Thursday, March 23, 2006

A Theory of Board Game Design: Definitions of Terms

In early 2003 I wrote a series of what would eventually become 20 or so articles on the topic of strategic game design. They appeared in my continuing Skotos column, Trials, Triumphs & Trivialities, and I later reprinted them as a group in my now-defunct RPGnet column, Thinking Virtually.

It's now been almost three years since I finished up the core of the series, and I haven't quite decided what to do with it. I've learned a lot about board game design since, so I could revise them quite a bit for posting here, but on the other hand there's enough material for a book if I could find someone to publish it.

So, I'm leaving those articles in stasis for now, but at the same time I do want to talk about some of the ideas I originated there. Thus, I've decided to write up this article, which is a summary of some of my design ideas, and a general outline of how I analyze board game design, part-by-part. Consider it a definition of terms. (And if you prefer to go read the entire original and somewhat out-of-date series, it's available at Skotos.)

Broadly I think a game can be defined using the following elements: components, activity, decisions, luck, and victory.


Four types of components define most games: environment, tokens, markers, and randomizers.

Environment is the board upon which a game is played, though it may actually be a tabletop, a score sheet, or something stranger. You can have abstract environments (like a chess board), representative environments (like most Euro boards), randomized environments (like Settlers of Catan), evolving environments (like Carcassonne), and more.

Tokens are pieces which are placed upon that environment. They differentiate themselves from environment by the fact that they're dynamic. They enter play, leave play, move, or in some other way change. Broadly they're often "pawns" and they're defined by who they belong to, whether they move, and what they depict. You can have personal tokens, shared tokens, public tokens, dynamic tokens, static tokens, representative tokens, and many combinations thereof.

Markers are game elements which exist outside of the environment and which effectively replace a tic upon a piece of paper. They usually help you keep track of some quantity, such as dollars. You can have scoring markers (VPs, money), trading markers (goods), and more.

Randomizers are game elements which also exist outside the environment, and which in some way introduce luck into a game. See the discussion on Luck, below, for different types of randomizers, but the most common types are random randomizers (dice, spinners) and arbitrary randomizers (cards). Randomizers can sometimes combine with other types of components, such as the Carcassonne tiles, which are arbitrary environment and the Diceland dice, which are random personal tokens.

Tokens and environment tend to be the two more important types of components. By cross-referencing them you can usually say a lot about a game. For example, you could define The Settlers of Catan as using "personal representative tokens on a randomized environment" while some The Seafarers of Catan scenarios instead have "personal representative tokens on an evolving environment". Which is a fancy way of saying you have settlements and cities in both, but in one the board is randomized, but set at the start of the game, while in the other the board is randomized and can grow during the game.


I define activity as the mechanics that describe component interactions in a game. In other words they're the rules that say if you're allowed to change one marker into another ("wood for your sheep?") or if you're allowed to remove someone else's token ("die evil red army!").

I'm not entirely happy with my activity descriptions from my original article, but I still think they inevitably are classified by different ways that you move your components around. Following are some off-the-cuff definitions that I'm using for the moment.

Token activity might include: token conflict, where tokens directly affect each other, like Risk; token movement, where you're moving your tokens, often to some deliberate endpoint; the simpler token placement, where you're placing tokens on-board from an off-board position; or token removal, which is often a type of token conflict, but with a predefined result, such as when a dragon eats a meeple in Carcassonne: The Princess & The Dragon.

Environment activity usually centers on environment exploration (and/or environment placement), where you're trying to figure out what's in an environment and/or take advantage of it, such as in Anno 1503 or Goldland. It could also include environment conflict, if you're having environment in some way fight each other, though I can't immediately think of any games which meet that definition.

Marker activity includes almost any type of logistical game and most resource-management games. I usually classify them as marker collection and/or marker placement. The Settlers of Catan, Parthenon, and many others feature one of these as their main game activity.

Inevitably the activity within a game is defined by activity points (or "action points", to use their more common name), which define how many activities you can take during your turn. There are two degenerate cases which actually define most activity: where you have 1 AP, and can thus do just one thing; and where you have infinite APs, and can do do as many things as you want, subject to resource exhaustion. These cases usually don't actually define their activity as using APs. However, many gamers' games (particularly those by Kramer & Kiesling) make their APs explicit, and give you 2-10 to spend on a turn.


Decisions are what make activity interesting, because they offer different choices for you to make. You can usually define activity as a set of decision sets, each of which has two or more options.

Decision sets can cause Analysis Paralysis, which can be the downfall of a good game, and thus it's important to constrain them and thus make them more manageable. Some of the constraints I suggested in my original article include "constraint by turn phase", "constraint by game phase", "constraint by ability", "constraint by needs", "constraint by attractiveness" and "constraint by results".

