Click "Sleep" for a dark background.
Click "sleep" again if text isn't dark.



Mulling Over Multiple Goals

Winning in video games is hard. Generally, the deeper and more complex the game, the harder it can be to achieve victory. To help ourselves, we break down options, tactics, strategies, and gambits and assign a value to them according to how well each will help us achieve victory. Whether risky, difficult, exciting, or mind numbingly tedious, we evaluate all actions according to a value scale. But how do we evaluate game actions when there are multiple goals?


2 Kings? That means there's 2 ways to win? 

If this question seems simple to you, then allow me to illuminate this quandary. Let's start with a clarification on what I mean by goals, multiple goals, and value scales

  • A goal is a gamestate that players strive to achieve in order to obtain victory of a level. Though each game organizes their level structure differently, if a game can be won it must have a goal. In multiplayer matches, each round or match has a goal.
  • Multiple goals is not the same as multiple options, tactics, strategies, gambits, or paths to achieve victory. Instead, a game with multiple goals (multiple ways to win) is one where the designed method for achieving each goal cannot be evaluated on the same value scale. Also, victory is achieved absolutely by achieving any one of the game goals. 
  • Value scales are the measure by which we evaluate every game action. Our scales can be very personal and subjective (how fun or difficult something is) or they can be quantified and universal (how many points you stand to gain or lose). In terms of game goals, the scale must have at least one quantifiable element because the game must recognize the victorious game state when it occurs. 


Before explaining exactly why the valorization (assigning a value) of actions in a multi-goal game is quite unique, I wanted to go over some examples. 


  • Quidditch. As wiki puts it "The team who catches the Snitch wins 150 points, and only the capture of the Snitch will end the game." There's only one way end the game and the Snitch is it. Catching it gives you a whopping 150 while throwing balls through the other goals are only worth 10 points each. Since each Quidditch match and the overall season is determined by points, this game only has one real goal with points as the key value scale.
  • Section 8. This game is mainly a multiplayer FPS that supports up to 32 players. This is one of the most interesting games I know to study multiplayer gameplay dynamics. Basically, there's a wide range of actions that each player can do to earn points. Everything you do from earning kills with specific weapons, to earning assist kills, to spending money on structures/vehicles, destroying enemy structures, repairing structures, healing allies, capturing bases, completing objectives, and preventing the opponents from completing objectives has an assigned point value. The team to reach 2000 points first wins. All the dynamic mission objectives, strategies, and gameplay challenges that are a part of the emergent combat are all measured on the same value scale of points. So even when there are multiple objectives occurring simultaneously, there's only one goal. 
  • In Super Smash Brothers Melee/Brawl or Street Fighter, common tournament rules allow players to do battle within a set period of time. If the time runs out on a fight, whoever has the most health wins. Some people embrace this rule and have developed very effective run away/defensive playstyles. For these games, even though it seems like there are two ways to win (ie. 2 goals), there's still only one. No matter which path you choose, you must have more health than your opponent in the end. Both strategies are dependent on the same value scale of damage. 
  • In Super Mario Brothers, getting to the end of each level is the goal. For the castle levels, Bowser is designed to be the a dangerous obstacle guarding the end. There are only two ways of overcoming Bowser: 1) Touching the axe. 2) Killing Bowser with fireballs. Because doing damage to Bowser has nothing to do with how close you are to touching the axe, these two options are on different value scales. Unfortunately, overcoming Bowser is a level challenge not the level goal. Therefore, this example doesn't qualify as a multi goal. 


The reason I presented these four seemingly apt examples that upon closer evaluation don't qualify as games with multiple goals is to illustrate that games with multiple goals are a bit of a rarity. I was only able to find a few examples including Age of Empires II, Pikmin 2 battle, Dragon Quest Wars (DSi), NSMB DS 2 multiplayer, and Advance Wars

In Advance Wars multiplayer, a match can support up to 3 goals. The most common goal involves destroying all the enemy units. Capturing the enemy HQ is also a common goal. And finally, you can set the match so that whoever captures a set number of properties first wins. Notice that each of these three goals have distinct value scales. In other words, destroying any number of enemy units doesn't necessarily get you any closer to capturing the enemy HQ. Seeking to capture the enemy HQ doesn't get you any closer to capturing enough properties. And capturing properties doesn't help you destroy enemy units. 

