The Art of Defeat

Lost in the reams of design documents and technical specifications is that we are trying to make games. Mixed with stories and graphics and sounds, sure, but at their heart we are still trying to negotiate some sort of molded puzzle. And one of the first things a designer tends to gloss over is what happens when the player fails. We want the player to have fun, and most of the time that means we want the player to succeed. But in game design, the one thing we fundamentally cannot count on is the way the player will react, and that means we have to plan for the worst.

We have to plan for the player to lose.

A lot of modern design seems to draw from the ancient coin-op origins of defeat. You lose, pop in some more money to continue. For the arcades, this made some sense. The shops want to make money, and the only way they can do that is by encouraging a steady flow of gamers to beat top-scores. The developers were banking on failure; winning was a loss proposition.

At the home, the theory changed. There’s no coin-slot, just a huge up-front payment. The player has an expectation – not discrete, but still there – that they will get a certain amount of enjoyment for those dollars. Those once useful game over screens amount to nothing more than a digital slap to the face for getting something wrong. Instead of encouraging replay, the gamer begins to question whether they should leave the game running, not to mention questioning whether they made the right purchase in the first place.

That’s not to say negative input is a bad thing. Like sides of a coin, we need to punish and applaud the player for the actions they take. But the old school of loss theory – the original art of defeat – is one step too far. Instead of punishing the player within the context of the game, it punishes the player for playing the game. We’re stripping them of their experience like parents would ground an unruly child, and it’s a practice that has to stop if we want to make experiences that don’t lose the audience after a few inevitable mistakes. Consider these game examples:


Irrational‘s rethinking of today’s FPS included a new take on the art of defeat, actually working the details into the environment of the story. The moment the player dies, he is reincarnated at the nearest vita-chamber, with all of their progress intact. If you try to apply some sort of story logic to it, the idea fails miserably, but as a gameplay element it is marvelous. And to the player who never dies, it’s simply invisible.

Beautiful Katamari

Here’s a fresh example of how the art of defeat was not taken into account. Fail during a game event, and not only does the event end, but the player is thrown onto a giant-sized pool table, with the game’s king and narrator shouting profanity while shooting pool-balls at the player as he scrambles around. The player is left to scramble for their life until the the retry screen appears. In execution the event is quite funny, but over time it begins to sound more and more like an angry diatribe aimed at the player himself.

Picross DS

Here’s a more cut-and-dry example: a simple puzzle game. Each puzzle asks the player to draw the picture according to the numerical clues. Drawing a pixel in the wrong place results in a progressively higher time penalty, and the player has only an hour to solve each one. But what if they go over that hour? No sweat. The game still allows the player to complete the puzzle, only their time record is not saved. The designers realized that there was more value to letting the player learn through completing their struggle than through penalizing their attempt at it.

So what is the modern art of defeat? Should we baby the player through each obstacle or relentlessly punish their mistakes?  I say neither, for the modern art of defeat is that of teaching.

Teach through repetition, as Picross suggests by allowing uninterrupted play. Teach through redemption, as Bioshock suggests by allowing the player to clean up his mistakes. Hell, teach through instruction – Uncharted: Drake’s Fortune (PS3) not only auto-reloads, but also offers a hint button that triggers after significant player confusion.

As developers, we shouldn’t be trying to stop the player from playing our games. We should be trying to teach them the line of thinking that we had when we designed our games in the first place. (Even a sandbox title like GTA relies on the player knowing the basic mechanics!)

Because if the player doesn’t understand how to play, how are they supposed to have any fun?


  1. Tyler Sigman

    I definitely like the trend of games being more friendly to the player–Bioshock being a great example with the VitaChambers but also the goal arrows and the hints it offers you when it thinks you’re running a bit astray. Using those hints is still my option, but it’s nice to know they’re there if I want to speed up my progress. Time is my most precious gaming asset these days.
    The old days of “bad player, now DIE” don’t have quite their charm for me anymore.

    However, I still think the designed failure was a great part of arcade games , though, or modern takes like Geometry Wars. Playing Gauntlet was only amazing fun when you knew you had X quarters to spend. Failure had consequence. Having free credits robs that style of game of its tension and the result is the same thing as removing conflict from a story–a boring experience!

