Skip to content

You’re Dead, or I Wish You Were

February 5, 2008

Yesterday, I discussed Dan Gilbert’s talk about happiness. In it, he posits two things: a) that we are happier when we are bounded (a more limited set of choices) and b) when our choices are irreversible.

From a design perspective, this has multiple implications, but the one I am interested in now specifically addresses his latter point – we are happier when our choices are irreversible. In games nowadays, few choices are actually irreversible. Reload your save, and you’re good to go. There is far less investment in that choice, and therefore, far less caring. At the same time, this has also made games accessible to a much larger group of people that may not have otherwise considered playing. In effect, we made death a more casual experience.

Back in the late 1980’s, there was much discussion in designer circles that the player should not have to learn via death. This especially applied to (and may actually have arisen from) certain adventure games that presented the player with a “choice” – one choice was correct, the other was death. This is bad game design, obviously. In an ideal world, players should be able to spend their time thinking about which way forward is the best way forward, not eliminating possibilities via death/options/load save game. That’s not fun. If you’re going to kill me, at least give me an idea that it’s a possibility so that I can assess the risk.

This “no death” principle soon found its way to other games where players were effectively piles of hit points, and death became a temporary experience, and then, not really an experience at all. Consider this transition:

  • 1981 – Harsh: Your characters die, and you’re done. A new party can try to go down and get them, but by the time they are able to reach them safely, your new characters are high enough level that you’ll probably just stick with them. As a bonus, if you do get them, you get two chances to resurrect them. If you fail on the second try, the character is erased from the disk. As a bonus, there was also no saving in the dungeon.
  • 1990’s – Mild: Your characters die, but you can resurrect them. No one would think of erasing your characters from the disk. That’s just mean. In fact, resurrecting isn’t that hard anymore, either. Developers are starting to think, “What’s the point? They’ll just reload their save game.”So, they make resurrecting easier. You still have to walk back to where you died, though. I mean, there has to be a cost.
  • 1990’s + 1 – Check Points: “Okay, players are getting annoyed having to walk back through the level. So, we will add check points and autosaves. That way, we’ve got their back if they do something stupid. They can just reload the save and start at the last check point.”
  • 1995-ish – Don’t Bother, Part 1: “Why are we even making players reload a save game? That seems unnecessary. We know players are going to do dumb things. So, let’s just automatically save the game for them and automatically reload it for them if they do something stupid. This should remove all risk from the occasion.”
  • 2000’s – Don’t Bother, Part 2: “Why are we even reloading a save? Why don’t we just make death kind of abstract, a little tiny bother. Just pop ’em back in where they were, and off you go.”
  • 2008 – Immortality?

When players know that death of some sort is a real possibility – whether in avatar death or in terms of some investment in time lost – it creates a degree of intensity that is otherwise missing from many games. Truly, though, I don’t know that today’s average gamer would appreciate the old-school “you’re dead and you’re done” style.

What do you think?

40 Comments leave one →
  1. February 5, 2008 4:11 pm

    I think that it really depends on the type of game. In a heavily story driven game I would want lots of autosaving because I dont want to back-track 20 minutes just to replay the story I’ve already learnt about.

    But, I’ve been playing lots of 2D shooters lately, particularly of the bullet hell variety, and they play entirely differently because you have to learn the patterns that the game throws at you and there’s no way to learn other than try-die-try-die-repeat until you win or give up.

    Similarly I’m developing an online roguelike game at the moment and I experimented with including an autosave – Which I then removed because it took away all sense of danger from the game, and it became, well alot less fun.

  2. David permalink
    February 5, 2008 4:18 pm

    If death is to be intense, it makes sense to me that it should also be rare. Anything that occurs frequently loses its effect, similar to the psychological effect of Habituation. The reason death is such a powerful concept IRL is because it only happens to us once (barring near-death experiences where people are technically dead for brief periods of time). If we could die and resurrect at will, death would be no different than brushing our teeth.

    The best example of significant death in games I can think of is table-top RPGs. The game master usually fudges the numbers or otherwise saves players from death-causing situations unless it’s an extremely dramatic moment or the players have been behaving like complete idiots. When death finally happens, it’s extremely meaningful. CRPGs, where the game master is a cold, heartless computer that plays purely by the rules, results in much more frequent and thus much less meaningful death.

    I personally prefer games that explain away death as part of the meta-narrative, such as Prince of Persia and Assassin’s Creed. I also like the time-reversal mechanic of PoP, which allowed me to avoid frequent deaths, but that mechanic can only properly be used within the narrative of a limited subset of games.

    An alternate solution is to make death a part of moving the story forward. Planescape: Torment did this rather well, I think. What I would really like to see, though, is some sort of mini-game that triggers when the player dies. The player must complete the mini-game to return to the regular game. It’s important that the player not be able to fail this mini-game, though, so as not to impede the flow of the game. The “penalty” for death is having to complete this mini-game, with loading save games being disabled. One could even add small bonuses to the game proper for performing well in the mini-game, though this would have to be capped and/or properly balanced to avoid the player farming the death minigame.

  3. February 5, 2008 4:19 pm

    I’d really love to be able to punish the player more. In my games, the executive producers aren’t letting me punish the player AT ALL for doing anything wrong, other than reducing an abstract score at the end of each level. In other words, they just can’t mess up, no matter how hard they try.

    And yet, in the focus groups we’ve done, the players are ASKING to be punished for screwing up. They want some sort of reaction to their actions, whether they’re right or wrong. They want more than just positive feedback. They recognize that the game is pretty much playing itself at this point because the challenge is reduced so much. It’s like having God Mode on all the time!

    So is it really the right thing to go the immortality route? Is it what the players want?