Because achieving any one of these goals most assuredly ends the game in victory, how do we begin to compare two actions pursuing two different goals? In other words, how do we compare two unrelated or orthogonal ideas? It's the age old apples to oranges question. And the answer is quite complicated. Basically, in Advance Wars every decision either consumes resources or wastes opportunities. Also, each decision may affect the the conditions and viability of pursuing another goal. This push and pull interplay between pursuing multiple goals is as variegated, deep, and dynamic as the game at hand. So, instead of thinking linearly along a value scale, you have to think more in a flexible 3D graph.  

Unless you understand and think about your moves according to how well they pursue multiple goals simultaneously, you'll probably overlook costly holes in your decisions. In the end, multiple goals can make a game deeper and much more difficult to play. Reaching one goal is hard enough. But playing with multiple goals is like playing multiple games simultaneously. 

« DigiDrive Full Analysis | Main | Elegant & Brute Force Solutions »

Reader Comments (8)

You are wrong!

ALL those game only have one goal: WINNING. The effect is still the same and each of your "fake goal" can be evaluate against that ultimate goal. Winning also have a "shadow goal" >TO NOT FAIL.

Think about it, even a game like mario 64 does not provide multiple goal and non linear progression, star are just a resource to get closer to "completion" and just like old SMB you cannot "go back". The difference is that the progression in abstract (number of star) rather than "physical". Star are the MEAN to advance rather than MOVING to the goal.

Compare to game like the sims, animal crossing, harvest moon, Princess maker 2. Those game have no PRESSURE, there is no winning and no fail (but still punishment and reward). If you do nothing you end up in the worst "game state", and the best "game state" is still harder to achieve AND maintain. And sometimes you have multiple game state to balance sometimes mutually exclusives, sometime not.

If you think about it, these game provide you not with objectives (goals) But subject you answer by creating goals:

In princess maker the subjects is RAISE A GIRL, whatever you do you raise your girl, even if you do nothing actually. There is 72 endings (career the girl get) that depend on how you raise her (from lazy housewives, to princess of darkness or dancer). The game provide you with tools to create your definition of raising a girl and let you observe consequence. Going for a warrior some tools became obstacles (ex that job pay well but raise too much sensibilities) while going for a dancer these same tools became relevant and those you were using for warrior became obstacles (you still need them), if you shoot for a dancer and end up with a warrior you fail, but only on your own perception. Choosing a goal and balancing element to achieve it and answering the "subject" these game are about freedom. The player judge his own progression on his own value.

You cannot have a freedom with a goal from the game, you always have to go for it, whatever the mean and this is the judgement the game pass on you, imposing value, the stake of the "play".

February 6, 2010 | Unregistered Commenterneoshaman

@ neoshaman

It's good that you bothered to respond in detail.

However, I have to say that you are completely off the topic.

You didn't read the article carefully enough, and you didn't understand the definitions I posted.

A goal is more than a objective. In the examples I described, there are multiple ways to win. The open ended, sandbox style games that you mentioned don't really qualify.

There's not much more to say here. Clear and exact definitions are important so two people can talk about the same things/concepts. I have a lot of problems with the way you define/used "goals" in your response.

February 6, 2010 | Unregistered CommenterRichard Terrell (KirbyKid)

Well i have reread everything 3 times, maybe language barrier play a role in the lost of subtlety (i'm not English speaking native) but you have offer better define element in the past. One reason i enjoy your blog is that we can use pragmatically use your thought in a design without questioning it too much and have concrete result! That's how strong you used to be (well not always, like your praise of wii music was exactly your complain about american experience aesthetic!).

However if i understand well, a goal is the end of a certain progression that end the game. They may share resource (the unit you move as example).

The reason the quid-ditch is not a goal in itself is that it's shortcut (bonus) to the main progression (having most points), as well as browser is an obstacle (still need to physically reach the hammer, option are to avoid him or to confronts him).

However in advance war how much the capturing unit is close to the opponent HQ is not the same as coming close to the property capturing count. The progression are independent bu have the same consequence. It's like basically playing "different games" with a share design space and resource.

If i'm right It's essentially a problem of semantics, what i call goal, you call it objectives (and the other way around too). and i use the terms progression rather than value scale.

The reason behind my semantics is simple, games have only one goal in my definition. Achieving a goal is winning. Objectives are "options" on how to obtain the goal, you just need one to beat the game. Most time you can make "objectives" by simply adding an abstract layer of progression and giving an "events" a decisive amount of progression in the value scale (as in quidditch). What you called "goals" in the advance war example, like capturing HQ, is really just "events" (independently of the value scale) were the consequence is winning (my goal, the desired game state), objectives is not the events themselves but "reaching them".

