Subtractive Design

Subtractive design is the process of removing imperfections and extraneous parts in order to strengthen the core elements. You can think of a design as something you build up, construct and let grow, but it’s pruning away the excess that gives a design a sense of simplicity, elegance, and power.


Make everything as simple as possible, but not simpler.
—Albert Einstein


Make everything as simple as possible, and then a little simpler.
—Hectóre Blivand


First let's look at the theory behind this idea to see why designers in many fields often think in terms of negatives (subtracting things) rather than positives (adding things). Then let's look at several successful subtractive designs so we know what to aim for. Finally, I'll discuss why subtractive design often breeds controversy.

Why Subtraction?

Designers in many fields, not just games, often think in terms of negatives (subtracting things) rather than positives (adding things). Design is creating a form (a game in our case) that fits a context. There isn’t just one boundary we have to check between form and context though, there are infinitely many. Is our game easy enough to learn? Does it have the desired amount of strategy or depth? Does it appeal to the intended age-group? Is it cheap enough to make in both time and money? Is it aesthetically pleasing? Do the aesthetics help the player understand how to play the game? Do the mechanics work well with each other? Do they require the desired amount of dexterity? The list goes on.

We first come up with a design that might fit all the requirements. Sometimes this comes from the intuition of a designer who has internalized all those forces and somehow spits out a new answer. More likely, we start with something pretty well established so that we know it solves many of the requirements already. That’s how genres, sequels, and remakes help us make good (but not necessarily new) designs.

Once we have something, we have to evaluate how good our design is. Does our form actually fit the context? Architect Christopher Alexander had some choice words on this subject in his Notes on the Synthesis of Form:

We should find it almost impossible to characterize a house which fits its context. Yet it is the easiest thing in the world to name the specific kinds of misfit which prevent good fit. A kitchen which is hard to clean, no place to park my car, the child playing where it can be run down by someone else’s car, rainwater coming in, overcrowding and lack of privacy, the eye-level grill which spits hot fat right into my eye, the gold plastic doorknob which deceives my expectations, and the front door I cannot find, are all misfits between the house and the lives and habits it's meant to fit. These misfits are the forces which must shape it, and there is no mistaking them. Because they are expressed in negative form they are specific, and tangible enough to talk about.


Alexander explains that when a misfit occurs, we are able to point at it specifically and describe it. When we instead try to explain what a good fit would be like, we’re often reduced to generalities that are hard to act on.


With this in mind I should like to recommend that we should always expect to see the process of achieving good fit between two entities as a negative process of neutralizing the incongruities, or irritants, or forces, which cause misfit.



When Fumito Ueda designed Ico, he did not start with a list of everything the game should have. Instead, he started with the core idea that it should be a platform / puzzle game about a boy and a girl, and that the game should have emotional impact by creating an environment that had its own believable reality to it. Using other platform and puzzle games a point of reference, he then started subtracting away everything that was extraneous to his core idea.

Other games might use a nine act structure where the story starts in a village, then you go into the forest, then you find a castle, then escape back to the forest, and so on. Ueda was conscious of this, but cut everything except the castle, so that the castle could be fully realized, fully polished, and seem to be a character of its own. Other games have NPCs that stand around and give you hints, but when you see the same character say the same lines over and over, it takes you out of the fictional world. Ueda stated from the beginning that he would have no such NPCs. Other games may have an army of different enemies, but Ueda found that puzzles were enough, and only one type of enemy was needed. He also removed the health meter, inventory screen, and even background music from his design—all things that come standard in other games.

Another misfit that was on Ueda’s mind was bad animation. If the core idea is to show a boy and a girl escaping a castle together, and we want a sense of immersion and reality, then nothing in the boy’s or girl’s animations can stand out as strange. His team spent a great deal of effort on the character animations, especially those where the girl and boy interact, because any imperfections there would have been glaring.

I learned all this from Ueda’s 2004 presentation at the Game Developer's Conference, but there was one more detail that stuck with me. He showed us a screenshot of one room in the castle and asked us what was wrong with it. I thought it looked pretty good. Ueda then pointed out that there was a chair in the screenshot that didn’t look very good. He said when something like this happens, you have to decide whether you have the time and resources to fix it (make the chair more beautiful and believable) or cut it. In this case, he cut the chair.

For all Ico’s cuts, you’d think it would be a game with nothing much left. And yet it received high critical acclaim as powerful game. The important elements are executed unusually well, and the unimportant elements—I couldn’t find any.



Jonathan Blow’s Braid shows a similar reductionist design with a similarly powerful result. The core concept behind Braid is the manipulation and rewinding of time without the need of a meter to limit the mechanic. You can rewind time as much as you like, as often as you like, and the difficulty of the game comes in puzzles that test just how clever you are with this manipulation. Because time manipulation is the core concept, Braid explores time mechanics fully. On one level walking left reverses time while walking right moves time forward. On another, some objects are immune to the time shifting. Each level investigates a new idea.

What’s remarkable about Braid is how many things Jonathan Blow cut away. There are no “lives,” because the entire idea of lives is incompatible with having infinite time-rewind powers. There are only about five types of enemies, much fewer than is usual for a 2D platform game. There are only two action buttons: jump and rewind time (though there is a third button used later in the game to put down an object).

