I love the idea of setting stakes for dice rolls, so that a failure still means something happens. Instead of dramatic events getting vetoed by failed dice rolls, I'd love to have it in my games that players all sit up when dice are rolled, because something is going to change.

Now many forge-ites will recognise this, so guys, how do you do it? Are there any tips or "rules" that can help people to set interesting and appropriate stakes, without discouraging action?

That's the general problem, here's my combat/conflict mechanic that flagged it up:
Both roll under their skill, with a bidding of dice to escalate the difficulty of combat. Every player has their own set of stakes and rewards, perhaps balanced like this:
"Ok Jared is going to kill him"
"But if he fails he will die"
"Whoa, ok, I'll just try to injure him then"

The trouble comes with this idea that people can take soft consequences for failure, in order to force the other player to fail, and I want to have it that everyone has something they don't want to loose. How could you set minimal consequences in a situation like this, if someone's objective is only self defence?

I've considered stopping bidding going stupid with a fatigue system based on the number of dice rolled, that might work, but it doesn't really solve the main problem, which is matching dangers to rewards.

Matching dangers to rewards... you know, those two need not be matched all the time. Specifically, in a narrativist context, they can't be matched because you as the game designer or GM don't know what the character or the player values in the fiction. He's a black box to you. The whole purpose of choice-oriented conflict resolution in a narrativist game is often to go through different situations and find those that become difficult choices for the player - the ones where the risk vs. reward is really finely balanced.

That being said, there are many ways to make conflict resolution work, and the best way to learn about them is to get your paws on some games and try them out. Different games have slightly different procedures for making the conflict situations work. For instance, many games with full-blown conflict mechanics do not have explicit pre-roll stakes setting at all simply because they figure the stakes out in some different manner. To pick an example, my own Zombie Cinema does not really have stakes setting because the stakes are derived from character intents - the characters in the situation each want something and declare it, the dice roll is to there to find out whether they get what they want. The stakes of each situation are simply defined as whatever tensions are on the table in the given scene - we continue the post-resolution narrative judgement until there is nothing else to resolve in the scene, at which point we can end it and move on. At no point is it necessary to define stakes, and there is in fact no effort from the side of the system to match risks and rewards with each other - the players in fact do that themselves because the game rewards interesting conflicts and punishes frivolous ones.

That example goes into the secret inner truth of stakes-full conflict resolution: the secret is that stakes as something predefined before the dice roll are really not a mechanical "thing"; they are just communication that clarifies character intent. The vast majority of successful conflict resolution games don't have very elaborate stakes-setting mechanics simply because all that fiddling with possible conflict results is not very fun. You wrack your brain for potential results and then after the dice have been rolled you don't actually have anything smart to say anymore - you already put out your ideas before the conflict, you preplayed it in a way.

Many successful games get over the need to prenegotiate conflict consequences by having a mix of soft fictional consequences and hard mechanical consequences in play as potential results. For example, Dust Devils: in that game you always lose some ability scores when you lose a conflict, no exceptions. The conceit is that there is no such thing as nonviolent conflict, and the result is that we don't really need to negotiate stakes before the conflict procedures. We know that somebody is probably going to get hurt, the mechanics tell us how much, and the post-conflict narration can show us how that hurt came about.

Even games that do not have this sort of mechanicalization of consequences get by fine without explicit stakes negotiation. Primetime Adventures is a good example of a game that one might read as having a sort of stakes negotiation in it: just like your game each character has separate and semi-independent success and failure in the game. But unlike your model, in PTA the players don't negotiate specific outcomes that match the success and failure of the forthcoming conflict - instead they just choose the topic of the conflict, the issue that is "at stake". The issue might be what happens in a given battle, for example; there is no need to specify in detail what happens as consequences, because when the post-conflict narration comes up, the rules require us to resolve the issue - the battle is over, now we need to narrate how the victory or defeat in the conflict reflects in the fiction. Thus the conflict consequences are created after the conflict resolution, not before.

--

Now, after all that we have one thing down - the majority of successful conflict resolution games do not, in fact, do explicit consequence-setting before the conflict. How do we then ensure that failures have consequences? In many games this is accomplished simply by pacing and dramatic coordination - I write about this specific issue a lot in the Solar System, which has a conflict resolution system without mechanical stakes-setting. The way it works in that game is that if your character fails in doing whatever it was that he was trying to do, then not only you can't try it again, but the world around the character goes merrily on its way, the situation changes and the opportunity is lost for good. So if you're trying to kill another character and fail? The failure narration not only describes how you fail, but also how the other character escapes the situation and the whole scene ends for good. If you want to kill him, now you're going to have to find him first, probably while he's doing the exact things you wanted to prevent by killing him.