  2. Richard Knight

    Oh, I definitely agree, which is why I mentioned how we can’t entirely remove negative feedback from the equation.

    The trouble is, games like Gauntlet and Geowars are what we’d call simple-math conflict equations. Shoot or die, progress or fail. That’s it. Apply a narrative or volume progression (eg: Katamari) to a game, and I don’t think the old art of defeat works.

    A good example lately? Super Mario Galaxy. Wonderful game, but I still just don’t get why there is a life count except as a nod to nostalgia.

    We still need conflict in today’s games, but at the same time, we need progression. And we can’t have both if we restart the program every time the player makes 3 mistakes.

  3. Tyler Sigman

    Good points, and maybe it just comes down to two main classes of game:

    1) I think that games where entertainment value is measured in linear, progressive hours (e.g. “a 40 hour experience!”) need to be progressive friendly. “Bad player you fail” is truly antiquated, since the point of all the content is to give an experience, not simply a challenge.

    Games where entertainment value is measured in repetition of smaller experiences (e.g. you might play Geometry Wars 240 times @ 10 minutes each and still get the same 40 hours out of it) can build failure into the design because that’s where the tension and excitement comes from.

    Thinking back to playing Super Mario on the NES, the game was still interesting even after you could beat it at will. You were just challenged to do so more quickly, or more efficiently, or with more brute force (no warping). The risk of dying was integral to the game because if you died too many times, you couldn’t finish during that session.

    I’m with you–trying to keep the player from finishing a progressive experience is just kinda dumb. Challenge is good; too much penalization isn’t.

  4. Richard Knight

    Yeah, I think we are on the same page.

    In Mario (or countless other NES games) progression was more of a measuring stick from the title screen than anything else. I’m…7 levels from the title screen, and so on.

    The trouble – and hence the theory of this article – is that a lot of supposedly story or content-driven games are trying to use the same twenty-plus year old milestone as a sort of afterthought.

    In a game with a cutscene, why should I have to watch it again? Why should I even have to skip it? Can’t the game-flow itself allow me to progress without that detail?

    A great semi-modern example would be from Metal Gear Solid 3: Snake Eater. At one point in the game, you can be caught in a Sniper duel. Get shot, and instead of reliving your trip to the duel, you wake up in a nearby prison cell. It’s up to you to promptly escape, head back to the duel, and finish the job (which still reflects any progress you made!)

    The entire game isn’t quite so smart about it, but the section strikes me as a great way to let the player progress from failure and learn a few things on the way.

  5. Radiant

    I agree with Richard on the point of MGS3. I was about to say that for the Katamari example, instead of punishing the player by removing him from the event, why not “level down” and reset the timer appropriately? Why not just let the player continue through learning?

    I like how Prince of Persia (xbox, ps2) handled punishing the player for mistakes. You had limited number of sands that could be replenished, and your mistake can be reversed a number of times. How is that different from simply “you lost 1 life, now go back here.”? Simple – the control is in the hands of the player. He can rewind to which ever part he feels necessary (to a point). The player is still given the sense of challenge but without the frustration. It has the elements of Bioshock’s Vitachamber in there, but with the limited amount of “retries” which can be replenished at the player’s abilities.

    Other ways this is dealt with is extra resources, or “health packs” that keep the game-overs in the player’s control.

    The failings of course is inevitably, you may run out of these and you see the game-over screen. I still prefer seeing that and dropped to a nearby checkpoint than the vitachamber though. Otherwise I might as well have God Mode turned on. I don’t like the idea of being babysat through.

    Another way of dealing with it is like flOw does it – dynamic difficulty adjustment. The player is penalized by being sent back to the previous level, with only one health, but always a chance to redeem and win with dynamic difficulty. I think the key thing is to punish the player, but yes not so much as to frustrate him, and flOw does an excellent job of this.

  6. Little

    What about the Psycho Mantis battle in Metal Gear Solid 1? The game basically forces the player to fail one or two times because it gives the final clue that is actually impossible to figure out unless you know or someone told you about it.
    Of course afterwards it’s a very cool idea, but I think a lot of people could be very frustrated and maybe stop playing after the first failure.