What’s most strikingly minimal of all in Braid is the level design. Each level is as small as it can possibly be, containing as few elements as it could realistically contain while still being interesting. This gives the game’s construction the feel of a tight short story: every part is there for a reason and there’s nothing extra. Even the quantity of levels follows this logic—when the game is done exploring new time mechanics, it ends. It feels no need to make us fight hundreds of blue slimes in order to level up to fight red slimes.

By trimming the fat in action buttons, UI, enemy types, level size, and level quantity, Braid feels vigorous in the way Strunk and White meant in The Elements of Style, 4th Edition, as seen in his discussion of omitting needless words.

Vigorous writing is concise. A sentence should contain no unnecessary words, a paragraph no unnecessary sentences, for the same reason that a drawing should have no unnecessary lines and a machine no unnecessary parts. This requires not that the writer make all his sentences short, or that he avoid all detail and treat his subjects only in outline, but that every word tell.



Valve’s game Portal is another example of a compact, distilled design. The core idea is that you can shoot two kinds of portals (orange and blue) and then walk through one to come out of the other. You have no actual weapons, no inventory screen, no NPCs to talk to, not even any enemies aside from the occasional turret and the final boss. The controls are as simple as they can be, with action buttons only for shooting the two kinds of portals, for jump, and for using objects (open door, pick up crate, and so on).

Portal’s environments are sparse and sterile, containing practically nothing except for elements that are part of puzzles, elements that offer you visual cues as hints about what you should do, and elements to convey the story. That there’s nothing extra puts all the emphasis on the portal mechanic itself, which is incredibly fun. Like Braid, Portal explores its mechanic fully, doing just about everything you can think of to do with portals, then it gracefully ends without overstaying its welcome or subjecting you to filler content.

Team Fortress 2


Valve’s Team Fortress 2 has a lot of things going for it, but it’s specifically the approach to map design that stands out as a case for subtractive design. Most games of this type would offer as many maps as possible. More is seen as better by marketing departments, after all. Valve deliberately limited the game to only six maps when it shipped, though.

One benefit of fewer maps in a multiplayer game is less fragmentation of the player-base. If there are hundreds of maps it can be hard to find anyone who wants to play the particular map you do. But more to the point, Valve knew that in most multiplayer games, the community settles on just a very few maps they play endlessly. If this is a known phenomenon in so many games, why make tons of maps?

By sticking to only six, an unusually small set for this type of game, Valve had time to make these the best, most polished maps they could. Fewer maps means each one received more attention from playtesters, artists, and designers. The process of playtesting a map is, itself, a subtractive design exercise. You play it as much as possible in as many different ways as possible, looking for bugs, exploits, and defects that make the gameplay less fun or less strategic. The more you limit the number of maps, the more defects you can fix in each one.

Game designers should look outside the field of games for inspiration and ideas, so I'll present two examples of non-game software.

Google Chrome

Google has a simple elegance in many of its products, and the Google Chrome web browser is a great example of subtracting the debris that other browsers had. Why do we need two different fields at the top of a browser (one to search the web, one for the URLs) when they could be combined into one? Google Chrome does this to save space and reduce clutter. There is no chance of confusing the two uses in one field anyway, because search terms have spaces between them while URLs have things like “.com” in them.

The core idea behind Google Chrome is “get out of the way and remove debris whenever possible.” The “find” field only shows up when you press cmd-F, otherwise it’s not even there to get in the way. When you mouse over a link, the status bar showing where the link points to shows up in small box in the corner, but this box fades away entirely at other times. Chrome also got rid of these things from Internet Explorer 6:

  • file menu
  • edit menu
  • view menu
  • favorites menu
  • tools menu
  • help menu
  • homepage button (you can turn it on in the options)
  • search button (just type in the URL bar to search)
  • favorites tab button
  • history button
  • mail button
  • print button
  • edit button
  • messenger button (what is this doing here??)
  • the word “address” labeling the address bar
  • status bar

It added these interface elements:

  • menu button for options about the current page
  • menu button for options about Google Chrome in general
  • new tab button.

Does this simpler interface with all the debris removed mean that Google Chrome is less powerful or less advanced than its competitors? Quite the contrary. Under the hood, it separates each tab into a new process, meaning that one pesky website can’t cause your entire browser to crash. It also does a better job of preventing memory leaks than its competitors with better memory garbage collection. The power is out-of-sight, and the browser UI is as minimal as possible so that everything is there for a purpose, everything with no purpose is gone, and the things that are there are high quality.

Other browsers have since become more like Chrome. Unified search/URL bars and sandboxed tabs are now standard.

Apple's Time Machine

Apple also has a long history of simple and elegant products. It might be the best company in the world at making complicated things simple and elegant. Apple has done that several times by crafting new kinds of interfaces. They created the first personal computer to use a mouse. The iPod’s scroll wheel was a simple way to scroll through hundreds of songs. The breakthrough of the multitouch interface on the iPhone shook up the phone industry. I’ll use a less familiar example though: Apple’s Time Machine software.

The core concept behind this built-in part of a Mac’s operating system is “make it so easy to back up your data that you’ll actually do it.” The problem with data backup software is not that it doesn’t do enough things, it’s that the average user is too lazy to ever actually do it at all.

