Open Fells

The Triangle of Strategy

‘Strategy’ is a vast category that houses wildly different sub-genres: The empire expansion of 4X, the nation management of Grand Strategy, the frantic back and forth of RTS, and many more. These groupings are informal and malleable, just useful enough to give players a broad idea of the experience they are getting into. However, when designing we want our language to be more precise, to have a way to talk about the subtle differences between comparative games and systems. Having some sort of framework for how we understand the landscape of games helps us to question if a mechanic or idea will enhance the experience we are creating, to avoid the pitfall of having systems that are accidentally undermining each other.

Every designer approaches this differently. Soren Johnson of Mohawk Games and Johan Andersson of Paradox Tinto were recently guests on an episode of the strategy game podcast Three Moves Ahead, where they and the hosts discussed an alternative way of mapping out the differences between strategy games, “'Board Game' vs 'Simulation'-style Strategy Games”. This raised interesting questions about the player’s agency and the transparency of mechanics, with the ‘Board Game’-style giving the player a lot of agency and being especially clear about the results of their actions. By contrast, ‘Simulation’-style games offer a wider range of ‘levers to pull’, although the effects may be more subtle and some systems may be hidden, and place the player in the midst of a much larger system of moving parts to create a vastly different experience.

These definitions echo an earlier conversation in the 1950s when there was a schism in the world of tabletop war games, the precursor to strategy videogames, which saw players divided over their approaches to the hobby. Some decried the dramatisation of after-battle reports, saying that descriptions of “smoke drifting across the field” and “wounded soldiers dragging themselves away” were unnecessary details. This was furthered by arguments published in the magazine War Gaming Digest on whether a game should be ‘realistic’ or ‘simple’, with articles such as the sarcastically titled “Is this a hobby or a military exercise?”.1 At the core of this debate was an ongoing question of what the experience of a game should be, with players and designers arguing for their favoured play style. 70 years later and the conversation has thankfully progressed a lot. As well as being much more amicable about the divide than the writers of War Gaming Digest, the podcast also reached a clear consensus that the aim of a designer is ultimately to create an experience rather than to simulate history, even in the more simulation-style games. The conversation was no longer about an argument in finding a single 'proper' way to play, but rather exploring the range of different experiences and types of play that a game can encourage.

Another of war games' children, roleplay games (RPGs), also inherited these same questions and challenges of needing ways to talk about games and experiences that they create. In a small part of The Aesthetics of Play, Brian Upton tells the story of the 1990s RPG forum The Forge, a collection of players and designers “concerned with how RPGs function to construct meaning”.2 Some of their ideas were consolidated in a series of articles by moderator Ron Edwards, which put forth that players are drawn to playing a game for different reasons and each has a different expectation for the experience they want out of it. This is now a well known concept, with many studies and talks on player types, often with personality test styled names like “The Explorer”. However, Edwards distilled it to just three ‘agendas’ that players use: gamism, simulationism and narrativism.3 Upton takes these agendas and adapts them in his own model of how people navigate media through play. He argues that we play by traversing a shifting path of contextual actions to reach desired game-states. The agenda is the player’s approach to play, the reasoning that determines which game-states are desirable or not.

This concept of the three agendas (which I prefer to call ‘styles of play’) stuck with me and has grown into a useful tool in my work for the framing it provides for strategy games and, more importantly, the people who play them. The names and ideas have evolved with time and use, but I present my current version here in the hope that the framework may be of help or may provoke thought about your own vocabulary of design.4


Goal Play

The first of the three styles is the most obvious and the most common. It is “Goal Oriented Play”, or in other words “playing to win”. Most current videogames, boardgames and sports work like this. There is a win state and a lose state, with players testing their skill to try and beat some ‘other’, whether that is opposing players, the computer, themselves, or an arbitrary score. Goal Play can be clearly seen in players who like to min-max, where their decisions are determined by the optimal move to get closer to winning. This requires a well defined goal to move towards, otherwise it is easy for a player approaching a game like this to feel lost and for the game to feel meaningless. It maps well onto the idea of a ‘Board Game’-style game where there is transparency in their rules and systems. For effective Goal Play, in board or videogames, the player needs to be able to understand their options and the impact of their actions so that they can make decisions that move them closer to victory.

Role Play

Whether a character, an entire nation or an ambiguous god-like figure, the player inhabits some kind of role in the game’s world through which they take actions. Role Play is a form of play where players place priority on actions that fit within their own internal constraints of what they think ‘should’ happen or how a character ‘should’ act. It is seeing through other eyes to make decisions. This is possible even without playing as a character; in Historical Strategy games, there is a looming idea of “what happened” and the player can decide how closely they want to lean into their perception of the time and place that the game is set. Even outside of history, there is a sense of what is grounded or believable in the game’s fiction, as players quickly build their own understanding of the rules of the world and the motivations of those within it. Whatever the setting, agency and expression are key. The player needs to be able to act in the way that they feel is consistent to their role and supporting that often requires a wide range of possible actions.