  4. Jim Preston permalink
    February 5, 2008 4:20 pm

    Just a quick point about Dan’s talk. I don’t think he was saying “a) that we are happier when we are bounded (a more limited set of choices) and b) when our choices are irreversible.” Instead, I think he was pointing out that the prefrontal cortex is a wonderful evolutionary psychological adaptation because it acts as a coping mechanism. The brain has this sort of adaptive feedback loop that can shut itself down in certain regards. In the instance of the students picking prints, they no longer contemplate the option of not having the second print. In the case of the paraplegic, they no longer contemplate the option of walking again. In extreme cases, there are examples of the brain simply shutting down much of its secondary processes due shock. And some people think that clinical depression may be instance of this feedback loop where the brain is unable to stimulate itself back into a healthy balance.

    I take his real claim to be this: the brain has a coping mechanism that “dampens” the feedback loop of choice anxiety. I find that to be a pretty persuasive claim. I think it is a bit of leap to say “we are happier with irreversible choices.” Would it make more sense to say “we are able to cope better with our chances when they are essentially made for us and we do not have to deal with the choice anxiety”? If it were truly the case that one was “happier” with limited, irreversible choices, then women would seem to be the happiest in arranged marriage societies, but somehow I doubt that is really the case.

    What Dan’s talk teaches me is the amazing power of the brain to adapt to circumstances; I don’t think I’ve learned too much about the nature of happiness. In terms of application to game design, since I don’t think irreversible choices lead to “happiness” as much as they lead to “coping”, I don’t think having games waste my time with permadeath or something has ever increased my enjoyment as a gamer. So, yes, as games continue broaden their appeal, they will continue to soften the penalty for failure, which is just fine by me. There are other ways to challenge and reward the hardcore than with “you’re dead and you’re done.”

  5. Nat permalink
    February 5, 2008 4:21 pm

    I think Gilbert’s comments have more to do with good choices. If the experiment was run with pictures 5 and 6, I doubt the order of the top four change. 5 and 6 would maintain their ranks in irreversible test and swap in reversible ones. If the test was with pictures 2 and 5, 5 would most likely move up a spot or two but 2 would probably stay 2.

    Applying this towards reversible death, using Gilbert’s test as an analogy, death would probably be represented by picture 6. As such, I don’t think you can really affect the player’s happiness much when you are comparing it with a theorectical picture 5 and given the choice between 5 and 6, the player will select 5 anyways.

    Addressing the question directly, you need some sort of consequence for failure otherwise you dilute the lessons learned. Just from my personal experience you learn much better from failure then someone holding your hand and spoon feeding you information/experiences; provided you are given the chance to correct your prior mistakes. If the players aren’t failing at all, the game is probably too easy. They’ve already ‘groked’ all there is to know about your game. Which isn’t to say that it still can’t be fun. Just that it probably isn’t AS fun as something with just the right amount of challenge. Mass Effect was pretty harsh with “you’re dead and you’re done unless you’ve been religiously saving your game.” Auto-Save points spanning over an hour of gameplay? If you remove death, you need to replace it with another consequence to facilitate learning. Perhaps that theorectical picture 5.

  6. February 5, 2008 4:39 pm

    This was an issue which created the burning embers of a heated debate at a conference I recently attended. I believe that death needs to be present in the minds of gamers as they make a choice.

    Being a writer, I can’t stand it the moment when characters “return from the dead,” especially without consequence (Gandalf disappeared forever eventually, so that was ok). Once death becomes a minor nuisance, (or the player is so powerful not even a threat), the drama of the moment is sucked away. You worry less about the characters and therefore care less about them, because every situation, no matter how dangerous, you know they will make it out unscathed.

    While I do use it more often than I should, I believe that the player’s ability to save whenever they want should be eliminated. I know there are gamers who proceed turtle-like: Kill an enemy, save, reach spooky corridor, save, enter spooky corridor, nothing there, save, etc. etc. Using this method, characters will explore the spooky corridor, regardless of their current set of skills, and will do it without fear. This is because they only have something to gain: if they are strong enough they’ll survive the spooky corridor, but if they fail, now they know what’s going to happen, and can avoid it, as if players gain a new “skill” called foresight. At that point, why don’t you just give free strategy guides with each game?

    I think a good solution is the old resident evil method of having limited saves (which in this case players can save whenever they want).

  7. Patrick permalink
    February 5, 2008 5:03 pm

    I think this is an issue because of our traditional generic constraints. Let’s look at the wider medium’s possibility space. How many interactions involve a risk of death? A tiny subset.

    Also, forcing people to replay over and over again is a hold-over from the arcade business model.

  8. February 5, 2008 5:14 pm

    “Also, forcing people to replay over and over again is a hold-over from the arcade business model.”

    Or is it? It’s a traditional component of board games going way back. You’re done? You’re done. At the same time, and like Jim notes above, that’s a rather harsh way to please the hard core gamer. Maybe there is another way.

    One interesting point is that we do accept “death” in direct proportion to the amount of time invested before the death. In casual games, we have no problem with a regular restart. In fact, the idea of an 8 hr long game of Chuzzle seems, well, painful to me (though I’ll do it for an hour with no problem). It might also be tied to player attachment. You don’t get attached to … nothing in Chuzzle. However, you do get very invested in and attached to your character in MMOs.

  9. February 5, 2008 5:47 pm

    @Jim – Totally agreed on the adaptive power of the brain. The part of his talk I was referring to was right near the end.

    “When our ambition is bounded it leads us to work joyfully….”

    And in this, I am thinking that giving players too many choices amounts to just noise. They can select one of these possible paths forward, and off they go, not worrying if they’ve selected the right way.

    The other part was the study where people were happier when they realized there was no going back.