Apple’s solution is a “zero click” interface, though maybe it’s more fair to call it one click. When you plug in any drive, a pop-up asks you if you’d like this to be your Time Machine drive to back up your files (it doesn’t ask if you’ve already set a Time Machine drive, of course). If you say yes, that’s all there is for you to do. Time Machine will then back up all your computer’s files and keep running backups every hour for the last 24 hours, daily backups for the last month, and weekly backups forever, until the drive is full.

The interface for recovering old files is slick and useful, but it’s the “zero click” setup that makes the feature so practical. Now that's subtractive.

Common Themes

There’s something all these examples have in common. Apple’s Time Machine and Google Chrome are both very sophisticated under the hood, even though they present simple interfaces to the user. Team Fortress 2 is not more shallow for its decision to launch with fewer maps; it’s actually deeper because of that decision. Portal and Braid did not, as Professor Strunk would say, “avoid all detail and treat their subjects only in outline.” Quite the contrary. Portal and Braid each fully explore their concepts—more fully than most larger, bloated games usually explore theirs. And finally, Ico’s sense of reality, immersion, and emotional power is not less because it subtracted all the extraneous elements; it's more. In each case, subtracting did not leave us lacking, it enhanced the experience.

The Controversy

Subtractive design is not all rainbows and puppies though. By fully committing to this idea, you are more likely to encounter resistance on your game development team, with your publisher, and with your players. The reason is that when we use vague language, it’s easier to get an agreement. When we use very honest, precise language, it’s easier for someone to realize that they disagreed all along.


“Some amount of collateral damage is expected in the mission.” Sure, ok.
“We are going to kill innocent people on this mission.” Wait, really?


When we distill a design down to the core concepts and remove the extraneous, it forces us to admit and agree what the core concepts actually are. For example, as designer of Street Fighter HD Remix, I made the statement that performing difficult moves is not part of the core concept of the game. It’s an imperfection that should be removed, so that there can be more focus on the essence of the game: strategy. Clearly, that is a troublesome statement if you believe that performing difficult moves is part of the essence of the game. I think subtracting some emphasis on that aspect enhanced the final product though.

Likewise, in StarCraft, the ability to play not just fast, but extremely fast is highly rewarded. When Blizzard made StarCraft 2, it had to decide what the game is really about at its core. Maybe a game in the “real-time strategy” genre should focus a bit more on strategy and less on extremely fast clicking? Creating more user-friendly features such as multiple-base selection, auto-mine, and no cap on the number of units you can select all point to them thinking that. (Though they also added other features intentionally meant to add more clicks, so I don't know.) The point is that during development, the mere proposal of multiple-base selection and automine raised deep questions about what the the game is really about. Is StarCraft really about strategy? Or is it equally about rewarding the most actions per minute that you can enter? It’s easier to agree that we like StarCraft overall (vague) than it is to agree on whether a new version of the game should or shouldn’t remove the emphasis on certain skills.

The card game Magic: the Gathering has an even deeper conflict about what it’s really about. When it comes to the wording on the cards, the Magic team has made great strides over the years to remove unnecessary words, creating as many simple, elegant cards as they can. Compare the original wording of the card Control Magic to the current wording:

But on a more zoomed out level, what is Magic really about? Is it about delivering the most fun gameplay experience possible to its players? Or selling collectable items that have artificial scarcity? One gets in the way of the other, as it stands. I propose that the essence of customizable card games is the gameplay, and that collectability is purely a barrier between players and the game. Making such a statement naturally creates a firestorm of argument because it forces us define what the essence of a game is. That can be uncomfortable to do, but necessary for knowing how to design something.

For my customizable card game called Codex, I'm subtracting out all the chaff cards, and subtracting out the entire concept of collecting rares. It delivers the most gameplay possible in the fewest cards, which makes it the opposite of the standard approach of bloat in CCGs.

Closing Thoughts

It’s easiest to get people to agree on vague concepts. “The game we’re making is a platformer with exploration, but also with fast action and time pressure. It has an epic story of course, but also a personal story. The game is really challenging, but it’s for everyone to enjoy. It has 20 enemy types and 20 weapons including a kazoo and a kitchen sink.” What’s not to agree with? There’s something for everyone.

A more direct approach would be to find a starting point, whether it’s another game’s design or something you generate yourself. How to do that is another topic entirely. But once that core concept is in place, stay on the lookout for things that get in the way. Is every button press really needed? Every menu item? Every HUD element? Are there features that you just don’t have time to do justice? (Let’s add multiplayer!) If the game is about testing player skill, is it testing the only the skills you want to test? If it’s about story, are all the scenes contributing to that story? Do you need all the mechanics you planned, or will a smaller set (easier for the player to remember and learn) suffice? Is there a way to make your levels smaller or shorter by removing or shrinking areas that do not have much purpose?

Getting rid of all that stuff means there are fewer things for the player to misunderstand, and makes it more likely that the vision of the game in your head actually ends up in the player’s head, too. It takes some courage and pain to commit to a specific idea and subtract the rest away, but I think Ico, Braid, Portal, Team Fortress 2, Google Chrome, and Apple's Time Machine all demonstrate that doing so can lead to powerful, memorable experiences.

Slippery Slope and Perpetual Comeback

If a game has slippery slope, it means that falling behind causes you to fall even further behind.

