Category Archives: Games Theory

On toxicity

The word ‘toxic’ appears to have gone viral, at least in the competitive games community. It’s unsurprising, really, as it provides a lovely, pithy way of describing a certain kind of player or even person, to their backs at any rate.

As far as I can tell, Riot coined the term for PR purposes, as a way of defining the typical player who is not… socially considerate when talking to their community. In that role, it’s a fantastic tool, highly expressive and intuitive as to what it means. Of course, so are the words ‘fun’ and ‘game’ and don’t we all know the trouble that’s got us into, eh?

As much as I love Riot, this is one thing I can find serious issue with in their MO. It appears the company has collectively drunk the koolaid of believing that this PR jargon they made up is an actual thing, not just a neat way of describing a stupidly complex phenomenon. What’s made worse is that every definition of toxicity I’ve seen from them is different but for one fact: they all have a negative impact on consumer retention. We can fluff that up and call it damaging game experience and so on, but when it boils down to it, this is not about game design any more. It’s about audience selection and making wads of cash money. This isn’t inherently evil, of course, but it’s certainly not unambiguously good either.


So let’s get this straight. There is no such thing as a ‘toxic player’, at least not in the sense of actual flesh and blood players. It’s alright to use the term in pure hypotheticals and label a certain kind of action group toxic and consequently a hypothetical player who embodies that group a toxic player, but that’s not how the real world works, folks.

The worst outcome of this misuse is that people, actual people, are diagnosing themselves as toxic in fits of hypochondriac glee. Unsurprising given the peculiarly western tendency to syndromise any kind of problem, but distressing none the less. Personally, I am terrified of this development. There is no kind of psychological treatment in which repressing and punishing antisocial urges has anything but a styming effect on personal development and enjoyment for the individual being targeted.

Now I hear toxic being bandied around on unrelated podcasts, in design articles, interviews with devs, even potential research briefs. It’s a catchy, catchall meme. It’s also a delicious label to stick to someone you don’t like, a perfect tool to brand ‘the other’. It’s one of those inconspicuous little words that can worm its way into the consciousness of a populace and create divisions, excuses for unethical behavior and justifications for the unjustifiable. The sort of word that is one of many engraved in each paving stone on the road to hell.

Toxic, put simply, is toxic.


I stress that I’m certainly not saying that there’s no such thing as a player whose play is detrimental to the fun of other players, nor that a developer shouldn’t talk about this or even design things to prevent or mitigate that. Riot, for the most part, is doing the right thing- implementing systems in their game which gently encourage players to play in a way that is tolerable to other players. The problem is mostly in the discussion and the misappropriation of the term.

The most dangerous result of this is the idea that a certain kind of play is inherently ‘wrong’. This is… it’s just plain disturbing. Play, at its core, is a process of experimentation and discovery. Play exists to push boundaries and explore what is normally not acceptable in ‘real life’. It also serves to train and develop, to instill contentment or joy and create wonder. Fundamentally, though, human beings, in fact all animals which play, play in order to do what would otherwise not be possible. To enforce a moral rhetoric upon this is almost always damaging to the player, because that concept is antithetical to the core purpose of play.

I’m not talking about something that’s nice and fluffy here. As much as the activity of play can be about exploring love and trust it is also about exploring betrayal and anger. It’s a careful balancing act, but one that is fundamentally necessary for human mental health and self development. In the end all those people calling you a noob faggot have a largely positive outcome on you as a human being, hard to believe as that may be at the time. What it doesn’t help is the developer’s profits, because you’ll probably drop the first few games where you’re exposed to that sort of antagonistic play. When these two collide we arrive at the problematic concept of toxicity. The utopian idea that all players can live in harmony and sing around the campfire at night. Like all utopian concepts it is enticing, dangerous and utterly wrong.


If I hold this to be true, I also have to hold that it is reasonable to create exclusive communities of play, where entry requires certain rules to be accepted and adhered to by all players, that is after all the essence of a game. The important distinction between a ‘toxic’ player and one who infiltrates such a community and then breaks the rules is that the latter is simply a rulebreaker, and can be expelled as such. A toxic player, on the other hand, implies someone whose play, while acceptable within the rules, is filthy and unclean. There is a double standard there- you’re welcome to come and play however you like, but good players only play like this. If you don’t, you’re a bad player and everyone is allowed to judge you.