    Ironically, he also notes that when our fears are unbounded it leads us to act recklessly or cowardly, which could say something about the dead-and-done death thing, too. 🙂

  10. February 5, 2008 5:47 pm

    Killing the player isn’t fair unless it’s tied to meaningful choices and mistakes. Consider:

    Principle: Sometimes a player’s mistakes aren’t really mistakes. If a player has no good reason to believe a particular action might lead to failure, then failure should not be blamed on the player.

    Failure by surprise occurs when the player’s actions do not appear as if they should lead to failure, because the relationship between action and outcome is either obscure or counterintuitive, or because the element that makes a particular action lead to failure is presented too late for the player to react to it. King’s Quest and Dragon’s Lair are known for this kind of failure mechanism, where the wrong move can easily result in unexpected death or failure.

    A game should generally provide enough clues for players to anticipate danger, and should otherwise give players a fair chance to react to any surprises.

    The following questions concern the problem of meaningful choice:

    – Why is X a better choice than Y?

    – What if I make the wrong decision? Can I try again, or am I stuck with it?

    – What would have happened if I’d done X instead of Y? Would it have lead to a better outcome?

    – Why must I choose between X or Y, when Z is obviously the superior choice? The only reason I can’t choose Z is because the programmer didn’t bother to implement Z.

  11. February 5, 2008 5:53 pm

    I liked Diablo 2’s hardcore mode. If your character died, nice try but you are going to have to start over again. I really don’t see what is so wrong with it. At least it gave me the option to kill off my character. Sure I could play it on a easier mode where death was temporary but I personally like to lose sometimes. I never beat it in hardcore mode, I got close a few times but I always died and had to start again. I spent a whole lot of time without gaining the ultimate sense of accomplishment…. It makes me wish it wasn’t stolen from me.

    Losing is ok. I’ll play a game knowing ahead of time I’m going to lose and that is fine by me. I like losing because I can observe the winner’s methods of play. I can emulate what they have done to become victorious and I can now better understand why I have lost. But a lot of players don’t like to lose. They are spoiled with respawns, autosaves, unlimited lives/continues, or something else to be sure everyone stays alive. Everyone likes to win, all the time and if they don’t win it seems as if they get angry and frustrated :*( . Really I just think it is game designers spoiling the gamers. Kind of spoon feeding them the experience as if they were infants. It reminds me of my mom asking me to let my younger brother win because he wasn’t as good as I was. Of course I’d eventually let him win (when he started playing correctly or when he simply became too upset) but when you lose you are supposed to learn why you’ve lost not cry about it. I used to get really really upset when I lost at any game with the NES. I’d get so pissed and my father would simply tell me it is just a game and if I wanted to win I’d have to try figure out what I did wrong and then correct it(or to go outside and play).

    I love shooters like Socom II. In the standard online mode you start out the round alive. If you die, you are dead for the entire round. Your body sits there and you respawn only when there is a winning team is declared. Now there are the respawn rooms but they are unranked and don’t add to your stats. I would use respawn to warm up, or to play drunkenly but when I wanted to compete and feel any sense of victory it had to be a room where death mattered. Nothing feels better than sweeping another team and going 25+/ 0 (Kills / Deaths) in a high ranked room. (High ranks mean better players, for the most part) You just feel like you dominated but if it is a respawn room you can simply get kills be camping spawn points, or camping in general. Respawn games don’t really hold as much competitive value to me.

    But of course if every game were to suddenly kill off the forgiveness level, gamers might get frustrated and since they are babied so much they will probably turn off the game and thats not really what you want to happen. But I do predict that babying gamers will ultimately result in a desire to have games with more of a risk factor so over time the risk factor will arise. Sounds weird but when you start making death less serious you tend to feel less rewarded when you are victorious. “Yay, I win, lets play another game now.” has now become the victory roar. Why can’t it be “HOLY SHIT, I BEAT IT…. @#$^ YES!”. Or a overly excited “IN YOUR FACE! IN YOUR @#$^ING FACE.”

    I also think maybe allowing players to win so easily takes away the desire to invest in purchasing the game. Why buy the game when you can simply rent it and beat it over the weekend? Then you don’t have to buy it and can simply rent it.

    Maybe I am dead wrong and games will eventually just replace loss with meaningless warnings and there will not really be any negative consequences of death. Maybe losing isn’t as valuable as I think. It is just about keeping gamers spoiled…. However, I personally would like to be provided with the option to actually lose in a meaningful way. Otherwise I eventually won’t have any choice but to win and that isn’t fun to me. It would be like I’m in hell.

    As a novice game designer I don’t feel any remorse killing off players. When I hear a person say “But that would mean the player would have a harder time trying to survive.” I usually respond, “Good, if they deserve to die then let them die until they learn.” Since when isn’t it the job of the game designer to kill the players? I don’t mean that in a literal sense by the way.

  12. February 5, 2008 6:10 pm

    Adrian – I don’t understand how killing players is unfair. If they make a move that dooms them to failure, I see that as a mistake and they should suffer for it. I don’t really know how else to interpret it. In chess If I leave my king exposed then that is a mistake and I should lose because of it. If I leave my queen exposed in a meaningless way that is a mistake and I should lose my queen. It will be surprising when the next person makes their move but that is totally the players fault for making the move. Or perhaps it is the reward of the other play by forcing them to make that move.

    Unexpected failure is marvelous!(Besides the infamous FPS death immediately upon respawning flavor) You have to grant the player that ability to discover why they failed unexpectedly but there is no reason to eliminate failure by surprise. I see this as an opportunity for the player to use their brain and to figure out what they can do to avoid this from happening again. Some call it strategy, I call it fun.