For example, imagine that every time your team scored in basketball that the opponent’s team lost a player. In that game, falling behind is doubly bad because each basket counts for score AND it makes the opposing team less able to score points of its own. The actual game of basketball does not have this screwy feature though, so real basketball does not have slippery slope. Scoring in real basketball puts you closer to winning but does not at all hamper your opponents’ ability to score.

Slippery slope is another name for positive feedback—a loop that amplifies itself as in a nuclear reaction. Because people confuse the terms positive and negative feedback so easily, I prefer the more descriptive term slippery slope.

Slippery slope is usually a bad property in a game. If a game has a powerful slippery slope effect, that means that when one player gets a small early lead, they are more likely to get an even bigger lead, which in turn makes them more likely still to get yet an even bigger lead, and so on. In a game like this, the real victor of the game is decided early on and the rest of the game is futile to play out (or to watch).

Examples of Slippery Slope

StarCraft and Chess both have slippery slope. They manage to be good games anyway, despite this anti-climactic property.


In Chess, when a player loses a piece, their ability to attack, defend, and control space on the board is slightly reduced. Sure, there are many other factors in Chess—positioning, momentum, pawn structure—that determine if a player is actually “losing,” but losing a piece does have an effect. Clearly, losing a lot of pieces, say 8, puts a player at a significant disadvantage. It’s pretty hard to make a comeback in Chess, and a game is usually “won” many, many moves before the actual checkmate move.

This is why there are a lot of forfeits in Chess. Good players don’t actually play out the pointless part of the endgame when they recognize the opponent will definitely win. Chess players would say that forfeits being a regular part of the game is fine and not awkward, but it’s a disappointing quality compared to games without slippery slope.

Chess 2

Chess 2 doesn't have the problem of anti-climactic slippery slope situations because of the midline invasion rule. You win if your King crosses the midline of the board. (You can also still win by checkmate.) This means the game ends before one player falls too far down a slippery slope. It's better to be put out of your misery than stuck in a hopeless situation for many turns.

StarCraft (and StarCraft 2)

StarCraft also has slippery slope. When you lose a unit, you are penalized doubly. First, you are closer to losing (having no units at all is so crippling as to be virtually the same as the actual loss condition of losing all your buildings). Second, you are less able to attack and defend because the unit you lost was not just part of a score, but also part of the actual gameplay of attacking and defending.

StarCraft has even more severe slippery slope when it comes to the game’s economy. Imagine that your opponent rushes you (sends an early attack to your base) and you fend it off. Let’s say you each lost about the same value of units in the exchange, except that you also lost one worker unit. In a different type of game, this might equate to being one “point” behind. But in StarCraft, that can be a crippling loss because gathering minerals is nearly exponential. Your opponent is ahead of you in the resource curve, increasing their earnings faster than you are. You’ve fallen down a very slippery slope here, where an early disadvantage becomes more magnified as the game goes on.

In basketball, the score is completely separate from the gameplay. Your ability to score points doesn’t depend at all on what the current score is. You could be ahead by 20 points or behind by 20 points and have the same chances of scoring more points. But in StarCraft (and Chess), the score is bound up with the gameplay. Losing units pushes you closer to loss AND makes it harder to fight back.

Fighting Games

Fighting games don’t usually have slippery slope. In Street Fighter, for example, your character still has all of his moves even when he’s about to lose. Getting hit puts you behind in life totals (in “score”) but doesn’t limit your gameplay options in the way that losing a piece in Chess does or losing a unit in StarCraft does. An unusual example of a fighting game that does have slippery slope is Bushido Blade. In that game, getting hit can cause you limp around or lose the use of an arm. This is extremely rare in the fighting game genre though, and for good reason.

While it might be "realistic" for a nearly dead character to limp, move slowly, and have generally less effective moves, it's not fun. (At least in Bushido Blade's case, this part of the game lasts only a couple seconds, then you lose.) Meanwhile in Street Fighter, comebacks are frequent and games are often "anybody's game" until the last moment. Street Fighter does have some very minimal slippery slope aspects (if you're very near death you have to worry about taking damage from blocked moves which aren't a threat if you have full life), but overall it's pretty "slippery slope neutral."

There is one fighting game that stands out as an exception: Marvel vs. Capcom 2. In this game, each player chooses 3 characters. At any given time, one character is active and on-screen, and the other two are off-screen, healing back some lost energy. The off-screen characters can be called in to do an assist move, then the jump off screen again. The main character can attack in parallel with the assist character, allowing for a wide variety of tricks and traps. The player can switch the active character at any time, and he loses the game when he loses all three characters. But here, slippery slope rears its bitter head. When one player is down to their last character and the other player has two or even all three of his characters, the first player is at a huge disadvantage. The first player has can no longer attack in parallel with his assists, which often means he has no hope of winning. Comebacks in MvC2 are quite rare and games often "end" before they are technically over.

This was specifically addressed in Marvel vs. Capcom 3 by adding the X-Factor mechanic. That's a mode you can activate once per game that powers you up a huge amount for several seconds. It powers you up more the fewer characters you have left. Activating X-Factor when you're donw to your last character is so powerful that you can reasonably hope to defeat one or two of the opponent's characters before your X-Factor runs out.