That’s a far cry from a simple, professional: These are the rules, to maintain the game as it is intended to be played, the rules must be adhered to. If you don’t, you’ll be removed from the game. No hard feelings, that’s just how it has to work. If it’s in the rules, the players are free to exploit it however they so choose. The responsible party for toxic outcomes is not the player, it is the designer.

So I simply ask that we should be careful with the word. It’s not entirely without merit, but it’s powerful and dangerous. It shouldn’t be thrown around to label everyone and their kid brother. Nor should it be used in formal discussion unless clearly and meticulously defined and justified.


As a kind of postscript, I will say that Riot seems these days to have shifted their attention from toxic players to toxic environments. Unfortunately, the earlier discussions have created the implicit assumption that the former creates the latter, where Lyte’s recent analysis at the GDC shows that is clearly not the case. I applaud them for bringing that salient little fact to our collective notice. Notice it.


ERI #1: The experimentation behaviours


For basic information on the ERI framework, please read this article



Experimentation behaviours are the first of the kinds of behaviour used in ERI. They’re also perhaps the most common and useful for analysing modern game design. E behaviours are motivated by a particular form of pleasure reward, derived from pattern recognition and creation.

They manifest in various forms based on the properties of environments. Environments that have a lot of unknown features trigger open experimentation behaviours. These are characterised by lack of a focused goal and match animal and child play forms in that they appear aimless, spontaneous and compulsive. As environments become apparently well known and understood, behaviour shifts towards rigorous experimentation behaviours, which tend to be highly focused, formal and generate abitrary rules and limitations.



Experimentation behaviours really form the ‘core’ of the ERI framework. This is mostly because they’re the behaviours which modern game design revolves around triggering, so they’re the ones you’ll deal with most often when analysing videogames. They’re also the easiest to break down and isolate, so they’re the easiest to base a theoretical design around.

Raph Koster’s theory of fun does a really good job of outlining the basic premise of these behaviours  so if you’ve read that you’ll be in a good place to see why they’re important. However, as I came up with the framework before I’d read Raph’s work, the different angle I came in from helped me notice a few interesting things that Raph doesn’t talk about and I think are pretty important.

For those who haven’t had the pleasure, Raph sums things up by saying that ‘Fun is just another word for learning’. Similarly, experimentation behaviours are roughly synonymous with learning. Experimenting is, at a really basic level, actively inducing change in the environment in order to see what happens. In seeing what happens you are able to make more educated judgements about that environment. In many cases ‘playing with something’ and ‘experimenting with something’ is used interchangeably, though it’s probably unwise to do so when considering play on a broader scale as we are here.


The element of ‘fun’ comes in thanks to a particular biological effect that I would have liked to study in far more detail than I have been able to. There is a distinctive form of pleasure associated with recognising and understanding patterns of all kinds, be they entirely new or the recognition of similarity in two or more patterns previously thought distinct. I’ve yet to meet someone who didn’t immediately understand what I was talking about when I described it: the ‘aha!’ moment, the excitement of discovery. I call it the ‘Eureka sensation’. The important thing to understand is that while the most recognisable experiences of this sensation are quite sharp and distinctive, the same principle applies to all pattern matching, be it highly conscious and deliberate or almost unconscious and incidental. The sensation of slow, steady learning and progress is subtle and softly invigorating, like slow burning carbs to the sugar rush of the sudden, powerful discovery. They are still both a result of the same motivating factor- pleasure from learning.

It’s common to assume that the pleasure of discovery largely comes from the material benefits this may bring, but games scholars for the better part of a century have found that this is not the case. Win or lose, so long as the player’s understanding increases, they’ll feel this particular kind of pleasure. This isn’t strictly true, actually. The trigger is the player recognising their understanding is increasing. This leads to the possibility of someone getting the effect from thinking they see a pattern where none actually exists. This is a fairly common and well documented phenomenon in psychology which illustrates one of the important subtleties in the behaviour. So long as the individual thinks they’ve recognised a pattern, they receive the biological kicker. Actually being right is incidental. Remember that finding a negative- being wrong- is also discovering a pattern, and so can yield just as much, if not more, of this sensation,. This is why surprise is often so ‘fun’. There’s another rarer possibility in that someone is recognising patterns but it is subtle enough that their preconceptions override the reality of the process, leading to frustration. This is one of the more common interactions between E behaviours and R behaviours, so I’ll cover it in more detail in a later article.