  13. February 5, 2008 6:29 pm

    Killing players isn’t inherently unfair; It’s only unfair if players have no good reason to believe a particular choice is incorrect. Designers should give players the opportunity to make the right choices based solely on the information that’s available to them up to the moment the choice comes up.

  14. February 5, 2008 9:11 pm

    “What I would really like to see, though, is some sort of mini-game that triggers when the player dies. The player must complete the mini-game to return to the regular game. It’s important that the player not be able to fail this mini-game, though, so as not to impede the flow of the game.”

    There’s a game called Prey that did exactly that. I believe the mechanism was even the main topic when they wrote a post-mortem of the project for Game Developer.

    “I think this is an issue because of our traditional generic constraints. ”

    Actually, I suspect it’s an issue that comes from Technical constraints. We added save points because “save anywhere” required too much memory on disk. Eventually it became a genre convention.

    “One interesting point is that we do accept “death” in direct proportion to the amount of time invested before the death.”

    I think that’s a big part of it. I don’t mind “dying” in a 90-minute board game because I don’t have a lot of time invested, but losing an RPG character that I’ve put 40 hours into just plain hurts.

    Another part of it is expectation. I know people who are perfectly happy playing board games that takes longer to complete than most CRPGs (some of these were even mentioned in your earlier Boardgames thread on this blog). But you know from the beginning that at the end, one player will win and the others will lose.

    If, by contrast, you’re playing a game and suddenly it kills you and erases your save, you’re like… WTF? Unless this was something you expected at the time when you purchased the game.

    FWIW, the concept of Permadeath isn’t gone, it just hopped genres from CRPG to Roguelike. In Roguelike games (most notably Nethack), if you die you’re dead. In some versions of Angband, not only do you die, but a new unique monster that’s an undead version of your earlier character haunts the dungeon, making it HARDER to complete the game 🙂

    Even a few console games take this harsh road. As an example, Nightmare of Druaga has an absolutely draconian save system: you’re only allowed one save, you can’t save in the dungeon, and if you die in the dungeon you lose all your equipment (you retain your experience level, but the majority of your character’s power and abilities come from equipment, so for all intents and purposes this is like dying). If you turn the game off without saving (i.e. in order to cheat the penalty for death), not only does it treat it as a death in the dungeon, but you’re given a ten-minute-long lecture on cheating.

    Lastly, in some ways this comes down to player choice. I know some people who play games in “hardcore mode”, where they will delete their own save and start from scratch if they die (they do this manually; the game doesn’t support it). Brenda says that today’s games are like having God Mode on all the time… but does that mean a game with permadeath that offers God Mode is okay? And if so, then are we really just arguing over the default mode of play, whether you start in “normal” or “God” mode, since you can essentially enforce either one as a player?

  15. Jim Preston permalink
    February 5, 2008 10:29 pm

    @bbrathwaite – I think you and I are probably fundamentally in agreement, but I just want to underscore one point: I think Dan has pulled a bit of a logical sleight-of-hand. First, I don’t think anyone is going to dispute that there is such a thing as choice anxiety. Barry Schwartz does a nice job of analyzing it in his ’04 book “The Paradox of Choice.”

    The mistake I think that Dan Gilbert makes is that he points out that there is such a thing as choice anxiety, and that if you eliminate those choices then you also eliminate that anxiety. So far so good. But it is quite another thing to say that the absence of anxiety is the presence of happiness. I don’t think there is a clear correlation there. If there was, then the more we eliminated people’s options, then the happier they would necessarily become. But surely working in a dead-end job with no options for promotion and our “ambitions are bounded” doesn’t sound like it will lead to joyful work.

    Just as the absence of vice doesn’t indicate the presence of virtue, the absence of choice anxiety doesn’t indicate the presence of happiness. In terms of games, I think it is possible to overwhelm players with choices, but the solution isn’t always to strictly limit their choices, but it could be to order, rank, associate or prioritize those choices rather than limit them. Burnout Paradise doesn’t give me the choice to restart the event I just lost, and this hard bound on my ambition is making me throw my controller into the couch! 🙂

  16. Nick Prince permalink
    February 5, 2008 10:45 pm

    The Metal Gear Solid torture scene was on of my favorite gaming moments. It takes place after you’ve completed about 75% of the game. If you die in the torture your save is gone, no more lives, no continues, just game over. And you were fully warned before hand. On top of all that, submitting or surviving the torture affected the ending of the game.

    Death should be meaningful. The whole save, try, die, reload cycle ruins any suspense the game may have had. I think 90% of games benefit from the checkpoint system. Saves happen only after the player completed a challenge from beginning to end.

    Another possibility is checkpoints that increase in difficulty after death. The player has learned something by trying and dying, so make it harder.

  17. February 5, 2008 10:58 pm

    @Jim – great point. There’s definitely a leap there. Somewhere between too much choice and too little choice, there’s a happy medium, I suppose. Good call on his leap, though. The topic’s rather fascinating.

  18. February 6, 2008 12:05 am

    @jim-“If it were truly the case that one was “happier” with limited, irreversible choices, then women would seem to be the happiest in arranged marriage societies, but somehow I doubt that is really the case.”

    Surprisingly this is exactly the case.

    “Yelsma and Athappilly (1988) studied marriage satisfaction and communication practices of 28 Indian couples in arranged marriages, 25 Indian couples in “love” marriages (marriages of choice), and 31 American couples in companionate marriages. They found that persons in arranged marriages had higher marital satisfaction scores, as measured by the Dyadic Adjustment Scale (Spanier, 1976), than either the love-married persons in India or the companionate-married persons in the United States. Furthermore, their results indicated that husbands and wives in arranged marriages were more satisfied with their marital relationships than were the husbands and wives in the U.S. sample.”