Game are define by "two kind of goals", failure and winning. Because failure is also a progression along a value scale, except there is no desire to progress in that direction, because reaching the end of that progression mean you cannot reach the "true goal". Failure then is not a goal in itself but i don't have a term (neither french or english) to define them (like winning is not exactly the goal).

Every move in a game is define by these two progression (falling, winning) according to a quadrant with the axis desire/fear and the axis losing/obtaining.

That's why i was confuse the first time by your post, i have a very formalize system for game in which i successfully plot the experience of gaming. I May have jump to fast to an answer (despite reading the whole post multiple time). I must admit i had some expectation in finding tangential answer to my struggle with formalizing what i call SUBJECT game.

Sorry if my first answer was a little harsh and rough on the corner, i hope this one clarify anything and spark a good discussion. Anyway, just keep up with your good work, despite blatant Nintendo fanboyism that plague the reflexion... hum... just like me ;) when i discuss game! It's Nintendo fault in making nearly perfect game in terms of design.

February 7, 2010 | Unregistered Commenterneoshaman

@ neoshaman

First of all, I appreciate that you made such a response. You clarified your position well, which should spark a good discussion. Here we go.

You pointed out that we have two different definitions for "goal" not to mention some other terms that are uncommon to each other. Yes, part of our problem is one of semantics.

"However in advance war how much the capturing unit is close to the opponent HQ is not the same as coming close to the property capturing count. The progression are independent bu have the same consequence. It's like basically playing "different games" with a share design space and resource."

This is one of the key ideas I wanted to communicate in my post. Having multiple goals (or independent events) is like playing two different games at once. Naturally this means that deciding what to do with limited resources can be tricky. After all, it's easy to concentrate too much going after 1 goal or spread your resources too thinly trying to cover both possibilities. The trickiest part is, because there are multiple goals/events, it can be far too difficult to weigh how different decisions pursuing different goals.

"Objectives are "options" on how to obtain the goal, you just need one to beat the game."

Based on your response, conceptualizing the entire game as having one ultimate goal is too ambitious and generalizing. This is not to mention how it fails to apply to all video games. Instead of thinking about the video game as a whole product, it's more accurate to think about the smallest games within the product. Think about it. Some games are collections of other games, mini games, and such activities. Even if you can complete 100% of the game's content, thinking of this as the game's only goal fails to account for the rich variety of smaller games (Wario Ware, Wii Sports, Mega Man Collections, Rhythm Heaven, etc).

Beating the game is also not very important to analyzing game structures and how these structures influence players. In actuality, players stratagize for one game at a time to complete one goal/level at a time. So with your terms, we need to focus on a game's events, right?

"Every move in a game is define by these two progression (falling, winning) according to a quadrant with the axis desire/fear and the axis losing/obtaining."

This is one way of evaluating player moves. This method though seems like a more complex version of the simple risk/reward concept. Fear/losing = risk. And desire/obtaining = reward.

My Critical-Rule of Thumb is it's better to zoom in and be clear/specific than to try and understand a game as a whole. Game design layers tend to build up to very complex and interesting systems.

From my understanding, we agree on some of the major points. It's a bit difficult weeding through the particulars of our language.

I'm curious as to what you mean by "SUBJECT game."

If you have any questions or criticisms about anything on my blog or otherwise, feel free to bring them up.

Thanks for the compliments.

February 7, 2010 | Unregistered CommenterRichard Terrell (KirbyKid)

Thanks for the answer :) here is some more (long) clarification (sorry about the length)

Regarding analysis, i have limit game to one goal for a pragmatic purpose. I have build upon pajitnov definition of puzzle (one one goal and one solution), game have one goal and many solution, the next step is SUBJECT game (waiting for a better word, have multiple solution and multiple goal).

Actually to be more precise my model for game is the same model than story, the actantial model of greimas. Only I have extend this model from 6 to many more elements. It is because fundamentally game and story are the same thing, except game create stories, they lie in different end of the same spectrum.