It’s easier (and perhaps more accurate) to think in terms of the reward being given for independently constructing patterns- be they proved true or false- rather than just recognising existing ones. This leads to a lot of interesting phenomena to do with imagination that I’ll again not go into here for the sake of brevity.

So the motivator for experimentation is this eureka sensation, a pleasure reward mechanism for indulging in the behaviour. Similarly to sexual pleasure it is powerful and not particularly precise, which means that, just like sex, we’re inclined to do it whenever possible, even if it’s not actually fulfilling the ‘intended purpose’ biologically. This contests purely utilitarian conceptions of play and, while not particularly important to game design, it does make for interesting lines of thought that have helped me more accurately figure out where these behaviours might be influential.


So what are these behaviours I’ve been going on about? They are a spectrum of tendencies that arise from this basic mechanism, depending on the environment in which the individual finds themselves. In this case, environment means any kind of recognisably distinct entity with its own set of rules*. So while ‘everything’ is an environment, so is a melee match of Starcraft, a relationship with a friend, a piece of fiction or a Rubix Cube. It’s a fluid thing that shifts with attention.

When an individual is placed in, or restricts their awareness to, an environment in which they have no information from which to create a framework to base their behaviour upon, they will engage in open experimentation behaviour (OEB). This is the more or less random stimulation of the environment by any means possible in order to rustle up some patterns to be observed, analysed and compared to existing knowledge of cause and effect outside the environment. The classic example of this in gameplay is someone reaching a point in a puzzle where they become frustrated. This is the point where they recognise they do not understand the properties of the puzzle and that patterns they assumed were true in this environment are not. They begin to register the environment (which is currently limited to the specific puzzle they are stuck on) as having predominantly ‘unknown’ properties and this triggers OEB.

OEB matches many descriptions of uncultured playforms, from Fagen’s ‘apparently purposeless activity’ to Caillois ‘spontaneous manifestations of the play instinct’. It is characterised by performing a variety of sometimes wildly variable, sometimes strangely methodical actions without a clear ‘goal’. From an internal perspective, the behaviour is almost subconscious, occurring when puzzled or unsure. You act ‘just for the sake of doing something’, move things around, manipulate them, re-arrange them, try and get more sensory information from them- taste, touch, sound, even reaction if the subject of uncertainty seems like it might be alive. By young adulthood this process is so intuitive in most people they don’t recognise they are doing it. It is such a fundamental behaviour that much of your social training and conditioning as a human being relates to controlling and limiting your own tendencies to engage in it, a point that becomes important when we deal with the taboo-breaking capabilities of games and play.

At the other end of the spectrum is rigorous experimentation behaviour (REB). REB occurs in environments where an individual is able to successfully predict the results of simple actions based on their established understanding of patterns- if I drop an apple, I can predict it will fall. If I begin running I can predict roughly when I will get tired.

Rigorous experimentation behaviour isolates and clarifies noted ambiguities in patterns- if I run to the shops, I get tired faster than running to the park. The shops are uphill and the park is downhill, so I’m going to run somewhere else that’s also uphill the same distance away. If I get tired at the same rate, I have more evidence supporting the pattern that uphill running tires me quicker than downhill running. You may note this is very similar to the basic form of the scientific method- hypothesise, isolate, experiment, speculate based on results. This is a fairly good model of REB. The main difference between REB and formal science is that the scientific institution pays a lot closer attention to not falling into logical fallacies than an individual’s natural REB.

Beyond basic isolation and investigation, REB’s have an incredibly important place in studying games, because engaging in REB is essentially creating a game. It is creating a set of abstract limitations and rules with an end goal- testing the initial hypothesis. Everything from doing pushups to see how many you can do to testing whether your new girlfriend minds as much when you leave the dishes in the sink compared to your old one is REB and, at a basic level, a game.