Fighting games with ring out such as Virtua Fighter and Soul Calibur are especially good at avoiding slippery slope properties. In these games, a player instantly loses if his character is ever pushed out of the ring, no matter how much energy he has. Basically, no matter how far behind you are, no matter how close you are to losing, you always have a 100% damage move: ring out. Long ago, I thought this concept was "cheap" and served only to shorten games while adding little benefit, but actually the threat of ring out adds quite a bit to both these games. Since the threat of ring out is so great, another whole element of positioning is added to the game. A player must fight both to do damage to his opponent, and fight for position to avoid ring out.

Limited Slippery Slope

Fighting games do have a very localized, limited kind of slippery slope that’s actually a good quality. If a game truly has no slippery slope whatsoever at any point, then it can feel like a series of disconnected decisions. It’s interesting though, if a decision you make at one point in a game echoes forward through time, and can influence later moves in the game. The problem is if this influence is allowed to snowball into a greater and greater advantage.

In limited slippery slope, there is a cap on how far you can slip and the effect is temporary. In Street Fighter, getting knocked down (hit by a sweep) does have a bit of slippery slope. You lose health (“score”) but you also have temporary limitations on what your character can do. Your character falls down, then gets up into what is usually a disadvantageous situation. The two things that are important about this are: 1) after the knockdown is over, you regain all your moves and 2) you cannot get doubly knocked down. (The card game Yomi features this same kind of knockdown mechanic, which also has limited slippery slope.)

Hitting the opponent with a sweep does echo forward through time, but this advantage is reset soon after and can’t snowball into “getting REALLY knocked down” because there is no such thing as degrees of knockdown. If you are already knocked down, you can't be knocked down "even more."

Another example is backing the opponent into the corner (the edge of the stage). If you do this, you have a natural advantage because the opponent has fewer movement options. But again, there’s a limit here. Once the opponent is in the corner, they can’t be “more in the corner.” There’s a limit to how disadvantaged they can get.

An even more basic example is anytime you block a move that has a fair amount of recovery. In these case, you recover from your blockstun before the opponent recovers from their move, so you have a few frames to act first. This gives you an advantage because if you both try to do a move of the same speed, yours will win (it will start first). Your good decision to block echoed forward into the future, but the effect is very fleeting. Even one second later, this advantage fades.

So fighting games are full of small, temporary slippery slope effects that actually help the game. And yet, on the macro level, they do not have the real kind of slippery slope, the permanent kind that snowballs until the game ends. Compare this to Chess where you don’t just get your captured pieces back a few turns later.

An RTS Without Slippery Slope

Here’s an idea for turning the full-on slippery slope (usually bad) into the limited kind (usually good). Both players start with the same amount of resources to buy units. When your units are destroyed, your resources are refunded. A delay in the timing of this refund combined with the build-time for making new units means that losing units really is a disadvantage, but that the disadvantage fades over time, similar in nature to getting knocked down in a fighting game. The real-time strategy game World in Conflict does exactly this, but I’ve never actually played it.

My point here isn't about whether World in Conflict is a good game, or even whether the exact refund system stated above is good. It just shows that it is possible to remove slippery slope from an RTS if you try hard enough. Someone very dedicated to that problem could probably come up with an even better way to remove it that results in a deeper game, rather than a shallower one.

Perpetual Comeback

The opposite of slippery slope, I call perpetual comeback. That’s just a more descriptive term for negative feedback. (Also, negative feedback sounds like a bad thing, but it’s usually a good quality in games, so it’s helpful to have a term that doesn’t sound negative.) A thermostat uses negative feedback to keep the temperature of a room from spiraling out of control.

Perpetual comeback, then, is a quality in which being behind actually gives you an advantage. I’d like to draw a distinction between two types of this effect, though. In one, when you are behind, a force pushes on you to help improve your position. An example of this is the Fatboy mutator in Unreal Tournament. In that first-person shooter mod, when you kill an enemy, you become fatter and easier to hit. When you die, you become skinnier and harder to hit. Multiple hits magnify the effect, so if you die over and over you get skinner and skinner. Note that even if you die a lot, you are still losing (your score is not helped), but you do have an advantage (harder to hit).

A similar example is any version of Mario Kart. The further behind you are, the more powerful the items you get. In last place, you can get the powerful blue turtle shell which has homing powers to zero in on the first place racer. Meanwhile, the first place racer gets only weak items.

Advance Wars: Dual Strike on the Nintendo DS has a similar feature. Each side has a powerful “tag attack” that’s tied to a meter. When you get attacked, your meter fills up at twice the rate as usual, so the losing player will have faster access to this powerful attack, giving them a chance to make a comeback.

In all three of these examples, the games have a force that help out players who are behind and hinder players who are ahead. This is generally a good type of force to have, because it makes games closer, and small early mistakes are not crippling. That said, maybe the effect is too extreme in Mario Kart, or maybe it creates strange artifacts such as avoiding 1st place on purpose for most of the race. And the power of the tag attacks in Advance Wars might be too extreme, making them dominate the game. Tuning issues aside, the concept is still sound and when it’s done right, it can make matches closer and more exciting.

Perpetual Comeback Extreme

There is a different type of perpetual comeback that is far more extreme and far more rare. That’s when getting closer to losing doesn’t JUST give you helping hand, but instead actually puts you ahead. I think the best example of this strange property is Puzzle Fighter.