  7. Richard Knight

    The Psycho Mantis battle is an interesting one, as the game is based on a variety of gimmicks. Sure, it’s possible to lose, but I’m not quite so convinced that it’s easy to fail.

    Let’s look at the criteria:
    -Instructs the player? Sure. Frequent codec alerts happen during the battle, including one that reveals the controller swap. I think you can learn this the first time through if you bug the Colonel enough?

    -Redeems the player’s failure? Again, it seems to. Your restart point is typically right outside the hall, and you can immediately put any tricks you learned to good use.

    -Learning by repetition? This is less clear; combat itself is repetitive but there are some situations where you just have to take the hit. Still, the sequences are fairly predictable and you can learn how to avoid most of the problems.

  8. Sieg

    I’m not sure why I’m stepping out front in the open to leave a conversive sample anywhere as I’m reclusive out of disbelief in self expression in handful of situations, usually as it earns you rejection or lack of acceptance. But I’ll take the leap and say I disagree with this philosophy entirely, with all due respect and an understanding of the probable mentality breathing life into it.

    I for one am not an exceedingly talented person. You could say I led myself astray and never knew how to play many games at a high extent growing up all the way into my latter teen years, this being one of which, and I couldn’t hold my attention to a tv screen to play a game much. I’m confident had I layed hand on katamari, bioshock or anything like that if I was still of that personality I’d get bored after half an hour; I’d prefer to sit down and poison my brain with manga and get lovesick rather than be a gamer, even if I grew up loving them.

    But difficulty and player creativity in games changed all that. I’ll admit to my viewpoint being niche, but if you’ve made it this far to hear me out, I can add to it that evolution is a cornerstone of human development and adaption is a corner stone of evolution. Difficulty, when designed properly, makes you adapt and feel you’re on the ropes to constantly invent and create. When you gain the self awareness to realize losing really is just the flipside of the coin as in your example and not something that burnt all the playdough you were twidling about in your fingers, you can see it as what it is – it gives shape to the playdough and adds rules to abide by, and thus makes you think so you can avoid it.

    Games that primarily rely on the experience – yeah, they’re decent. Before I got into games properly (Playing everything that fit my philosophy from Street Fighter to Viewtiful Joe to Final Fantasy Tactics, I play all day sometimes and own hundreds of games) and cared more about stuff like travel and the net, I’d like a game like mgs or ff to drown myself in the story, and stuff like Portal (had it been out) would interest me as an idea probably.

    But what really makes you think and gives you a reason to pay attention to the shape of a game carefully? The concept of fair, but unweakened and sour bitter defeat. A game is little more than playdough to mess around with – before a game ties specific demands and makes you play about in certain ways for enlightenment and reward.

    Bioshock is cool for any gamer even though I won’t hide from you that I think its much better for ‘playdough’ gamers – but the vita chambers actually made it less alluring for smarter people and difficulty would’ve helped it. The concept drew people in and made it good, it wasn’t the Vita Chamber that breathed life into the game – it was the structure.

    When you can lose, you can win too, or else it’d be unfair. Thats why a game thats hard but fairly balanced rather than broken or messy is more prized than the latter type, even if both were equally hard. Victory and defeat go hand in hand, and this goes into the design of specific things in gameplay. “If they can do this much, you can do this much to stop them.” ~ It complicates the theory when scaled down like that, but retains the ideal.

    Its because you can win though that theres a point in trying to win. If you make an eventual but inevitable reward for every player of every skill level, the concept of fighting for something and thus the ‘reality’ of the playability loosens in importance, and thats what bored me about games and threw my attention span down the drain.

    Its fine to have fun gimmicks, Mantis is for example everyones favourite boss and will likely remain mine as so. And I’m not saying everyone has to view games this way, but alot of people who love gaming as a hobby – not an interest that grows into a culture they’re nearly uninvolved with like *almost* everyone online – use this to dictate the reality of a game.

    They might play anything – but things that make you think are pivotal.

    I hope this doesn’t alienate you too badly, lest because of the length – that crept up on me, I rarely actually type much at all – but also because of the assertiveness of the disagreement. All due respects to difference though, I hope.

Comments are closed.