However in my answer i have forget to adress the problem of analysis scope. I don't study game as a product, but structurally, i am aware it depends of the bounds of the analysis (scope of a match or a round as ex.). When i analyse game, i don't see a game as a series of minor games, it's useless, but i do define a granularity to apply the model. Wario ware is less a collection of games than a collection of sequences if we review the main progression as a scope. Each theme is level you can pass towards completion. On the level scope, you have miss as pressure and many "mini game" as a "step", because a particular "mini game" is not needed but the number of success. By seeing game as progression and defining it's objectives, we effectively "layer" the analysis without any ambiguity as every element is define by that progression in that particular scope. It also give you clear delimitation about how a particular layer work into a higher one in a fractal way. By the way the definition (as presented here) imply that all game have a reachable goal but really it mean that game have a "direction", the goal may not be reachable (tetris as exemple as a goal which is place on infinite). The most famous exemple is pacman infamous level 256 which is a not intended ending.

I agree it does not cover all games (actually it does but some only partially). It does not describe fully the "SUBJECT game" i have talk about. SUBJECT game are really true sandbox game. Except that "sandbox" have very sloppy definition and expectation that plague pragmatic thought. Most sandbox game are just a "collection of games (quests)". Completing all the quests is the value scale towards completion, with a main quest as a lower level completion drive. They have one direction and you can only go there or stall. The rules of thumbs is that the fastest way to completion/victory is the value scale.

BUT game like princess maker, elite, the sims and animal crossing or even pokemon have NO direction, they may have some bread crumbing (tom nook debts, beat the elite four and filling the pokedex) but they can be played and enjoyed without them as it's entirely optional. The model still work but fail to address some design choice that structure these games. No wonder that despite their enormous sales, no one copy them because no one have a clue on how they work! And still they does not have or encourage "static space" or "busy work" there is some kind of overall progression of an undefined nature, something hard to grasp and certainly a hint toward paradigm shift.

Subject game seems to concern themselves with giving tools for the player to BE rather than DO, but it's a partial answer i have find, most game have a BE as a backseat experience to justify what we DO. These games do not rely on "quest" that can be completed but on "task" that can be repeat and more importantly these game ANSWER the player action not just with events but with different behavior towards him (acknowledging his BEING in the game). Being in those game is generally "camping" reversibly in a certain "state", they seem to work by pushes and pulls. I don't have a solid framework for that yet, my model need to be extend with a major concept i don't grasp yet.

Regarding the quadrant what about "fear to obtain something" (avoid) and "desire to lose something" (get rid), your example does not directly reward punishment, but more the evaluation of cost that led to it (keeping and obtaining). That is what the quadrant plot: direction of cost of an evaluate action or event. For example, if you have too many items and you need place in your inventory, you evaluate them in term of "get rid", while the skull bonus in bomberman is generally "avoid".

Let's back on that story/game divide to open a door:
Imagine a scene where a victim try to escape a killer. The distance between them became the value scales, if she reach the street were there is enough people to be safe she "win" over the killer, otherwise if the killer is close enough she lose and get killed. Any events that modify the value scale provoke an emotion, fear arise when she trip, relief arise when it's the killer. The distance of the street also provide an estimation of how close of winning she is, while the distance from the killer how close to losing. If she reach the street and it happen to be empty, it reset the progression and provoke tension (sorry our princess is in another castle).

This model (game as progression) does not help to structure game/story only, it also predict the kind of emotion we can have as a spectator or a player. Now game or story like this example are only set on a single interaction layer, the physical one. The model remain the same in the other major layer of stories (moral, psychological, social). Game have try to address those layer too but in a simple way. Imagine a boy which cannot speak to the girl he love because he lacks courage, the distance he must progress on is not physical only psychological, knowing our model it's easy to plot, not only the mechanics but also the experience moment by moment and tuning gameplay regarding the cost of any movement. We can even peek in a random game like god of war which have nothing to do with our thematics, and learn from his dynamics, because the model is abstract enough to be agnostic.

The problem, to rivals with stories as experience, is that stories mix "games" on these different layers interdependently and on many interlocked frequencies within the entire experience. We may reach a new level with game if we successfully answer these question of being and gameplay on different layer of characterization. Hope this last bit is clear.

February 8, 2010 | Unregistered Commenterneoshaman

@ neoshaman

You put a big emphasis on 100% completing games as being a game's ultimate goal. Whether you get all the stars, beat all the levels, or get all the achievements completing a game 100% doesn't necessarily mean a player has beaten all the challenges or experienced everything there is in the game. So, instead of trying to conceptualize a game through steps made toward beating the game, completing the game, or reaching the credits, it's much better to zoom in to the smallest "game" within the game. This means looking at the discrete levels. Even if you take the levels layer and build up to a larger game, it'll be more accurate than starting big and zooming in.

Ultimately, we have two ways of looking at games that are about the same.

>>>>"but they can be played and enjoyed without them as it's entirely optional."