    Choice is what makes interactive medium, well, interactive. Without choice, the player/viewer is only being presented with a linear narrative, not interacting with it. As simple a choice as “turn left” instead of “turn right” causes a split in this narrative. Some games, such as Prince of Persia, allow a bidirectional linear narrative; players may move forwards through the series of events, or move backwards. This is what autosave added to the interactive process.

    This, however, does NOT mimic real life. In real life, as we all know, time is linear; we are not allowed to go back and fix our mistakes. Why then should this be so in games? This “choice, die, reload, repeat” mentality goes against the prefrontal cortex; a marvel of evolution according to Dan Gilbert. Why should designers fight this?

    Designers should, however, implement some sort of system to inform the player of what effect their choices will have on the game, but the information does not to be regarding imminent events. I remember at the beginning of Black and White there is a boat. The player can choose to either help the people on the boat or destroy it. It has no immediate impact in the game. However, near the end of the game, the saved boat returns and gives a great boon to the player.

    There is a system already set in place in life that informs us that something we are going to do is bad: it’s called pain. A man much smarter than me once told me that “Pain is a great motivator. You want to do anything you can to stop it.” Perhaps designers should turn to it as an example of how to implement a feedback system. It is what we are used to as a means of communication information about an impending death.

  19. February 6, 2008 12:26 am

    “When players know that death of some sort is a real possibility – whether in avatar death or in terms of some investment in time lost – it creates a degree of intensity that is otherwise missing from many games.”


    What would Super Mario Brothers play like if baddies caused you no harm, and falling into the void lead to an instant respawn at the edge of the nearest platform? I expect it wouldn’t be any of fun.

    I like to look at this in terms of loss of progress. Click on my name for a long-winded post on the issue of progress and failure.

  20. February 6, 2008 3:16 am

    It seems to me that the problem with death in games is that it doesn’t change the game in any way. There’s no feedback loop. The player is simply sent back a certain amount of time and forced to repeat the exact same set of actions and pass the same set of obstacles. The only difference being the player now has a bit more information about the coming obstacles. Death should be a game-changing event, not a rewind button. The system that Ian described from certain Roguelikes where you actually leave an undead version of your character when you die is an interesting example because it does have a feedback loop (albeit a positive feedback loop that really only serves to make the player more likely to die next time).

    A lot of problems also stem from tying together saving the game in order to reload it next time you play, and saving the game in order to reload after player death. We take it for granted that saving and loading serves two purposes, but there has to be a way to separate those two systems. I just don’t know what it is.

  21. February 6, 2008 5:18 am

    The impact of choice is exactly why we feel different while playing different genres.

    For example, shooters usually have linear storyline and the only way you can play it is to go from point A to point B with less frustration as possible. Even if some choices are presented to us in FPS, like say should I hide behind that box or that car wrack, they loose significance because we can retry them and choose differently in case of failure (death). Story branching could increase the importance of choice in this genre while death is something we have to live with.

    On the other hand, some genres are not linear. Like strategies for example. They are chaotic in fact, and each and every choice we make has tremendous importance for a game outcome. How many times I felt happy while admiring my walls and precisely positioned defense towers in WarCraft just to be pwned by some kid who built his army before I did.

  22. Jim Preston permalink
    February 6, 2008 11:24 am

    @jeff – very interesting study, Jeff. I wasn’t aware of that. I wonder what the results of that study would say about me as a committed bachelor? 🙂

    A lot of interesting replies in this thread. As a designer, my thoughts now turn not to who is right or wrong, but how can I make a game that will appeal to both the challenged-starved and the challenge-averse.

  23. February 6, 2008 1:35 pm

    @Jim – “As a designer, my thoughts now turn not to who is right or wrong, but how can I make a game that will appeal to both the challenged-starved and the challenge-averse.”


  24. February 6, 2008 3:40 pm

    @David: “resurrection as a mini-game” has been already done, at least once. In Prey, when player is killed, he is sent to a spirit realm, where he must shoot certain amount of spirits flying around him. Then he is sent back to where he was killed. However, I do not think this was the best way to divert the problem of death-reset-try-again cycle, as it was pretty much the same thing; player is practically unstoppable.

    I think dead-and-done death mechanic are useful depending on the purpose of a game.

    The whole point (or 99%) of playing most of today’s RPGs (FF series, etc) lies in seeing the story till the end. Go through challenges to see what happens next. In this case, if having to lose save game file when player’s party gets wiped out, is more than extreme. It’s the game-stopper. This is as if books having challenges at the end of every chapter and having to read from the start if a reader fails. This is ridiculous, not even remotely close to fun nor meaningful. (Unless the book was a textbook?)

    If a game’s point of playing is to have experience until death, I think death-and-done mechanic may work quite nicely. TRPGs are the best example, as Dave mentioned, as the major reason of playing TRPG is to act as the character, experiencing the fantasy world guided by dungeon master, until the character’s death.

  25. February 6, 2008 4:05 pm

    Oh, one more: (should’ve thought of this while writing above comment)

    If the game’s challenge is to hone one’s skills and overcome difficulties to progress, penalty by death is probably not a good idea, as the game becomes unfair by not giving players chance to practice at getting better at the skills being tested. That maybe why we’re seeing less and less penalties with today’s games as the majority of today’s games are based on skills, such as FPS or RTS.

  26. February 6, 2008 7:18 pm

    If the punishment is too harsh, it could eventually take the fun away from the game. If it’s too light, the game would have to be so fun and amazingly designed that experiencing the world alone makes the player forget the lack of challenge. Think Half-Life 2: Episode 2. For my part, I want to play further if the reward makes the previous frustration worthwhile.