The purpose of all of these constraints is the same: to reduce an infinite set of options to decisions sets which have 7 or less options each (using the psychologist's "Rule of 7" as a good thumbnail for what an individual can easily concentrate on). It's much preferred to have a 5-option decision, followed by another 5-option decision, then to have a single 10-option decision, and that's where constraints come in.


Luck is primarily used to determine the outcome of an activity. As I mentioned recently, in my second article on luck, I define four broad categories of luck.

Randomness is essentially selection with replacement. You get a random result from a set that never changes. These are dice.

Arbitrariness is essentially selection without replacement. You get a random result from a set that shrinks as you select from it. These are cards.

Chaos is the way that other players affect you & your plans.

Uncertainty is centered on hidden information, which is to say things that you don't know, but that other players do, which could have affected your decisions.

Dungeon Twister was a game that struck me last year for its claim that it has no luck. But it does: primarily chaos. Your placement of characters and items and your opponent's selection of the same can dramatically benefit one of you or the other through no real strategy. There's also some arbitrariness related to the placement of the room tiles. If my opponent's Dragon ends up on a big open square where it can flame lots of nearby rooms, and mine ends up on a square that's surrounded by walls no matter where he goes, then my opponent got lucky and I didn't.

Which is a long way of say that no dice doesn't mean no luck.


Just as activity is defined by activity points, victory is defined by victory points (and thus victory markers). Some games define a single victory point and some define five or ten required to win, but most constantly give you VPs throughout the game, and you call that a "score". Perhaps more confusingly some games use victory points as a resource within the game too (such as in Oltremare and many others where victory is money).

Victory Points are usually given out for component interactions (or, to put it another way, for certain types of successful component activities). Broadly, VPs are awarded for: token creation, destruction, collection, or movement; environment creation, destruction, control, or exploration; or marker creation, destruction, or collection.

As with activities, my definitions of victories have changed a bit since my original articles, and they're probably not quite settled yet.


This week's article was mainly intended to be a glossary. I suspect I'll refer back to it when I talk about game designs, starting with an upcoming series of articles on the design of Carcassonne(s). I hope I didn't bore too much; clearly defining everything here should pay off in months to come.


Anonymous said...

You wrote that one form of luck is chaos, defined as: "Chaos is the way that other players affect you & your plans." While I agree that others' actions can throw a monkeywrench in the best-laid plans, can we really consider it "luck"? I've always thought that luck required some form of randomness or chance.

Shannon Appelcline said...

I think chaos is definitely randomness.

If a player happens to grab the junction you were building to in _Settlers of Catan_, or happens to want the same tile you do in _Alhambra_ or happens to build in a way that blocks you in _Alhambra_, all with no competitive purpose, that's chaos.

It's fundamentally a lack of control which has an unexpected effect on your strategy, and you can particularly note it in games where the chaos increases notably with the number of players, with _Alhambra_ and _Samurai_ being some prime examples.

Ava Jarvis said...

I like your breakdowns here. I look forwards to your analysis of the Carcassonne family/series, which is an excellent candidate for this sort of dissection.

Your reviews have always been deep and insightful, and I'm going to have to start getting through your older archives. :)

Anonymous said...

Great formal analysis of the basics of games; I've enjoyed reading your older TT&T articles and look forward to the rest of your series here.

As for an Environment Conflict example, how about _Reef Encounter_? If I understand your definitions correctly, the polyp tiles seem to be both tokens, since you place and remove them from the board, and environment, as they build up the board and don't necessarily belong to someone, with the conflict of coral attacks making large changes to the environment.

Shannon Appelcline said...

I actually kept coming back to _Reef Encounter_ whenever I thought about environment conflict. They certainly act like tokens because they come on and off the board, but they *feel* like environment, and you can place other tokens (shrimp) on them.

Pawnstar said...

I've always preferred to use "chaos" as a catch-all term for all "luck" elements (or more specifically anything beyond your decisions or control).

I despise the use of the word "luck"; mainly because it has always had too broad a definition and I got sick of people telling me how "lucky" I was when I had clearly controlled my own destiny.

None of this has anything to do with what you are saying; it's just my semantic preferences. I love all this ludic structural analysis.

CugineTycoon said...

Very informative, and illuminating for Variable changes to circumstances.. I enjoyed this article much . Thank you !!!

I would like to include what is your your thesis on
" Rules are meant to be broken , and broken rules are meant to be fixed "
To variable Changes to Luck , and decision .