Puzzle Fighter is, in my opinion, the best competitive puzzle game ever made and I felt that way long before I was lead designer of Puzzle Fighter HD Remix. The game seems standard enough—it's one of those games where each player has a basin that pieces fall into. There are four different colors of pieces, and you try to build big, single colored rectangles (power gems). You can then shatter those rectangles with special pieces called crash gems. The more you break, the more junk you drop on the opponent's side. When your side fills to the top, you lose.

Several factors come together to create perpetual comeback (the extreme version!) in Puzzle Fighter. First, each "character" (there 11 to choose from, including secret characters) has a different "drop pattern." A drop pattern is the pattern of colored blocks that a character will send to his enemy when that character shatters blocks on his own side. For example, Ken's drop pattern is horizontal row of red, followed by a horizontal row of green, then yellow, then blue. Every time Ken sends 6 or fewer blocks to his opponent, he'll send a horizontal row of red. Every time Ken sends 12 blocks, he'll send a row of red, then a row of yellow. Since the enemy knows this, they can plan for it. They can build their blocks such that Ken's attack will actually help rather than hurt. There's one catch: when you send blocks to the opponent, they appear in the form of "counter gems," which can't be broken immediately by normal means, and can't be incorporated into deadly power gems. After about 5 moves, the counter gems change into regular gems.

The other very critical property is that power gems broken higher up on the screen do more much more damage (send many more counter gems) than gems broken at the bottom of the screen. So consider what attacking is actually like in this game. Attacks are really only temporarily damaging, until the counter gems turn into regular gems. At that point, the opponent will probably be able to incorporate the gems into their own plans, since the opponent knows your drop pattern. Even if the opponent isn't able to benefit from your attack in that way, they can still "dig themself out" of trouble by breaking all the stuff you sent them. By filling up their screen most of the way you've basically given them more potential ammunition to fire at you. What's more, as they are nearest to death, their attacks will be the most damaging due to the height bonus. Gems broken at the very top of the screen do significant damage.

Puzzle Fighter has the extremely unusual property that "almost losing" looks exactly like "almost winning." Let's say you break a whole slew of power gems and send a large attack at your opponent. Your screen is now almost empty. You're winning right? Their screen is nearly to the top—almost full. They're losing, right? Well, they are on the verge of losing, but they have all the ammunition and they have the height bonus, whereas you have almost nothing left to defend with. In effect, your opponent is both "losing" and "winning" at the same time. Very curious, indeed!

It turns out the best way to play Puzzle Fighter is to not attack until you can make it count. All those little jabs you do just help the opponent in the long run. You've got to save up for a huge, 1-2 punch. You need to send a big attack that almost kills them, then immediately send another attack that finishes them off. 1, 2! The point is that Puzzle Fighter is a high energy, edge-of-your seat game. Your opponent very often has enough attack to kill you, so you have to have enough defense to stop them. Whenever the scales start to tip in your opponent's favor, they have also, weirdly, tipped in your favor as well, in some sense. A game of Puzzle Fighter is never over until the last moment. Comebacks are the name of the game, and the excitement goes to the very last second almost every time.

I liked this comeback mechanic so much that I included it in my tabletop game Puzzle Strike. In Puzzle Strike, you have a gem pile that fills up by 1-gem automatically every turn. You can crash those gems to send them to other players. If you end your turn with your gem pile totaling 10 or more gems, you lose. So the more gems in your gem pile, the closer you are to losing, BUT it also means you have more ammo to fire at other players and you get the benefit of the height bonus rule. The more gems you have (the "higher" your gem pile), the more chips you get to draw each turn. That lets you do bigger combos and enables you to crash even more gems to save yourself.

So Puzzle Strike has a similar perpetual comeback mechanism to the one in Puzzle Fighter. In each case, you're the most powerful right before you lose.


Slippery slope is a force that punishes players who fall behind, making them even more likely to fall further behind. Left unchecked, this makes for matches where the real victor is decided long before the game actually ends, leading to either boring endgame play, or lots of forfeits. While fighting games lack this overall slippery slope, they do have several forms of temporary, limited slippery slope that improves gameplay. This limited slippery slope probably exists in other genres as well, but could be a conscious design choice for future games. Finally, perpetual comeback, the opposite of slippery slope, is a force that helps losing players and puts the brakes on winning players, making for close matches. This property can easily go wrong if tuned improperly, but if done well, it leads to closer, more exciting matches. Puzzle Fighter and Puzzle Strike take this concept to an extreme, by making winning look almost the same as losing.

The Fun of Super Mario Galaxy

I've only said "Wow!" a few times in the last couple decades of playing games. One of those times was for the breakthrough Super Mario 64, a game that took action/platforming into a 3D world and made it work. It's fitting that I said it again over its (true) sequel, Super Mario Galaxy, a game that took action/platforming even more into 3D and made that work, too.

Why is Mario Galaxy so good and what can we learn from it? To borrow some terms from Nicole Lazzaro's four kinds of fun, Mario Galaxy has hard fun, easy fun, and social fun as well as the ability to evoke the emotions of surprise and wonder.

Hard Fun

Gamers know this kind of fun all too well. This is the fun of overcoming obstacles and attaining goals. When you succeed at an especially difficult challenge, the Italian word fiero describes the emotion you feel as you raise your fist into the air triumphantly. Mario Galaxy has 120 stars to collect, offering plenty of this type of fun.