Can't all games be played and enjoyed without the player knowing what the goal is or trying to win it?

When you talk about a game giving tools for the player to BE, it sounds like you're talking about how games can be highly engaging without necessarily providing challenges with goals. Animal Crossing in particular is a game that's like a mirror. Uniquely more so than other games, the mundane "everyday" living in Animal Crossing reflects the player. With people to talk to, letters to write, clothes to wear, and a house to decorate self expression becomes the most engaging element of the game.

By thinking of games like you do stories to help understand how different events affect the player/spectator, you put a lot of focus on emotional experiences. I tend to talk of games strictly according to their mechanics, structures, and function knowing that interactivity/gameplay can create a wide range of emotional responses. Because each player responds differently, I tend to refrain from bringing emotions into the discussion.

I'm afraid that if you don't buckle down and analyze games objectively by measuring the elements/layers of its design, you might get caught up in the emotional, experiential side of things.

I understand where you're coming from and what you're saying. For the most part, I don't have any major problems with your models. Do you have anything written on the internet in English that I could read? Have you analyzed any game that I've played? Can I read any examples of your models?

February 10, 2010 | Unregistered CommenterRichard Terrell (KirbyKid)

Hello sorry for the delay, i didn't refresh in a while and i'm back just out of curiosity.

Well i look at game structurally, the player may drop before he complete the game, can exploit mecanism to create his own game or just left at a design "closure" goal, but structurally the game is design the same. Once the player "consume" the goals there is nothing left in those kind of game. This kind of structure, even if the player is not aware, completly wrap options and play pattern, it define all the function the player will deal with.

I have quit calling OBJECT game or SUBJECT game, instead now i'm saying GOAL structure (focus on progression) or ROLE structure (focus on "management" of a state, but still subject to change). It seems more accurate as game is less define and more cloudy, and can feature those structure mix together, i have listen to your critics about the naming.

Regarding game without challenge, it's easy to see that the superficial elements play a role. But unless you carbon copy the mechanics it's hard to replicate those who are successful. We have mostly NO understanding of going on under the wood. Goal structure, so far, have show pretty steady result irrelvant of dressing up.

About emotion, i will use the joker card and use it to swap emotion with feeling. The kind of "emotion" i made reference to is the feeling of deceleration or acceleration of tension. Whether the player like or not is irrelevant in this case. Because game share with movie goal structure, and because movie rely on the manipulation of feelingsusing that structure, we can actually anticipate and design the particular effect of a mecanism (using my desire quadrant more specifically). "Push your risk" is a pattern that rely on this kind of experiantial feelings, and with that model we can break it even further (ie losing an disadventage, gaining an advantage to proress, getting away of losing or a real progress VS. being closer to fail, losing an adventage or losing progress). The mario kart and smash bros "rubber band items" is another way to turn feeling effect into design. Imagine a game where the health is feedbacked to the player in a fuzzy way, you can induce tension on the player by manipulating the intensity of the feedback where cheap damage can be turn into huge feedback. Because goal structure inform the designer both the internal structure dynamic and the experiential mecanics, it makes for a great tools to design game, it also a great tool to analyse how other kind of experience (movie) play with the audiance by inducing a perception that may change his behaviour. It help anticipate good practice in feedback and manipulate them with design intention.

I don't have yet any information on the net, except for some discussion, you may reach me on mails also. ether.n.aile at

Using my former answer i had open a topic on tigsource about game structure:
That's where i got with goal/role structure.

I'm a pretty big fan of formalization and try to crack serious basis for game. I still don't know how to operate a blog (with customize graphics) or hosting (i m a gamedesign i'm turning indie game maker) to show practical use. I'm always hunting for blog like yours with practical analysis and model. I want to make something as strong as graphic principle for game (see gesthalt theories) and create the equivalent of understanding comics trilogy for games.


February 18, 2010 | Unregistered Commenterneoshaman

@ neoshaman

Good response. You have really good goals you're working toward.

I read through your response a few times, and I think I know exactly what you're saying.

Blogs are easier to start than ever. I suggest starting with Google's Blogger service. It's free and easy.

I plan to write an article about "rubberband" design features like in Mario Kart sometime in the next month or so. It should make for an interesting discussion.

February 20, 2010 | Unregistered CommenterRichard Terrell (KirbyKid)

PostPost a New Comment

Enter your information below to add a new comment.

My response is on my own website »
Author Email (optional):
Author URL (optional):
Some HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <em> <i> <strike> <strong>