This is not to say that you can't have "stakes-bidding" as a mechanic, it's just that speculative play is not very fun, so games that apply those sorts of mechanics tend to discuss consequence negotiation more than stakes-bidding. For example, Polaris and In a Wicked Age both have conflict resolution mechanics that allow the players to negotiate the conflict consequences. Those negotiations are, however, held within the mechanical constraints of the game system, and they happen in the middle of the conflict resolution procedure - in fact, they are the procedure. Polaris, say: there is no dull pre-playing of conflict outcomes where you state outcomes that never come to pass; instead, you suggest interesting outcomes and the opponent gets to add to them, or gets to counter them by paying for such counters.

What Eero said.

Best, Ron

JoyWriter wrote: The trouble comes with this idea that people can take soft consequences for failure, in order to force the other player to fail, and I want to have it that everyone has something they don't want to loose. How could you set minimal consequences in a situation like this, if someone's objective is only self defence?

I've been mulling over these responses, because although I take Eero's point that you can't set "appropriate" consequences from a player perspective, because you don't really know what they want. Interestingly, this hasn't weakened my liking for the specific system, which suggests that there is something else going on there for me!

Even if there isn't a subjective balance I like the idea of incorporating the ironic/tragic consequence structure of old-school myths, where the spy is blinded, the liar is deceived, or the blackmailer is shamed. The classic version of this is "He who lives by the sword dies by the sword".

Here are some rules I have come up with for the consequences:

Action to someone else, consequence to you.
The consequence should stop retries of the attempt by in-game fiction alone
There should be a sort of inverted symmetry to the actions, so they are opposite but of the same type
If the consequences involve another player, they should be able to shape the result, so you don't get pacifists killing people etc.

On the last note, I find it amusing that I start making PG rated hollywood films when I try to imagine a pacifist fighter verses a hardened killer: "So he falls of a ledge, no, he gets blown up by his own superweapon, no he is electrocuted by a falling pylon, etc"

The interesting thing about it is that the consequences for the action "belong" to the person doing them, but they get shaped by the personalities of the players around them, and the nature of what they have created.

I don't want to make the actual narration occur before the roll, but have people create things like the above proverb, but suited to the character, so that by the inevitabilities of probability, their actions eventually come back to them!

It's only when you know the results of the two rolls that you can work out the narration, also taking into account somehow the "energy" of the conflict, which was set by the number of dice rolled. I almost wonder whether it would be possible to do an inverted wushu system, where you have to add details for each dice, but I think that that inversion misses the point of the incentives in wushu.

I'm not sure how to codify the differences between a quick one dice combat and an intense 5 dice one, I only hope that the players would be sufficiently invested in the high cost ones to do it naturally.

One further possibility, if more codifying is needed, is to put the different action types in the hands of the gods, or cosmic forces. Interesting, and gets ritualised fast, but not quite what I'm going for.

I'm not in favour that much of consequence bidding, because that requires ridiculously invested players, and then works directly against what they like: "Do you still want to participate in this game? How about now?" I'm not altogether confident that I can whip up sufficient investment to make that work, without risking overbidding and bankrupting the game! I suppose you have to tread very lightly, or have something like "you ask too much" from  Polaris. To take another angle, perhaps whipping up investment is the wrong way to think about it, but I'll leave that for another time/thread.

I agree about the flux of events as a way to make failure matter, but that has a very different feel to it; "take your opportunities before they pass.". It seems like that too, implicitly defines the world; the world must have it's own dynamics, and change itself even if you don't. A good for challenge for GMs, making the "you can't do it again" stick in an coherent way.

I heard a story from a PTA game (I believe from the Son of Kryos podcast) that seems relevant to this discussion.

The stakes for a conflict were that a player was trying to force an npc to tell him something.  He failed but won narration rights, but the stakes were only that he forced him to say something.  So the player has freedom to narrate him getting the information he wanted, but the npc says it of his own volition.

Huh? Doesn't that make the whole conflict roll to force the NPC to tell him something entirely pointless? You fail it, then just do the same thing with narration rights?