REBs tend to become more and more focused and sharply defined as time goes on. In the frequent case where the REB involves the cooperation of more than one individual, this necessitates clear communication and, inevitably, formal codification of the limitations of the behaviour and goals. You can see where this is going. Caillois identified this trend as ludus, opposite to the spontaneous and apparently direction less actions of paidia– my OEBs. Ludus is the tendency towards ever more intricate and formal limitation in play. The incessant desire to create new rules, to isolate and to emphasise.

This tendency once again derives from our addiction to discovering patterns. By artificially limiting and restricting things we can create new ‘micro environments’ to analyse. I can do 20 pushups, but how many can I do on one arm? On my knuckles? On the backs of my hands? Clapping in between? Combining these restrictions? When our natural environments do not yield sufficient material to satisfy our lust for patterns, we simply make new spaces with new rules. When they are well understood, we further delimit them, again and again and again.

These sorts of behaviours also perform an interesting and powerful function in biological terms. Isolating and developing specific areas of capability in turn generally yields better results than less deliberate, unstructured experience. Once again, not particularly important for game design, but helpful in understanding how this odd proclivity may have remained in our makeup.


Now, I describe experimentation behaviours as a spectrum. Based on what we have so far, I can define the ends of the spectrum like so:

At one end, in an environment where the individual has no understanding of cause and effect and no points of reference to other environments from which to predict, they will engage in true open experimentation behaviour- acting upon the environment with no goal other than to make something, anything, happen. This purely abstract, for it’s functionally impossible for such an environment to exist.

At the other end, an individual who thinks they know everything about an environment will immediately begin to ‘play god’ and add rules to it, changing the environment to create the opportunity for more patterns to develop and be observed. Once again, this is an abstract space, because no-one (some might say no-one sane) thinks they know everything about an environment. This is since the borders of any environment are a little porous and connected to others by common features, which are in turn connected and so on, meaning that to say you understand one environment is to say you understand them all. People implicitly understand that you can’t perfectly isolate anything from reality because… well, it would cease to be a part of it then.

All experimentation behaviours occur somewhere between these to states- absolute incomprehension to total knowledge, and it is this which determines their properties.

The closer to an environment which would spawn OEB a person is:

  • The more they will act based on no specific goal or innate assumption
  • The less likely they are to predict the outcome of an action
  • The less they will be inclined to arbitrarily limit their own actions based on abstractions
  • The lower the gap between action and observation

The closer to an environment which would spawn REB a person is:

  • the more focused and formal their explorations will be
  • the more likely they will predict the result of their action
  • the more defined and intricate this prediction, if made, will be
  • the more likely they are to artificially constrain or alter their own actions or the environment itself


To put this in a more narrative mode. Let’s say we dump someone in an entirely unknown environment. Their thought process is likely to begin with thoughts such as:

“I wonder what will happen if I do this

Over time as they explore their environment, the shape of the question will shift towards something more like:

“hmm, I wonder if that will happen if I do this”

As a picture of cause and effect develops more strongly it will eventually shift to:

“hmm, I’m pretty sure this causes that to happen, so that means, when I do it like so, this other thing should also happen, because that’s the way it works over there.

Or even further along:

“Ok, so this does that, and this other thing also does that. If I take the ‘this’ out of this other thing, does just other thing still do that? I don’t think it will, because I’m pretty sure it’s just this that does that.

And finally the stage of artificial alteration is reached:

“well, I know this does that for sure, and it’s definitely only this that does that, not other or thing or even elephants. So let’s see what happens if I get my axe here and see if thi works just as well as this

Note in the last stage the phrasing shifts back towards that in the earlier thoughts- ‘Let’s see what happens’, not ‘I think … will happen’. By making an unnatural change to the environment, (breaking the unity of this) this hypothetical thinker creates new unknowns to be explored and observed.

Between these stages, I’ve also observed there is a gradual shift from the next stage being present near the edge of the subconscious to being the dominant conscious process, but having only my own brain, those of a few close friends and anecdotes in literature and so on to build these observations from I’m not particularly confident saying this is an actual thing.

