western world will attest to, can vary from repulsive to remarkably clean.  But there are other factors that go into the seemingly simple decision to choose one cubicle over another.  Privacy, for example is the second factor, and one that can take into account dozens of variables in determining the level of therein.  For example, the location of lighting as to not cast a shadow onto the ground, the location of mirrors that may reflect through tiny gaps between the cubicle doors, or the closeness of the chosen toilet to the main entry into the bathroom.  All of these things weigh in, in varying degrees, into our decision to make the decision to walk into a cubicle, drop your trousers, and do your business.

But what happens when it all goes wrong and a third party enters the bathro0m and acts in a way you didn’t anticipate, as random, irrational people want to do?  I don’t know about you but it angers me to no end when someone enters the bathroom and chooses the cubicle right next to me, even though there are a number of cubicles to choose from that would maximise both of our levels of privacy.  The point is we make these decisions based on how we believe others will act – based on the knowledge of how we would act in that same situation.  But it’s imperfect information, so while we don’t know that Person X will come in and choose Cubicle Y, we can make guesses based on how we understand humans to behave when presented with a given set of factors.  It isn’t about observed human behaviour but rather perception and prediction of it.  Sometimes we’re wrong and we sit there irritated to the loud expulsion of waste matter in the stall next to yours.  But when you get it right, and no matter how many people enter the bathroom during your (hopefully) short stay it is perverse how much satisfaction you may derive.


What if games took this same logic and applied it to a game scenario?  Stealth games at present are about watching and reacting, often in a quick and decisive manner.  But what if they were more about predicting enemy movement in relation to an environment?  What if you had the time to assess the environment carefully and then make a decision before setting things into motion?  Imagine Lemmings where they have free will, or the Incredible Machine where instead of being bound by the laws of physics you’re bound by the laws of human nature and behaviour.  It would be a paradigm shift that could potentially give the stealth game a new genre.  It would take ‘understanding a game’s rules’ to a whole new level, in much the same way stealth games have done in the past.  Sure it adds some randomness and potentially some frustration – but on the flip side it may also lead to the creation of some of the most rewarding, intelligent and memorable experiences for players.  Then again it may be the impossible dream that our technology just doesn’t allow for.

What do you think?