    Bioshock could prove as an example of the modern “immortality” in games. Here, the player gets ressurected at a nearby pod if he/she dies. However, I found myself actually using the load function instead, even after being ressurected. I felt a kind of shame by being treated so lightly by the game, and I just didn’t want to see my character die in the story. Instead, I wanted to play through “flawlessly” by reloading.

  27. February 6, 2008 11:17 pm

    Adrian: “What would Super Mario Brothers play like if baddies caused you no harm, and falling into the void lead to an instant respawn at the edge of the nearest platform? I expect it wouldn’t be any of fun.”

    There is such a game. It’s called Braid, created by Jonathan Blow. It’s actually quite a lot of fun, and manages to make the player ask themselves why they put up with having to start over at the beginning of levels all these years.

    And that’s actually the problem with death: it forces you to start the game over, covering the same levels and enemies and patterns that you’ve *already solved*. It’s an artificial lengthening of time that it takes to finish the game, by forcing the player to do something they already know how to do (and therefore, from flow theory, the player becomes bored).

    So, on the one hand, the threat of death is a great way to create tension. On the other hand, the consequence of death is a great way to bore the player to tears.

  28. February 7, 2008 10:25 am

    I’ll have to play Braid to see what you mean. The description says it’s “an action-puzzle game”, which makes it a different genre than Super Mario Brothers. I just don’t think Super Mario Brothers would work if the only change were to the death mechanics. If you change other variables, you’re likely to get results that involve those other variables in addition to the particular variables that concern my claim.

    I have no doubt it’s possible to make good games that don’t require loss of existing progress, whether through death or any other sort of mechanism. Progress must still have a cost, however. For puzzle games and point-and-click adventure games, that cost is to be stuck (or be told the answer) if you can’t figure out how to solve the puzzle.

  29. March 9, 2008 6:46 am

    This has been an issue that plagued me for the whole of my holiday! Where do you draw the line. In my current game (I like to design games even though I am not a games designer) you are on this island and you must find out what’s going on and survive. The problem I faced was the whole consequence issue you talk about here. If a player just “respawns” instantly, then they won’t care about decisions. Likewise, if once you die, that’s it, it will, most likely turn the player off. In this game, without the consequence, the survival element will be diminished and I know that I can only be having fun if I feel like I am right on the edge of living and dying.

    I thought about letting the player save when they like, but of course, then if a decisions is bit risky, they will just save before. One idea I did have was, like in Alpha Centauri, you can only save if you quit. This works well because it means that if a player has to go to dinner of whatever, they can without having to find a checkpoint that may or may not be round the corner or loose their progress. On top of which, most players won’t bother to do that just to ensure your safety as they will have to restart the game.

    Another idea I had to work in conjunction with aforementioned the Iron Man saving technique was you have say 5 save points. You place these anywhere on the map and they are near undetectable by enemies. You can use these places to dump your stuff and save your game without quitting. This means there is still a penalty for dying because even if you just saved, you will have to walk back there, and because you can’t leave these in the middle of the enemy base, lest they get detected, so people will be forced to think about their decisions. On top of which, by only allowing 5, you introduce a bonus layer of strategy.

    However, I do think this is the holy grail of survival games at least because some players want the challenge whereas others just want to get through the game.

  30. Malinda permalink
    March 10, 2008 11:04 am

    Perhaps the ultimate question is how to respond to bad player choices without a character death option IF death is treated as an inconvenience rather than a serious error.

    As a very casual game player, not a designer, I tend to do game-breaking things like explore rather than stick to the designed gameplay script or task list. I’m typically more interested in the design and “world” than slogging through a task list UNLESS the gameplay is compelling enough to divert me from my poking and prodding and wandering.

    My free time is so severely limited that character death tends to be a signal to me that I’ve spent too much time playing a game rather than working on, say, my thesis. 🙂

    I think it is a solution that has to be adapted to the game storyline, first of all. If the game is a survival game, then yes, there has to be a possibility that the character will NOT survive. Making it more transparent before choices are made that grabbing for Door C or the green mushroom will kill you if you aren’t ready to handle them might be a wiser move and lead to less player frustration.

    This also ties into your player personalities discussions. Different game personality styles react to challenges and death in different ways. I see it as an interruption of my exploring, other players might see it as a nudge to “play the level more perfectly” and still others might shrug it off and cheerily backtrack after a reload.

    Some systems that seem to work are those that track your health and stats with you. As in real life, you usually know when you are too tired or ill-prepared to take on an unknown challenge. Making it possible for characters to access their abilities before entering a new level or a new goal setting hurdle, rather than having a binary Y/N result after clicking on something (you click the right option, you live, you click the wrong one and oops, death) probably treats character death more appropriately as a serious consequence of impulsive or unexamined gameplay.

    Save points are useful, I’d say, but need to be player-determined. Again, the player knows if their character is prepared or not (or should). Trying to beat a clock or soldier on in search of the next (hidden) save point can encourage sloppy gameplay, all because the character wants to avoid death and backtracking through a level.

    The game I’m thinking of escapes me at the moment, but one solution re: save points and death I recall is limited teleportation. If your character had enough energy or stamina to get from point A to point B, and s/he had BEEN to both points before, s/he could teleport. If conditions were not met, s/he could not. This system prevented the boredom of slogging through levels or rooms you’ve seen a dozen times before. If you could supposedly walk there, you’d be able to do a teleport, and bypass the time and deja vu of moving the character screen by screen.

    But I digress.

    I think the primary challenge when handling character death is making it less of a binary situation. Death, in order to maintain some gravitas, has to be a serious repercussion for repeated poor decision-making, not “Surprise! The Grim Reaper just mugged you, game over!”