The most important thing to keep in mind when dealing with this spectrum is that it will only ever trend one way. From a given starting point, over time behaviour will always move towards the REB end of the spectrum. In most cases this progress will be fairly smooth and one directional, though in some cases where a person makes some errant connections on entering an environment and starts with the assumption that they know more than they do, they’ll move forward, hit a wall, slide backwards for a time until they re-establish a solid grounding to work from then begin trending forwards once more. If a person isn’t making false assumptions, their understanding of the environment is genuine, they cannot slide backwards towards OEB, because you cannot ‘fake’ incomprehension. You can ‘fake’ something that looks like OEB, but picking apart thoughts reveals it to be just REB adding some limitations to the environment to give it some new, unknown properties and thus suitable for OEB once more.


As you can see, the E behaviours cover a lot of stuff, and do so at a very low level. Like the law of natural selection, when exposed to reality, E behaviours are simple and powerful enough to spawn a mandala of emergent, interweaving, intricate effects that have taken me the best part of a decade to even begin to get my head around over countless instances of careful analysis and observation. So I’m not going to try and give you much beyond the examples in the article so far already do. In later articles I’ll do some case studies under the framework which will give more specific insights into the topic.

Here, I’m just going to once again summarise the most basic effects of the E behaviours and their motivating trigger:

  • People want to discover patterns. They will actively and constantly act upon their environment to provide more information with which to observe or deduce patterns.
  • The reason people want to discover patterns is it gives powerful pleasure. This means that if this pleasure is blocked biologically they will not want to perform these behaviours. If acting upon the desire is restricted, it creates agitation, tension and unease.
  • The most basic effect of this trigger is that when you put someone in an environment with unknown quantities, so long as they can act upon that environment their understanding will increase over time thanks to this desire. If they can only observe their experimentation will be limited and consequently the trend towards understanding will be slowed and reduced in scope.
  • If an individual believes they understand an environment they will tend to artificially isolate elements or expend more and more time trying to break them down to temporarily create more uncertainty.
  • This tendency will lead to an ever increasing number of increasingly sharply defined artificial limitations imposed by the individual upon the environment. These can be wholly intrinsic, formally defined but intrinsically maintained or extrinsically imposed (see this article for more on intrinsic and extrinsic limitations/rules).


*I’ve been having some debates on this after my article on rule variance. I do not conceive of rules as hard boundaries. In reality, rules can exist in a strange space where they are both true and untrue. So while it is true that the rules of soccer say you have 11 players each team, we can play a game of soccer with 12 players a side. Game theory may state that, no, you are playing a different game which is a variant of soccer, but the way we conceptualise soccer under normal circumstances does not make this distinction. A similar construct exists in biology, the species distinction. In reality there is actually no such thing as a species, since there is no possible way to define the borders of a given species other than taking an individual animal and saying ‘any animal which can’t successfully interbreed with this animal is a different species’- not how we apply the concept mentally at all. Our brains create these sort of ‘both true and untrue’ concepts constantly.


Dealing with Ambiguity

In my very first lecture as an undergraduate the lecturer, curse his rotten soul, asked us a seemingly trivial question- “So, what is fun?’. As ludologists the world over shivered in horror, I began the process of trying to answer that question. Seven years later I’m still not there, but the very act of pondering has given me some insights that I’d like to share.

The first thing that should be acknowledged is that fun is an ambiguous concept. It is certainly not alone when it comes to the field of games and play: In fact, both the terms game and play are highly ambiguous themselves, enough so that when Brian Sutton-Smith wrote a career capping masterwork on play, he didn’t call it ‘the wonder of play‘ or ‘the importance of play‘, or even ‘the oddness of play‘. He called it ‘the ambiguity of play‘. Throughout my studies I don’t think I’ve come across someone more suited to attempting to define play and games and if he chooses to define them by their ambiguity, I feel that deserves intense consideration.

As a more practical type, interested in hands-on design challenges and establishing good practice in that regard, this ambiguity is a real sunnavabitch to deal with. Almost every other field of design is dealing with quite simple, unambiguous tasks. Houses need to provide shelter and durability, Chairs need to provide comfort and a pleasant aesthetic, tools need to be practical and efficient. Of course there are many more subtle nuances and complexities to these design tasks, but those same nuances are common across all fields of design, so a game designer has to cope with them as well. We also, however, have to cope with not knowing the basics of what we are trying to achieve. Comfort, durability, efficiency. These are not very difficult to predictively design for. Fun is.