Narrative Play

Rather than acting as a specific part of the system, in Narrative Play the player is writer and editor, seeking to find the most interesting and dramatic situations and building on them to tell a story. It is a more zoomed out style of play, with the player often reviewing events in hindsight. Narrative Play is probably the most difficult type of play to design for as most players have built a literacy of working towards a goal. It feels unnatural to go against instinct and be open to failure, which is a vital part of any good story. This is likely why the best examples of Narrative Play set expectations early, such as the Dwarf Fortress community’s “Losing is fun” slogan.5 Once these expectations are set, Narrative Play is naturally suited to strategy games, where the complex systems of moving parts can create wild and fantastic situations, generating rich narrative points for the player to discover and build upon. Randomness and orthogonal systems help to keep the intersection of different mechanics unexpected and outside of the player’s control to feed their imagination and encourage the brain's incredible pattern-finding capability, apophenia, to kick in.


Every game has some element of all of these. Mapping the three styles of play onto the corners of a triangle creates a spectrum on which to examine games by the styles they encourage and, crucially, what they leave out. This gives a language to speak about the nuanced differentiations between even similar seeming games and systems. The Grand Strategy games of Paradox nearly all take place on historically and geographically grounded maps, encouraging a strong element of Role Play through such a specific setting of time and place. However, within their games there is still a surprising amount of variance. Hearts of Iron has a clear victory condition for Goal Play (winning World War II) and very little focus on Narrative Play (the player may create stories about specific pushes or spectacular moves but little more). Conversely, Europa Universalis is much more focussed on Role Play and navigating the geopolitics of the early modern period, with much less of a sense of playing to win or achieving a particular goal.


Continuing with Paradox’s games, the Crusader Kings (CK) series is a famously fantastic example of Narrative Play, with the dynamic cast of characters leading to stories about the dramatic rise and fall of dynasties (and the occasional horse Pope). For both, the score is of virtually no importance and there is no defined goal. However, the changes between the two most recent games in the series show some of the subtle differences between Role and Narrative Play. In CKII there was a lot of post-choice randomness, with the results of events and the gain and loss of traits often changing due to chance. It was difficult to play into a particular character as a lot of what happened to them was out of the player’s control. However, the chaos created fantastic stories and consistently took the narrative off in surprising directions.

By contrast, CKIII has a strict limit to traits, which rarely change once they are set, and the outcome of most event options are set rather than based on chance. The game also introduced the Stress mechanic, which ties the gameplay into the idea of roleplaying - a character will have certain actions they would prefer to take due to their traits and acting against this will give the character stress. It’s a fantastic way of encouraging roleplay, but the focus being so strongly on one character and reduction of randomness means that it is that little bit more difficult to piece together stories through Narrative Play.


Turning to 4X games, the Civilization series is strongly focussed on Goal Play with light elements of Narrative Play in the story of how the different nations interact. There is very little encouragement for Role Play. Despite being based on history, the setting is abstracted to span the entirety of civilisation with immortal rulers and a randomly generated map each game. Humankind took this base of Goal Play and went even more abstract, with nations swapping culture as they progress through the eras of the game. However, some of the other systems encourage more Role Play through diplomacy and war scores - interesting systems that perhaps try to pull the game in different directions.

Old World has a more specific setting and time period than Civilization, although it is still heavily abstracted with elements such as randomised maps. However, rather than immortal leaders the player rules through a dynasty, playing as a sequence of characters with their own traits and relationships who live and die. Although the player is playing as these rulers, the focus is not on Role Play as might be expected - outside of narrative events there are limited actions that the player can take to inhabit and act as each different Queen or King. The focus throughout is still strongly on Goal Play, with the characters and event system adding texture through encouraging Narrative Play. A war will have a narrative reason it happened, and it is up to the player whether they lean into that or not.


I have said a number of times that games and their mechanics ‘encourage’ these styles. This is as there is no telling how a game is going to be approached; for every example given here, there will be at least some players who enjoy playing with the style that I say least encouraged. Players can also shift the experience towards or expand their preferred style through settings and mods, which are a good anecdotal gauge of the style of play that the players are most invested in and care about.

There are many ways of viewing the divide between different games and systems, the important thing is to find a method that gives you a language to talk about your own design. Approaching games on this triangular scale has personally helped to navigate the subtle differences between strategy games, regardless of genre. It is a bad idea to try to cater to all three styles of play equally - that way lies madness and a mediocre game - but thinking in terms of the experiences our players are seeking and what drives them through the game can help us to find the interesting areas to explore by adapting ideas and mechanics from other genres, while staying true to the essential experience of the game we are making.


“‘Genre’ may be a fine descriptive label for what is or has been done, but it's not much help in terms of what to do or what can be done.” - Edwards6

Until next time,


  1. Peterson, J. Playing at the World, p.299 - Seems to be out of print but can be borrowed from the Internet Archive here.

  2. Upton, B. The Aesthetics of Play, p.254-264

  3. There is an introduction article and then an article on each of the agendas - Gamism, Simulationism and Narrativism.

  4. A note on language - the terms I use are not perfect, especially with the prior existence of Roleplay as a concept and the loose description given by Narrative Play. If you want to trace back the ideas, it would be (from Edwards -> Upton -> me): / Gamism -> Goal-oriented Play -> Goal Play / Simulationism -> Coherence-oriented Play -> Role Play / Narrativism -> Closure-oriented Play -> Narrative Play

  5. The Losing page from the Dwarf Fortress wiki is worth a read.

  6. From Edwards’ GNS and Other Matters of Roleplaying Theory, Chapter 1 on The Forge.

#design #strategy games