    Rather than merely checking “alive / dead, y/n?” more satisfying gameplay probably checks more variables, making mistakes cost in different and less (supposedly) permanent ways.

  31. Daniel James permalink
    March 11, 2008 8:28 am

    I think the idea of death as a penalty differs greatly from genre to genre. Looking at two extremes, I’d never re-start a Final Fantasy game if death forced me to begin again. And on the other side, how boring would tetris be if every failure could be reversed?

    You basically have to break games up into two distinct groups, narrative games versus challenge games.

    Narrative games are games such as Half-Life, Final Fantasy. There is a built in narrative with progress from event A to event B, with a final conclusion. The variety you see in gameplay is often driven by the progression. If I have to start over the Warehouse level, it’s the same every time.

    Challenge games are games such as Starcraft, Counterstrike and Tetris. In these games you expect to play a session from start to end. This session could likely end in death. The variety seen in these games tend to be action-driven. Two games of Civilization with the same map settings may look vastly different.

    The difference is in how you measure your progression. In a narrative game I don’t want to repeat progress because I’m literally just doing the exact smae things over and over again. In a challenge game, starting over might mean the adaptation of a different strategy. Or repeating the same choice as the last game may not result in the same level of success.

    I don’t think that you can ever take a game like Prince of Persia and add perma-death. It defeats the purpose of that game, which is to tell a story. Painful death only works within some contexts.

    Also, if I might take some issues with the “Timeline”, I’d argue that as the game medium has matured, the number of options surrounding death have become more numerous. To suggest (as you do) that all games treat death now as a minor annoyance and just pop you back into action is a Straw Man arguement.

    Games today offer a wide variety of death penalties. Some (like Halo co-op) are very generous, only setting you back when the whole party goes down. Some, like WoW just eat time with corpse runs. Some, like Vangaard, make death a serious problem. And there are still games that offer “hardcore” playmodes.

    The answer, as with most problems, is not black or white but gray. As the methods of experiencing a game expand, so does the needed mechanisms to drive it forward. There is no “one answer”.

    Ironically, two of my favorite games are perfect examples of this, Dwarf Fortress and Planscape:Torment.

    In Dwarf Fortress you build a fortress against ever increasing dangers. There is only one save, and that’s when you quit. The fun of the game is to recover from disaster, not to prevent it. Eventually one too many disasters hit and you are left with the feeling of “if I had done x since the beginning…”, and then you start a new game. Losing is fun.

    In Planscape:Torment you play an RPG about a man who cannot die. There are essentially no penalties in the game at all. There is no “lose” condition. But the game works, because the narrative and settings are strong enough to carry it.

    Two different takes on death/penalizing players in two different systems, both which work brilliantly within their own contexts.

  32. March 11, 2008 1:00 pm

    @Daniel James
    Very, very true. I think you may have hit the nail on the head, as they say. I do still think that some stories can only be driven by fear of death, but your examples of Tetris and PoP, very clever.

    Another thing I thought of was, in Life (bit abstract I know!) most people don’t die until they reach old age (bit of a generalisation). However, people don’t do ridiculous things because, as far as we know, death is the end. So what if, in a game that needed to instil the fear of death, early on, had you die and forced you to restart the whole game. It would happen so early on it would just be like going back to a check point, but the user would not know this and think it was their fault they died and if they did it again, they would have to restart the whole game again. This would means that, as they progress through the game, they would not take risks. Then I thought, but what if they did die, then they wouldn’t pick the game up again. One option could be do have it so a “random” number was chosen, however, the number always led them to survive. The user thought he was very lucky to be allowed to survive, meaning he would still be very careful with decisions but without having to restart the game.

    But, like I said, I think Daniel has got the problem sussed!

  33. Malinda permalink
    March 11, 2008 11:45 pm

    @Daniel: Well said.

    You got me thinking more about what types of games I typically gravitate towards, which is a little off-track, but hopefully I’ll reveal the relevance before I lose my train of thought!

    Had you merely asked if I prefer narrative or challenge games, I would have responded immediately that I prefer narrative games, even those games where you create your own narrative if you choose (like Sims games). (Digression: PC Gamer UK just ran an article on Sims 3, and they are apparently adding in a “buff” feature…interesting. Some screenshots of the article here:

    Then you mention Tetris, and I still enjoy a good session of Tetris now and then.

    I also prefer cooperative games to competitive ones, but, oddly enough, most cooperative gameplay takes place in competitive multi-user games where you have to consider the needs and motivations of your team or group (or a computer-assigned posse that follows your character about). Cooperative games are few and far between, though some online art games and word games can be cooperative or competitive depending on the group involved.

    Other games add a veneer of group cooperation but you still have to succeed or fail on your own merits, such as Kingdom of Loathing or Forumwarz (you can choose to join clans, you can share items and buff other players, but your progress or lack thereof is mostly up to you alone; either you are adventuring or you are exchanging forum posts and messages back and forth, but you can’t do both at once).

    Seems like Permadeath is acceptable in challenge games that tend to involve very little character empathy (Tetris doesn’t even have a character, you are an unseen force guiding falling shapes), but is potentially enthusiasm-killing in a narrative game.

    KoL has a series of RPG-like quests and quest areas plotted out, but how players approach those, and the player’s stats, and whether or not the player has useful healing items on hand, all determine likelihood of success. “Death” is replaced by a state of being “beaten up” which requires the player to use adventure units resting, to use or buy healing items, or to try another challenge that is more level-appropriate.

    I think survival-type games combine narrative (you are in a bad situation and must think or fight or hide to survive) and challenge aspects (typically, a horde of enemies / challenger appears! …much worse than getting eaten by a grue for persisting in adventuring off-map or without a light).

    In multi-player games with clans and groups, you have character empathy and added pressure to excel and survive lest you let down or abandon your game team members.