Hard fun is so common in games that the only thing worth noting here is how well Mario Galaxy informs the player about exactly which goal he's going for, which goals are completed, and how many goals are left. I think this clarity magnifies the fiero aspect of the game. Putting the tally of hard fun at center stage (the number of Mario Stars, out of 120, you've collected) makes it all the more satisfying to achieve the goals.

Easy Fun

Ironically, this fun is much more rare in games. This is fun that's not bound up with winning or goals. The entire Nintendo Wii system has an advantage here because the motion-sensing Wiimote lends itself to easy fun.

Collecting the star bits (the colorful, glowing ammunition that bounces around everywhere) with the Wiimote's pointer is easy fun. Shooting the star bits at enemies is easy fun, though hardly ever required to achieve goals. Using the left-right-left-right gesture to do the spin attack is easy fun.

Another part of easy fun is exploration and variety. Some of the gameplay variety in Mario Galaxy includes:

  • Flying with the bee suit
  • Shooting fireballs with the fire suit
  • Creating frozen platforms and ice skating with the ice suit
  • Becoming a ghost who can turn invisible and float with the ghost suit
  • Jumping very high with the spring suit
  • Riding a manta ray on the water in a race
  • Riding a turtle shell underwater in many situations, including races
  • Balancing on a ball as you navigate through a level
  • Flying with the red star suit
  • Numerous tricks of gravity that vary across several levels

Just the moment-to-moment interactions involved with these things are fun, without even considering how they are used in the context of hard-fun-goals.

Social Fun

Mario Galaxy is primary a one-player gamer's game (lots of hard fun), but it includes a brilliant two-player feature that will surely become a standard. Some dismiss this feature as "tacked on," but something that strikes such an exactly correct note was surely a carefully considered feature. The two-player co-pilot feature is intended for a non-gamer to enjoy the game alongside a gamer. It adds a lot of social fun to a game that would otherwise have nearly none of that kind of fun.

The second player uses their own Wiimote, but does not use the nunchuck add-on (what non-gamer would want to anyway?) The second player gets their own cursor on-screen that can collect the many star bits littered throughout most levels. The second player (as well as the main player) can shoot these star bits at enemies. The star bits are basically like a shared pool of ammunition, and the second player can add to that pool and deplete it by shooting. 
The greatness of this feature is in the details. First, the main player never actually needs the help of a second player, so this isn't like forced grouping in an MMO. Also, the second player can enter and leave the game at any time without any annoyance or stop in the action. When the co-pilot is helping, they feel like they are contributing because collecting star bits and shooting enemies is at least somewhat helpful.

Also, there are several times in the game where a special NPC appears who asks you to contribute a bunch of star bits in order to unlock a new level. This means you can't completely ignore collecting star bits, and again, the co-pilot is contributing by collecting them. There are certain times when the main player is too engaged in hard fun platforming to be able to collect star bits at the same time, and this is yet another situation where the co-pilot can contribute.

The role of the co-pilot is kept from having too much impact because shooting enemies does not actually kill them (it momentarily stuns them). Also, even without a co-pilot, any hardcore gamer worth his salt would be able to get enough star bits that no co-pilot is needed. But the non-gamer co-pilot doesn't know that!

Finally, the co-pilot's role consists entirely of easy fun. There is no way to actually fail at anything as a co-pilot. You just collect star bits whenever you feel like it, and shoot enemies if it seems like it would help.

If at any point your co-pilot would prefer to sit there and do nothing or put down the controller and check on the stove, that doesn't cause any problems. Because the co-pilot has no pressure, it's easy to suck in a non-gamer. You get their in-game help, you get their observations about where a secret might be hidden, and most importantly, you'll actually communicate back and forth about things (aka social fun).

Surprise and Wonder

In addition to these types of fun, Nicole Lazzaro also mentions several types of emotions that come up in games. I already mentioned fiero, the emotion you feel when you achieve something difficult. Mario Galaxy also creates the very rare game-emotions of surprise and wonder. That's quite an accomplishment considering the game's genre is well-worn territory, but the twists on gravity are interesting enough that sometimes you just sit back and say "wow, that's cool!"

The surprise part is that you feel the wonder part several different times as the gravity tricks change. Just when you thought it was cool to run around the surface of spherical planets, you get to a room where "down" can potentially be any surface.

Then a level where you can flip switches to change the direction of gravity. Then a level where moving spotlights determine exactly where a different direction of gravity is "shining" on an otherwise normal level. There are enough surprises to go around, and I've already ruined most of them for you.


There's a few pet issues I'd like to point out that Mario Galaxy does right.

Opening Sequence

The opening sequence is only a few seconds long, before you get to actually move around. Compare this to over 15 minutes in the excruciating opening of Paper Mario. The original God of War has an opening sequence of less than 60 seconds, showing that there are better ways of conveying story than forcing the player to watch a long cut-scene before a game starts.

Mario Galaxy conveys some of its story simply by showing speech bubbles over NPCs as you run by. You also unlock chapters of an in-game storybook as you progress through the game and you can read them whenever you want or not at all.

Inertial Frames

When you jump straight up while riding a train in real life, you do not slam into the back of the train; you land on the same spot as you jumped from. Physicists say that you are in the same inertial frame as the train, meaning that you're moving with it and your walking or jumping is relative to it.