This is made even more demonic by the fact that, despite the fact nobody knows what fun, games or play are, everybody knows what fun, games and play are. Everybody. You can take a three year old out to the park, point at people doing things and say ‘is that fun?’ and you’d hear absolute certainty in the response. This particular ambiguity has been the subject of much groaning amongst scholars trying to deal with it. Some of these baffled exclamations are, however, insightful. My choice picks are:

The most irritating feature of play is not the perceptual incoherence, as such, but rather, that play taunts us with its inaccessibility. We feel that something is behind it all, but we do not know, or have forgotten how to see it.”

-Robert Fagen

“Play is one of those constructs that is obvious on a tacit level, but extremely difficult to articulate in concrete terms”

-Loyd Reiber 

This is one of the reasons I like to advocate game designers doing deeper reading into play studies, by the way, because play scholars inadvertently say the most wonderfully interesting things for us games people. Funny that.

Anyway, both these statements are just as true of ‘game’ and ‘fun’ as they are of ‘play’. This particular observation spawned an intellectual digression at the time into Kuhnian paradigm theory. The property of ‘tacit obviousness’, the idea that we somehow just know that all these things are games, or fun, despite not sharing one set of readily observable common properties, is something that many proto-fields suffer during the pre-paradigm and crisis phase of development. Another common property is the attempt to refine a whole bunch of things we now see as linked into separate schools or fields based on massive evidence gathering at a very shallow level- imagine if biology was broken into ‘horseology’ and ‘dogology’ and ‘bearology’, etc. This is not to say that these specilialisations are not rather valuable, we must of course understand the particular inner workings of ponies, puppies and pandas. But to think oneself a biologist when one is in fact a horseologist is… problematic (In the world of the paradigm, one is first taught fundamentals, then specifics. In the time of crisis, one first learns the specifics, then- if lucky- the fundamentals. A paradigm horseologist will be a biologist first, horseologist second. A crisis or pre-paradigm horseologist will be a horseologist first and discover what will later be called biology if they are lucky). I am often irrevocably reminded of this when I see the latest discussion on the nature of immersion or fun based on a new release in (insert game genre here).

Historically, escape from this crisis state has been some really, really smart person coming up with something that seems stupidly simple, but only once you already understand it. See- natural selection, classical physics, the atomic theory, the theory of relativity and so on. While I won’t assert a similar thing will happen around games and play, it’s interesting to note the similarities. Doing so has altered how I deal with the ambiguity of what I work with.

The main change has been to relax my desire to see these terms quickly defined and put to bed, because in a Kuhnian crisis scenario that leads to many problems- namely fragmentation, arbitrary specialisation and blindness to fundamentals as noted above. Like Sutton-Smith, I am for the moment content to accept the ambiguity of these concepts and in doing so, remain very careful about how I apply them generally. Echoing Chris Crawford’s view on fun, it is terminology so broadly used that it cannot be re-appropriated to be a strong design construct. Equally, neither play nor game are valuable concepts for a game designer because their general use and ambiguity is so pronounced that any attempt to reign them in will, instead of being insightful, merely limit the scope of your investigations. This has pretty much been going on since forever, even the granddaddy of all ludology, Huizinga himself, was a culprit. He was rightly taken to task by Caillois for failing to even mention the most concretely influential field of games on earth- games of chance. Caillois, in turn, arbitrarily ignored deep play and so on. It’s basically a story of dominoes, only with definitions.

The other thing I do to cope is to be incredibly self critical. When I think ‘games should be like this’ or ‘this would be a great system’, I force myself to really crack down those operative words. What do I mean by games in this precise instance? What is the measure of great? By ensuring I internally establish precisely what I’m thinking about without loaded and ambiguous terminology, I’m not only more able to ensure that I’m not making some hideously misguided assumptions, but it forces me to clearly see things in the most fundamental terms, which is fruitful ground for making connections to things which might not immediately be obvious and may one day allow me to introduce Jesse Schell to the Mendeleev he’s been waiting for. Every day is a new discovery, because every day I find something that I had taken for granted which is actually built on mist and alchemy when you dig down far enough.