    Balancing character empathy-building and challenges and level-ups seems to be key.


    1. Death should be a serious consequence, and yet avoidable if the player has been sufficiently perceptive to in-game information. It shouldn’t come as a consequence of opening an innocuous door and force you to retrace your steps, or boredom ensues.

    2. In games with a real time element, time outs may be just as effective a learning tool as deaths. If you have to sit out for a certain length of time and merely observe, you learn without having to backtrack, and you are still a part of the narrative. (Online game time-outs also serve to level the playing field a bit because no one can just grind for days on end or use bots to farm without limits of time or turns: once you use up your daily turns, you have to wait for roll-over or reset to play again, and parceling out your turns and heals wisely becomes a strategy.)

    3. If save points are used, employing them as part of strategy and possibly being able to earn additional ones, rather than relying on static or hidden game-set save points, may save some frustration if death forces you to fall back to your last save point. Save points should not be able to be set within a challenge track, perhaps, though it seemed to work in Psychonauts. You could, IIRC, save at certain points within a challenge area, but not when actively battling a group of foes. So save points when the character can be legitimately resting or exploring, but not in the middle of a fight when the player senses s/he might lose if s/he is hit again.

    4. Some narrative games I’ve played are mostly dialog choice-driven, meaning there are three or so responses you can choose to give in response to a NPC’s comments, some of which are right and some of which are wrong. The amount of coding required to tie up all loose ends in a game like this makes it simple to exhaust all novelty in a few play sessions. Players who like “perfect” games will be less inclined to explore “wrong” avenues; exploring players will be frustrated if there is no way to explore “wrong” branches and still learn something new or interesting, or if they must restart or replay previous areas.

    5. Open-ended games like Sims lack challenges. Once the player learns how to master keeping a Sim from peeing itself or dying of starvation, which is an easy task even if Sims are moronic, s/he has to create his or her own narrative or risk losing interest. What saves the Sims is character empathy. Players can design and play God with pixel people in a pixel ‘hood, and creating personal narratives is what keeps the game fresh once the so-called challenges and game play alterations introduced by new expansion packs have been exhausted.

    6. In a story game, the downer of death or other demotivators has to be matched with strong narrative. If the inconvenience of death is greater than curiosity about What Happens Next, the player will quit in disgust.

    7. So we have several factors to balance: empathy for PC, story, appropriate motivators and demotivators, and challenges that make sense and are not plopped into a game simply to make the game take longer to finish.

    8. I’d say uniqueness and mood are also useful hooks to keep players engaged. Post Mortem and Silent Hill both mastered mood-setting, primarily with sound and art that was a bit murky (or occluded with fog, or lit by flickering light sources, etc.). Uniqueness is a more difficult quality to master–you can’t step too far away from what works, or from standard or intuitive player actions and movement controls, and still find a market (or producer) for a concept that has not been tested or based on a known successful formula. With large companies swallowing up smaller ones that might be more willing to take risks, uniqueness may be even harder to come by in the future.

    Again, note that my approach is as a very casual gamer at best, but I am always interested in how things work, and why people like what they like, and what factors come into play when entertainment is created and marketed. Sims 3 may be incorporating buffs and demotivators that seem more standard to different types of games than “god games,” and some of the descriptions of how Sims 3 character personality types will be assigned sounds suspiciously like personality typing a la Myers-Briggs. Unless I misunderstood, it may be a case of players selecting traits from a list of adjectives, and some personality typing and career selection testing formats do the same thing.

    Death as a demotivator also ties in to whether a game is a subscribe-to-play type, free, or a CD or other media you buy (or acquire) to play. If death costs players real money or time, there is likely to be added dissatisfaction with arbitrary or surprise death traps.

  34. March 30, 2008 10:35 pm

    We often talk about ‘Death’ as if it means something. Remember that death in a game is merely a narrative device, signifying a penalty of some kind. ‘Death’ is the narrative component, the penalty is the gameplay component.

    This mapping from narrative concept to gameplay feature can be whatever we like: from a tiny penalty (you’ve wasted 3 seconds of gameplay, go back and do it again), to the largest we can imagine (no, sorry you can never play this game ever again [which has been done in at least one cross media game]).

    We get to choose the level of penalties in our game, the kinds of failures that trigger them, and we get to choose what we call them in narrative terms.

    As for the penalties: just like any context you want to think of (tax, education, relationships), you need to find the right level so that penalties act as motivators not demotivators.

    On the narrative side, wrapping up your biggest penalty and calling it ‘death’, imho, is a hackneyed narrative device. A large number of videogames don’t need death to function narratively. Particularly since, as you said Brenda, we’ve largely castrated it of any pathos or narrative force by associating it with minor incoveniences in gameplay.

    It would be an obvious cliche nowadays to implement ‘three lives’ outside of a retro-shooter, but somehow we’re still addicted to character ‘death’…

  35. sgtphoenix permalink
    May 16, 2008 11:07 pm

    It’d be nice if death was a little more of a threat to players these days like it used to be, but like you said, a lot of games make death just a tiny inconvenience. It doesn’t feel as important or dreaded an event as it used to. It almost feels like cheating if the game does half the work for you and autosaves everything for you. There should be consequences for making bad decisions and while death should be the ultimate one, it isn’t, anymore.


  1. The Good, The Bad … and the Ugly » Blog Archive » : Hardcore
  2. Forgotten Lore » Blog Archive » Playing For Keeps - Tomb of Horrors
  3. Still Dead, And Still Playing « Applied Game Design
  4. Randy Smith’s Article on Consequence & Choice « Applied Game Design
  5. Jeff On Games » Blog Archive » Your chocies have consequences

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: