2006-03-06 : Unpopular Idea #2: Punish the Loser

These aren't dangerous ideas - in fact they're quite safe - but they're good ideas and maybe they've fallen out of fashion.

2. Reward the Winner, Punish the Loser
GM: He hits you! Lose [rattle rattle] 1.6 wound levels! What's your new impairment penalty? Also, he takes the initiative and gets an extra d4 for his next attack!

Yep, that's what I think you should have in your game. Color it suitably, of course - "1.6 wound levels" is terrible - but that's what you should have. If I win a roll, it should be to my mechanical advantage, and to your mechanical disadvantage, in the immediate future.

Why? Two reasons.

1. Because then I care who wins and who loses. Something tangible is riding on every moment of resolution: every moment of resolution is consequential.

2. Because that's how the world works, even in fiction. When people or characters go on to achieve victory after being set back, their victories include overcoming their setbacks. We can talk about counterexamples if you want, if you've got some.

A game dynamics reason and a fiction reason. And what you're designing is a game for creating fiction - could you ask for a better dovetail?

Objection 1: What about death spirals? They suck.

Answer 1: So don't design a death spiral, duh. That's easy enough.

Objection 2: Yeah but, how do I make sure that the bonus and penalty a) are significant, but b) aren't decisive? I want there to be suspense even when we've determined the bonus and penalty.

Answer 2: That's the same question again. So design your game so that the bonuses and penalties don't swamp the rest of resolution. It's not that hard.

Objection 3: Well this isn't the same question again: in lots of fiction, a character does suffer setbacks and then does go on to triumph. Won't advantaging the winner and disadvantaging the loser scupper that?

Answer 3: That IS the same question again: no, it won't scupper it, as long as you don't design a frickin' death spiral.

Just exactly how important is it to you, by the way, that every single time anyone plays your game, they create fiction where the characters suffer setbacks and then go on to triumph? That's not, after all, the only shape fiction has. There's also fiction where the characters suffer setbacks and then struggle to triumph but never do, for instance. Also fiction where the characters suffer defeat when it matters despite winning every fight up til then.

We're talking about alive fiction, fiction where you don't know what's going to happen. Can you stand the thought of creating a piece of fiction, just by luck and play, where your characters lose after all? I can. You know I can.

For example: Dogs in the Vineyard
To see the rewards and punishments for winning and losing in Dogs, you need to not get distracted by the game's reward system, that being fallout and escalation. Instead you need to look right at the atom of resolution: the single raise-and-see exchange of dice.

When I win, a) I spend my dice efficiently, maximizing their usefulness over the course of the whole conflict, plus b) you have to spend your dice inefficiently, wasting some of their usefulness. When you take a blow, it doesn't just mean fallout - it also means that you have fewer dice on the table, less ability to stay in the conflict. When you reverse a blow, the opposite: you preserve your dice on the table and I throw my dice away.

(If you want to work upward and look at how fallout works, rewarding you when you take the blow, notice this: I make a big raise, you look at your dice. What you aren't thinking: "I don't want fallout, so how can I avoid taking this blow?" What you are thinking: "I want fallout, but I also want the stakes, so how can I take this blow and still win the conflict?")

Bonus unpopular idea #2a: Hit Points Work Fine
Hit points don't do what you think they do (unless you're Ben Lehman*, to whom all this is old news). That's why you think they're broken.

Hit points tell you how long a conflict can last. They're a pacing mechanism. They're a perfectly good one, right there alongside Trollbabe's "a conflict lasts 3 exchanges," Dogs' "a conflict lasts until somebody runs out of dice," and Primetime Adventures' "a conflict lasts 1 roll." You need some way to know when the conflict ends, that's all.

* For purposes of this post, you are Ben Lehman if this is old news. I expect that this definition means that several of you are Ben Lehman.

1. On 2006-03-06, Tris said:

Given my "contribution" recently, I feel almost disappointed that I agree entirely.


2. On 2006-03-06, luke said:

Doh! I was hoping I was the only one who noticed 2a. My secret planz are revealed to the verld!



direct link

This makes...
BL go "Welcome to the club"
Chris go "Whenever I go out..."*

*click in for more

3. On 2006-03-06, Tom said:

"Also fiction where the characters suffer defeat when it matters despite winning every fight up til then."

In point of fact, a great deal of game-inspiring fiction works on that line:

The bad guys win every battle and then some farmboy in a star fighter drops a proton torpedo down your exhaust pipe and *boom* the war is lost.


4. On 2006-03-06, Ninja Monkey J said:

2a has neat repercussions. I'm a-thinkin' about it.


direct link

This makes...
SDL go "Beats..."*

*click in for more

5. On 2006-03-06, Troy_Costisick said:


This is great stuff Vincent.  I'll be applying it to my future designs.




6. On 2006-03-06, Roger said:

Why are most games so much better at punishing the losers than they are at rewarding the winners?


direct link

This makes...
BA go "Because games aren't supposed to be fun!"*
Chris go "Injury vs. Reward"*
luke go "hee!"*
NinJ go "Those are the only possible things he can dream about..."*

*click in for more

7. On 2006-03-06, Vincent said:

Roger: for example?

Here's how I figure it: rewards and punishments are relative to some baseline created by the resolution rules. Is there any in principle difference between giving the loser -2, giving the loser -1 and the winner +1, and giving the winner +2? Not really, especially not when you presume that the loser and the winner are going to keep fighting just one another.

Is there any difference in feel? Of course there is.

But would I go on to say that the best-designed games favor rewarding the winner, and that favoring punishing the loser is poor design? Nope. Instead I'd say that most games are poorly designed, so if most games favor punishing the loser, that doesn't teach us anything worth learning. If most well-designed games favor punishing the loser, now that would be worth noticing.


8. On 2006-03-06, Matt Wilson said:

How to avoid the death spiral, just off da top o' me head: 1) have the punishments and rewards apply only to the conflict, not to the rest of the game; 2) apply bonuses and penalties that limit options instead of ones that hamper all options, like, "ha!, you're hit, so now you lose your special sword ability"; 2a) apply penalties with a time limit, like "you can't use that relationship in the next scene"; 3) have a buyoff option, like "you're at a -1 for the rest of this game session unless you spend 3 points."

Otherwise, re: the rest of it, yeah, that's what I'm a doin'.


direct link

This makes...
MDS go "Upsets"*
PB go "Sorcerer"*

*click in for more

9. On 2006-03-06, Jeph said:

Here's a 3a) for you, Matt:

Apply penalties that characters can ignore or suppress for a while. Examples: gritting your teeth and steeling yourself against pain in BW, rolling Will to use Stamina dice in Sorcerer.


10. On 2006-03-06, Vincent said:

I'm personally an ever-increasing fan of Sorcerer's diminishing carryover bonus: a success in this round becomes a bonus die next round. There's no death spiral because every round you lose some large fraction of last round's bonus - but not necessarily all of it.


11. On 2006-03-06, Roger said:

For example... hmmm.  There's a few games which reward the winner in the short term—TSoY, Sorcerer.  But I can't think of many.

It seems like in most games, the character is at optimal strength when he first rolls out of bed.  If he enters a conflict, the most he can hope for is to break even.  He never gets better able to deal with the next conflict (except with long-term gains like level increase.)  In anything short of a complete victory, he's worse off.

Maybe it's a pacing thing, like hp.

But speaking to the point "punishing the losers is equivalent to rewarding the winners", I think there can be a significant mechanical difference.

It becomes evident when you consider characters who don't do anything.  If the people who try and fail get punished, then the people who don't try are, in some sense, rewarded.  But if the people who try and succeed get rewarded, then the people who don't try are punished.

Maybe it's the games I've been in, or maybe the players I've gamed with, but in many cases, maybe even the majority, the transcripts could be exhibits in the textbook definition of "social loafing."  Punishments are specific, but rewards are shared, so coast along with the group.


direct link

This makes...
JBR go "Also, the Gauntlet"*
NinJ go "Good point, Josh..."*

*click in for more

12. On 2006-03-06, Ninja Monkey J said:

Uh, before everybody's all, "Oh, Vincent, I love your ass, pleas, please, let me kiss it!!!"* I want to bring up a counterpoint:

In Shock: the loser gains a Feature (or a Credit, when the Antag loses) while the winner accomplishes hir Intent. You have to choose between the fictional gains and mechanical ones. That is, adversity builds character in quantifiable ways.

This happens because the Protagonist is supposed to eventually be able to confront the Antagonist in a finale, where secrets are revealed and final fights are fought. The Protag has to have a decent shot at getting what you want for hir. Your stats are never decisive, though; bad luck can put you down even if you've got great numbers; the key is that you, the player, get to say what you want either way, by having the Protag get more and more refined as the story continues.

Even more appropriate than this is a game fully about "adversity building character". When I play a game about heroes, I want my scars to be what terrifies my enemies, I want my wounded soul to be what shows all I've done for my beliefs, and when the guy goes down, I want the story about the dude's many exploits, not his cowering in a corner because he got a whuppin' one time too many and couldn't get the dice to throw down against the dragon.

Now, you say, "Well, don't build it with a death spiral!" But that's what positive feedback loops do. Either that, or the resources you gain don't help you win future conflicts, which... I dunno, that's not really a resource, is it?

Oh, I suppose those "winnings" could go toward something else, like scene framing responsibilities, or some sort of tournament score, where they're tallied for who gets to tell the ending, or whatever, but you're just feeding into a larger-scale conflict and reward cycle; the guy who starts to win first is most likely to win later.

* This should be said in a falsetto voice while rolling your eyes up into your head and flapping your hands in front of your shoulders like they're dead fish.


direct link

This makes...
MSW go "We love Vincent.!"*
NinJ go "Oh, yeah? Well I think you're a Terrorist provocateur traitor communist!"
MSW go "ha!"*
NinJ go "They're not Commiewiccan!..."*

*click in for more

13. On 2006-03-07, Gregor Hutton said:

I feel I should put something in here that goes *me too* ... so, uh, people management courses tell you a surefire way to demotivate someone is to punish their success or reward someone's failure—you instantly encourage failure and not giving a rat's ass about succeeding.

My management trainer was real big on rewarding success, for the simple reason that it makes you want to succeed, it's a motivator. And that's something deeply wired into our brains. I think.


14. On 2006-03-07, SDL said:

To echo Joshua here:

It seems to me that as long as you have real character advancement - not just equivalent character change, but actual losses or gains in ability -  you are going to have a "death spiral", one way or the other.

I was suprised to see that tSOY and BW had explicit mechanics for removing powerful characters from play - but now i'm guessing that this is necessary to address the "runaway spiral" issue.

Not that these spirals cannot be good - they give you an Endgame!

(Or are we just talking about the exchanges within a conflict cycle?)


direct link

This makes...
SDL go "aka 'power curve'"
NinJ go "It's not the character change, it's the direction."*
SDL go "Counter-direction of the rewards, right?"*
NinJ go "Here's what it's for in Shock:"*
SDL go "but Shock ends there, right?"*
SDL go "What i mean is..."*

*click in for more

15. On 2006-03-07, Ben Lehman said:

Here's the best reason why you should punish the loser, reward the winner:  Capes already did rewarding the loser so well that you'd be hard pressed to compete with it.



16. On 2006-03-07, Walt said:

From a basic game design point of view, the reason a death spiral is bad isn't that it leads to death. It's bad because the outcome becomes predictable. A "success spiral" is equally bad. Any time lag between when an outcome has become predictable and when it is decided upon is bad, because it's an interval in which play no longer sufficiently contributes to the outcome.

Another thing that's bad if if the outcome remains too unpredictable for too long. If it does so, it means that early successes aren't giving enough advantage to the side that achieved them—in other words, having the better results so far doesn't put you sufficiently "ahead" ( = more likely to win in the end). Which means the early rounds haven't sufficiently contributed to the outcome.

So, the Scylla and Charybdis of "reward the winner punish the loser" and "avoid death spirals" are simply the balance between predictability and unpredictability, or as I often phrase it, stability and instability. Stability establishes that the leader is enough more likely to win to make getting into the lead worthwhile. Instability establishes that the underdog has enough chance to come back and win to make continued play worthwhile.

I've often heard that "game balance" is a useless concept because there are so many different and conflicting definitions of it. But to me, the balance between predictability (aka stability, positive feedback, reward-winner-punish-loser) and unpredictability (aka instability, no-death-spirals, chance-of-an-upset) is the All-Ecompassing One True Definition of game balance. Most other definitions I've seen of game balance are special cases or instances of this.

This balance can apply to role playing games on many scales of play from small (e.g. within a single multi-round resolution) to large (e.g. across the character advancement or reward cycle, or on entire character arcs). The most typical instability components are longshot chances of results extreme enough to generate reversals. But there are many other methods.


direct link

This makes...
NinJ go "Hey, that's interesting, Walt."*
BL go "I had a game design once"*
FSF go "I call the "death spiral" situation "It takes too long to lose""*
WF go "I agree."*
SF go "Positive feedback = instability, not stability"*

*click in for more

17. On 2006-03-07, Walt said:

One more thing:

You have to look at the whole package of rewards and punishments. What looks like a reward for the loser (such as, points credited toward earning a future reroll) might only offset rather than outweigh other inherent punishments for losing, and therefore can be a necessary or appropriate instability mechanism.


18. On 2006-03-07, Calithena said:

I am Ben Lehman, and have been for many years. Who knew?

Hit points also don't have a death spiral effect.

They also can be a resource to use across several conflicts, especially if healing is tight.

Maybe if one emphasized all three of those things and called them "prana" in a design said designer would get to be a Cool Kid.

I am curious though: you can have _long term consequences_ that _don't_ effect the local conflict as real, non-death-spiral effects. And you can have a resource like hit points. But what are some good examples of meaningful short term conflicts that don't necessarily induce a death spiral?

If you have multiple avenues of conflict (like, maybe, I don't know, that magic dueling game Schechter's been working on forever, where there are five elements of magic) then I suppose you can hit some with penalties, which isn't exactly a death spiral, just forces you to be cannier about your choices of tactics.

What else?


direct link

This makes...
C go "i meant "short term effects""
BL go "Wounds..."*

*click in for more

19. On 2006-03-07, Neel said:

Hi Walt, your comment that "Any time lag between when an outcome has become predictable and when it is decided upon is bad, because it's an interval in which play no longer sufficiently contributes to the outcome." doesn't ring true for me.

If the pace of decision in your game is too fast, then it won't be fun. By "pace of decision", I mean how many interesting choices the players get to make before the conflict is resolved. Critically, note that you can have interesting decisions even when the final outcome of the conflict is a completely foregone conclusion.

Take DitV as an example—pretty much any conflict with 3 Dogs and 1 NPC will be decided in the Dogs' favor, simply because the Dogs have so many more dice. However, those conflicts are still interesting, even though they are lopsided, because the NPC still has the potential to take meaningful action before the whole conflict is resolved.

For a negative example, look at Exalted's social rules. There, basically, you roll a whole bunch of dice, and then you succeed. This is not much fun, and the reason it's not much fun is that there's no place for anyone involved to make any interesting choices on the way to the conclusion.

Feng Shui also does a really good job at this, and is the system where I learned this principle. There, you can have an AV 13 character fight an AV 18 character, and everyone going in knows beyond a shadow of a doubt that the second character will win. However, the player of the first character will always have the opportunity to make a choice about whether they try to run, or try to accomplish something before they die, or try to stall long enough for allies to arrive, or beg for mercy, or something.

My own criterion for a well-balanced game is that wide differentials of power lead to conclusive victory for the more powerful side, BUT that such differentials shouldn't strongly change the pace of decision of play.


20. On 2006-03-07, Charles S said:


It seems to me that what you are describing (good systems where you know that one side will win, but you don't know how, and you don't know if the other side will get in some good licks first) is still the full outcome not being predictable. While the general shape of the outcome is known, meaningful parts of the outcome are not. Once all of the meaningful parts of the outcome are predictable, then a lag to the outcome happening is a bad thing (and thus fr'instance, the freedom to give in Dogs is a good thing). I think that, weirdly, simplistic mechanics for complex situations (like Exalted social rules) are generally unsatisfying for specifically this reason. Since they allow no room for complexity and detail of outcome, the outcome is predictable before you even reach for the dice (and simplistic mechanics tend to squash elaboration and detail, since you know going in that your elaboration and detail will be functional ignored). Boring.

But I agree that your elaboration of the point is an important one.


21. On 2006-03-07, Mendel Schmiedekamp said:

In defense of the Death Spiral.

Much like hit points, it doesn't do what you think it does. A self-building process that leads to an ultimate transformation. It sounds just like most of the end-game mechanics people have been finding so useful as of late. The only difference is the diminishing effectiveness of the characters within.

A death spiral is a descent mechanic, what can take it from being annoying to embraced is what happens in the depths. In the metaphor, death is change, and the death spiral is an ideal tool to build the tension and interest as it nears. But to do that the players have to realize that the "death" doesn't mean losing.

In Savagery, my game of social combat, you get three choices in how to play:

Upward spiral, where the characters try to improve themselves.

Making a change, where the characters try to change their lives into something simply different.

Downward spiral, where the characters fall kicking and screaming into madness and depravity.

The later is clearly a death spiral, and yet it forms a compelling type of play. It helps that the death has become madness, and has its own albeit twisted advantages. It's not simply an ending, it's the culmination.


direct link

This makes...
NinJ go "Finding the ending in there is key..."*

*click in for more

22. On 2006-03-07, Peter Nordstrand said:

I'm just happy to be Ben Lehman. He seems like a cool guy ... or I do, I mean.


23. On 2006-03-07, Roger said:

Maybe it's time for a bit of definition:

Death Spiral, or Vicious Circle, or Positive Feedback Loop:  A property of game systems by which the result of one resolution makes that same result more likely for future resolutions.  (See also attractor.)

Example 1:  If a character is hit by his opponent in combat, he suffers some condition which makes it easier for his opponent to hit him again.

Example 2:  A character's ability to increase his wealth is a function of his current wealth, so that the richer he is, the easier it is to become richer still.

The attractor in Example 1 is character death.  The attractor in Example 2 is infinite character wealth.

Counter-Example 1:  If a character is hit by his opponent in combat, he loses some amount of HP currency, but his ability to retaliate is not affected.  There is no feedback mechanism in this system.

Counter-Example 2:  If a character is hit by his opponent in combat, he loses some amount of HP currency, but gains some amount of Rage currency, which makes it easier for him to retaliate.  This is a Negative Feedback Loop.

Counter-Example 3:  The richer a character is, the easier it is to become richer still, but wealth becomes less and less useful as it increases.  This is Diminishing Returns, which moderates (but does not eliminate) the effects of a Positive Feedback Loop.  (It moves the attractor, in other words.)



24. On 2006-03-07, Ninja Monkey J said:

In Counter-Example 3, you see the reason wealth seems to increase greed past a certain point.

So, good definitions, Roger. Either one breaks games.

My favorite strategy games (e.g. Pirates of the Spanish Main et al., Mechaton, Go) give you the opportunity to strike at your opponent's offensive capabilities. That means that sometimes it makes sense to not blow up that guy; it's a waste of time, or even the small risk isn't worth the benefits. Nonetheless, unless you play things juust right, hitting you this turn doesn't increase my chances of hitting you next turn. Finding those capabilities is part of the fun of the game.

I was thinking, back in the early days of the Mechaton redux, about making it so that sensors could be used for other sensors; that is, once one spotlight's on my guy, other spotlights can find my guy more easily. Even that (which turned out to be totally unnecessary) was more of a gentle slope than a death spiral, though; once you actually used the sensor dice, the numbers were all used up; even if you could use sensors to target sensors, it would only do so much.


25. On 2006-03-07, Arturo G. said:

It seems I'm becoming also Ben Lehman. Nice, because I was thinking on writting a game called Polaris about chivalry tragedies, doing extensive use of hit points, of course.

About death-spirals... I'm also finding differences between long-term ones (which may produce/force a given kind of story, or at least a given kind of pace in the story) and short-term ones, which are directly related to how the resolution system works, and how uncertainty is managed during resolution.


26. On 2006-03-07, Chris said:

It's not hard to see the correlation between what we're discussing here and the issue of "predictable stories".  A lot of stories tell you up front whether the hero is going to win or lose, whether you can expect to feel good or have a tragedy, but people can still enjoy them.  Where people lose interest is where the scene-to-scene fiction is predictable.

If we're talking about the Fruitful Void, endgame mechanics don't "paint in" the canvas- they frame it.  But at the point of no return (or rather, high improbability of no return) in a death spiral, the canvas is pretty much already painted in, people are now just wasting time rolling dice/drawing cards/whatever and what's at stake has already been resolved.

A key design concept that allows you to punish the loser/reward the winner without falling completely into a spiral is to give the player tactical options in the resolution in order to offset the penalties and make comebacks.  While lucky comebacks is one way to do it, actual choices keep people engaged.  The loser struggles harder instead of giving up, the winner stays on his or her toes to not lose the advantage.

This tactical choice can be crunchy like Riddle of Steel or Burning Wheel or it can be really loosey-goosey like earning Fan Mail, Gift Dice, or Sorcerer's roleplay bonus dice.

The problem is historically in many games, there hasn't been much in terms of tactical choices that really matter- hence once death spiral kicks in, it's game over.


27. On 2006-03-07, Michael S. Miller said:

In With Great Power... I punish the loser in the short-term—in both game mechanical and fictional ways—but reward them in a long-term game-mechanical way. In the short term, you lose cards from your hand, your opponent wins their Stakes, and your opponent chooses which of your Aspects suffer. In the long term, you take away some of the GM's advantages (wild cards and such).

I've found that presenting the players with the choice of how to lose results in really gripping play.


28. On 2006-03-07, Matt Snyder said:

In fashion: Giving the loser some special say in how things go down. (See: Great-grand pappy Dust Devils, Gramma Trollbabe, Poppa Dogs, and all the kiddies.

Out of fashion: Nine Worlds. No room for losing here. I like it that way. Winners rewarded? Check. Losers punished? Check. Death spiral? Um, jury's still out. Dont' think so, but it ties into ...

Tell you how long a conflict can last: Check—the winners get to decide each time. One phase? Sure. Three phases? Ok. Just keep paying attention to those Muses.

So, the death spiral can happen, particuarly if the winner's feeling merciless AND can keep the story moving as it relates to his Muses. I have learned, however, that the reward system limits often keep winners from doing that to extremes. Pure serendipity!


direct link

This makes...
NinJ go "Keith Senkowski is: JOHNNY ENGLISH"*
Matt S go "Not a death in sight"*
BL go "The one problem with 9W"*
NinJ go "Matt, it wasn't death..."*
Matt S go "The End ..."*
NinJ go "Yeah, gotcha, Matt."
Matt S go "Ungodly typo"*

*click in for more

29. On 2006-03-08, Sydney Freedberg said:

The design toolkit really needs to include both positive feedback (to push situations towards an extreme) and negative feedback (to pull situations back towards a balance)—and then you need to know how to combine them, which is often, as people have been noting, about which cycle to nest inside the other.

And, when you try building these systems, it's startling how much complexity can emerge from simple variations on simple systems—I'm rather proud of this mental exercise:


30. On 2006-03-08, gains said:

The hope of chance is what I often see as the reason a player sees a downward spiral through it's course. Why do they keep going?

Our Dead Hero: I kept thinking my luck would change. That I'd get a good roll and upset.

So, rather than take the punishment as a reason to plan differently the player lets it ride until he's broken the bank. I see this from my players so often it infuriates me. In fact, I'm trying to avoid letting chance have such a broad effect in Contract Work so that you HAVE to plan and manage if you want to survive.


RSS feed: new comments to this thread