This may sound like a lot of work for the joy of uncertainty but really it has helped me improve my ability to understand and design, with no small hint of ironic emphasis, games. Just because you don’t quite know what you’re doing, doesn’t mean it can’t be amazing. Ask Miyamoto.

Leave a comment

Posted by on March 23, 2013 in Games Theory, Quick Reads


Introducing ERI

ERI is something I’ve been working on for a long time. It’s a personal set of tools I use for all sorts of things and has been incredibly valuable to me as a designer. I was originally going to outline it all in one massive post, but I think given the scope of the framework it’s smarter to do things bit by bit. Also diagrams, because they are the shit.

So what is ERI? It stands for Experimentation-Repetition-Imitation and is a way of deconstructing interactions between players and things they are playing with. It’s an attempt to stand alongside tools like the MDA framework, Koster’s theory of fun and Jesse Schell’s lense system. These are all tools for visualising the complex relationships that result in the experience of play and the entities we call games. ERI doesn’t specifically relate to games, but it was inspired by my investigation of them and is incredibly useful for analysing them

In my studies of game design frameworks, I’ve always found they focus heavily on the object of design, be it software, ruleset or so forth. Game design is primarily shaped by rules gleaned from experience- maintain a high level of feedback, clearly communicate objectives, actively engage the player’s attention etc. However, while there is ususally an explanation for why these rules should be followed, it’s often fairly abrupt. Depth is added by the design community endlessly picking apart individual games and pondering their unique properties. I feel this is something of a legacy of the old days of videogame design, when the task was more to do with software and hardware management than understanding the interplay between a complex interactive object and the people interacting with it. You can bet your ass that the guys who made pong didn’t spend weeks pondering how to optimise player experience. They were too busy figuring out how to make it look like pixels were bouncing off other pixels. Modern game theory is excellent at high-level categorization and analysis, but not very good at reconciling or constructing games at a low level. My history as a gamer of all trades and a researcher into pre-ludology game and play theory has inclined me towards other angles of approach.

ERI is an attempt to provide a new perspective. It serves primarily as a troubleshooting tool, one capable of breaking down interactions into their components. Every action we take is caused by something, and those causes are what the ERI framework spins around. If you can understand what causes someone to seek out powerful weapons in a game, or throw up their hands in frustration, you can control or engineer those situations. These may seem simple tasks, but how about figuring out why a particular part of your game is particularly prone to people griefing, or why you’re getting wildly conflicting feedback on some scene or mechanic?

Like most analysis tools, I’ve found that, having worked for ages developing and figuring out ERI, I can use the same cognitive toolset assertively. Rather than waiting for trouble to show up, I can run prototypes or concepts past a checklist using the ERI framework before I even begin to work. Consequently, I’m able to identify problems before they occur and even, joy of joys, identify and improve on things that are bad, but so nuanced that players wouldn’t be able to pick them out as a specific problem so they would slip right on by playtesting and iteration.

Despite all this, the framework is young. It’s at the stage where it needs to get out into the real world and get beaten into a more robust, practical form. So don’t be afraid to dig in, modify it, improve it and adapt it. I’d like this to become a useful tool for anyone interested in games and play.


I’m not going to go into detail here, since I’ll be posting big crunchy posts every week for a few months on the framework and examples of its use, but I’ll try and give a super quick overview.

The framework breaks down three powerful behavioural motivators, based on established biological and psychological tendencies: we explore and experiment, we repeat patterns of thought and action constantly and we subconsciously copy people who we think are cool. Each of these tendencies has a set of properties that make them somewhat harder to predict than you might otherwise think. For example, when you repeat something a lot, you get better at it. However, you only get better at it in certain specific ways. In some activities, getting really, really good at doing a component can actually make you worse at doing the overall activity.

By focusing on the causes and the subtleties of these behaviours, ERI allows you to take extremely complex interactions and pick them apart quickly and efficiently, isolating the most probable source of an issue. It also trains your mind to avoid making some false assumptions that seem pretty natural until you look really closely at how the gears spin.

Over the coming few weeks I’m going to look in-depth at each of the three behavioural mechanics in turn. I’ll also try and do some design analysis or maybe even an analytical lets play showing how I use the framework to examine designs.  If you bookmark this post, I’ll be updating it with links to the new articles as I go.