You all know this instinctively and yet almost no platform games know this. I remember actually being shocked in the game Spider-Man 2 when my Spider-Man was on top of a car and I jumped straight up and landed on the car. "Wow, they know about inertial frames!" I said. At long last, Mario Galaxy knows about them too. You can finally jump straight up while riding a moving platform and land on the platform without worrying about it moving out from under your feet. 

Wall Jumping

This is a small thing, but points to an important idea. In Mario 64, the wall jump move required good timing. You had to press jump just as Mario touched the wall, no sooner and no later. In Mario Sunshine, The New Super Mario Brothers, and Mario Galaxy, it no longer requires timing. When Mario touches the wall, he starts to slide down and you can press jump at any point during the slide to activate a wall jump.

Someone might say that the original harder wall jump was better because it "required skill." No one actually says that, though. Being able to do your moves is fun, and Nintendo realizes that making a move hard to do is a bad way to add challenge. Even when it's easy to execute a wall jump, there can be plenty of difficulty coming from the level or the situation you're in.

Maybe you have a time limit, or maybe there are some flame jets you have to wall jump past, or a hundred other things. Incidentally, this is the same logic I'm using in making the moves easier to perform in Super Street Fighter II: HD Remix. The moves themselves aren't meant to be the source of challenge, it's how and when you use those moves in the context of the game that's challenging.


I used to think that moving the camera around while you are in the middle of platforming was part of the game in Mario 64. I was good at this, and I considered it one of the skills the game was all about. Mario Galaxy removes this "skill" almost entirely because it has an amazingly good camera system. Almost all the time, the camera is pretty much where you want it to be. This is a similar concept to the wall jump mentioned above, in that the game is much better off creating difficulty in other places than wall jump execution or camera fiddling.

Mario Galaxy's camera is actually an amazing accomplishment. I saw a GDC lecture one year about camera systems in games from the guy who did the camera for Metroid Prime. That game also has excellent camera handling (and the best mini-map ever). You might say, "But it's a first-person shooter! There is nothing to the camera."

What you don't realize is that Metroid Prime has over 20 camera modes. When you're in an open area, it's a regular first-person camera. When Samus rolls into a ball, it's third person. Some ball-rolling areas have a side-view camera and basically turn the game into 2D gameplay. Going through a tunnel has a special camera, and some boss fights have another camera. 
A Mario-style third person platform game has even more demanding camera needs than Metroid Prime. In 1996, I would have not even been able to imagine a camera for a 3D Mario game that was basically in the right place almost all the time.

When you consider that Mario Galaxy presents far more challenges to camera design than any other 3D platform game ever, it's that much more impressive that it succeeds. No matter which way gravity is going or which kind of crazy thing you're jumping around on, the camera seems to know where it should be. This is undoubtedly the result of endless hours of hand-tweaking of camera paths and some very smart logic to boot.

Even More Excellent?

It's a real jerky thing to take an excellent game and say, "I'm knocking it because it wasn't excellent in some other area that it didn't even attempt." I already cringe at that being done to me someday, so I apologize in advance for this, but I do wish Mario Galaxy were even more excellent.

Before I say what that is, I'll tell you what I think is one of the best surprises in a video game. I mentioned this in my Pacing for Impact article, and I'm about to say the same spoiler now, for Castlevania: Symphony of the Night.

When you beat that game (which only takes about 10 hours), you are lead to believe that you really reached the end. The game has been showing you what percentage of the map you've uncovered and it gets closer and closer to 100% as you work your way toward the final boss.

After you defeat him, the big reveal is that the entire castle where the game takes place turns upside down, and you have as much more gameplay ahead of you as lay behind. This isn't some cheapy "play the entire game again and get the pink weapon" trick like Ghosts 'n Goblins uses, though. All the stairs and chandeliers and everything else are now upside down, creating all-new puzzles even though the territory is familiar. The enemies are also all replaced by harder enemies. It's surprising and amazing that it works.

Back to Mario Galaxy. It has plenty of surprises of its own, but those take place within each of the many levels. If you take a zoomed out view of the game and just look at the structure of it, it's incredibly predictable. You very quickly realize that each world has 5 galaxies (levels). You realize how many worlds there are from the way the blank spots are arranged on the map.

Even though particular levels are surprising, the overall exercise on the most zoomed-out level becomes monotonous. I played probably the last third of the game on low volume while I watched reruns of Frasier and The Golden Girls on a second TV. (A less honest writer would not have admitted that!)

Mario Galaxy's purple coin missions were especially boring and tedious, even though some of them were very difficult. These missions have you return to familiar levels, but this time the levels have 100 purple coins in them that you must collect.

It's like a 10 cent version of the Castlevania's upside-down castle gold standard. I really wanted Mario Galaxy to break out of its own formula and surprise me on the macro level as much as it surprised me on the micro level. If a big, paradigm-shifting surprise belonged in any game, I think it's this one.

In some strange synchronicity, note that even this very article took a completely different direction than it started on. You expected it to be sheer glowing praise all the way through, then I started giving you may strange fantasies about what the game might have been. Nonetheless, as I mentioned earlier, Super Mario Galaxy iterates on its core mechanics in some of the most clever, polished, and beautiful ways possible.