Task Resolution v Conflict Resolution

[angelfromanotherpin]

Now, I know Ron Edwards is held in very low opinion on these boards, but all the Apocalypse World discussion got me thinking about something I read by him which I always thought was pretty valid.

Imagine the following situation: a player-character James Bond expy has snuck into an adversary's office, believing that important plans are in his wall safe. The player declares intention to crack the safe, and rolls for it.

The task is 'open the safe.' The goal is 'get the plans.'

The conventional expectation is that there are two likely outcomes:
1. Task succeeds, goal reached: The safe is opened, the plans are got.
2. Task fails, goal not reached: The safe is not opened, the plans are not got.

But there are two others to consider:
3. Task succeeds, goal not reached: The safe is opened, the plans are not got.

Once opened, the safe turns out to be empty, or have only money and personal papers, or a porn stash or something.

4. Task fails, goal reached: The safe is not opened, the plans are got.

Defeated by the lock, the spy turns away in frustration, only to find the plans sitting on the desk, or in the printer, or the trash or something.


The second pair of outcomes are likely to feel like plot-nudging contrivances, even if the specific location of the plans was put down by the MC in writing months beforehand. As a player, there is the expectation that success will come with a reward, not with an anticlimax (or indeed with bear attacks) and that failure will come with consequences, not with rewards.

The difficulty is that a task resolution system by its very nature can only tell you whether or not the safe gets opened. And on that level, it is trivial for in-game circumstances, whether intentionally contrived or accidentally arrived at, to render that task meaningless. The arbitrarily-declared location of the plans matters more than the safecracking roll can matter.

Ron's suggestion was to shift a game's focus to goal-based resolution, where the roll determined the outcome of the stated goal. At the same time, it was acknowledged that depending on the desired scale and pace of a particular game or scene, the stakes of any individual declared goal could be grand (win the battle), moderate (defeat this enemy) or modest (cause normal weapon damage to this enemy).

One result of this train of thought that I really liked was a greater freedom to reflavor successes and failures. As an example, let's go back to the James Bond expy, who is now attempting to charm a lady so he can get access to her husband. The player rolls for it and gets the worst possible critfail, so he's not only getting no access, but he's getting negative consequences on top of that.

From a task-based perspective, Jimmy Bond has to completely foul up his charm attempt. Which is unfortunate, because having an established smoothie like him start stammering like an idiot and spilling his drink into the hostess' cleavage is an immersion-breaking experience.

On the other hand, from a goal-based perspective, all that matters is that Jimmy doesn't get the access he wanted (plus an extra downside). So maybe he does charm the lady, but when she goes over to her husband to talk about how cool this guy she just met is, the husband gets jealous instead and has Jimmy escorted out of the party. That outcome respects not only the result of the roll, but also the integrity of the character.

Thoughts?

[shadzar]

4a. Task fails, goal reached: the safe is not opened and the plans ARE inside, but the safe is stolen to be opened later to extract the plans

which creates 3a also in the same vein but really not required, unless you take into account for 4a, that the safe now missing causes something to happen not unlike the safe being opened.

[Foxwarrior]

Well, it means that the player no longer has to think about which tasks they want to attempt, only their goals.

While it does work for your James Bond example better than the D&D implementation where the DM decides whether the husband is jealous and James Bond has +$TEXAS to Seduction, consider this counterexample:

The player wizard is fighting a Dragon.

In task-based resolution, the player says "I fly up to the Dragon and cast shivering touch on it."

In goal-based resolution, the player says "I roll to defeat the Dragon."

[angelfromanotherpin]

In task-based resolution, the player says "I fly up to the Dragon and cast shivering touch on it."

In goal-based resolution, the player says "I roll to defeat the Dragon."

This sort of thing was addressed.

At the same time, it was acknowledged that depending on the desired scale and pace of a particular game or scene, the stakes of any individual declared goal could be grand (win the battle), moderate (defeat this enemy) or modest (cause normal weapon damage to this enemy).

Goals can be big or small (and so can tasks). I don't think anyone is on board with 'roll to win campaign' or 'roll to nock arrow,' but there's a lot of room in-between.

[Foxwarrior]

The goal was already a larger scope in the Jimmy Bond example ("get access to the husband", not "charm the lady"), and if the Wizard successfully casts shivering touch on the Dragon, the fight's basically over.

But actually, I'm tempted to change my mind about Jimmy Bond. A D&D-style system where Jimmy Bond simply cannot fail Seduction checks (a necessary thing for representing the character) can have plenty of strategic thinking (given the right rules), where Jimmy needs to decide how much effort to spend researching the lady or her husband.

[Sashi]

This is a really weird way of dealing with what is actually an MC-player communication issue.

Defeated by the lock, the spy turns away in frustration, only to find the plans sitting on the desk, or in the printer, or the trash or something.

The real problem is that this is a shitty resolution. How did he find the secret safe without searching the desk first? Why does he automatically assume the plans are in the safe when there's also a computer, file cabinet, and waste basket in the room? What information did the MC give him to create this expectation?

If the plans are laying out in plain sight, then Bond should have found them with the search check he used to find the safe. He shouldn't be opening the safe at all except to get at Blofeld's secret persian cat porn stash.

The plans could also be in the safe, which Bond fails to open, but there's a partial set of plans with a printing error left in the trash by a stupid underling. They're not as good, but it's all Bond can escape with and live.

The fact is that you have to overturn player expectations sometimes. It may be to surprise them, it may be because their expectations are dumb. But the idea that player expectations must be fulfilled at all costs, or that it's simply better to not have player expectations than it is to overturn them is dumb.

3. Task succeeds, goal not reached: The safe is opened, the plans are not got.

Blofeld purposely leaked that the plans were in the safe. It was actually a trap and the safe releases knockout gas into the room when opened.

4. Task fails, goal reached: The safe is not opened, the plans are got.

Bond fails to crack the safe and sets off an alarm. Bond hides, Oddjob rushes in and opens the safe to evacuate with the contents. Bond fights Oddjob and wins, grabs the safe contents for himself, and fights his way out of the mansion.

[Sashi]

Though thinking about things in terms of goals and tasks might be a good way to divide up complex and discrete checks.

Discrete tasks (punch a dude, jump 10 feet, open a door) should be single checks.

More general goals (interrogate a dude, climb a mountain, hack a mainframe) should be complex checks.

[silva]

One advantage of conflict-resolution - specially in AW implementation of it - is how fast the game flows. It may get even frantic at times.

After having a couple of exciting combats taking 10 minutes of real gaming time, we dont wanna go back to Shadowrun hour-long combats ever again.

[Parthenon]

It does mean that everyone has to verbally state what their aim is and what their action is that they are attempting. Otherwise Mr. Cavern needs to guess what their aim is and so sometimes get it wrong and have them fail their aim when by the roll they should have moved towards it.

Which means it isn't possible to have many forms of inter-party conflict or have hidden aims. Some of the best stories and best experiences are based on players with opposing aims which doesn't work when they tell each other their aims.

[silva]

It does mean that everyone has to verbally state what their aim is and what their action is that they are attempting. Otherwise Mr. Cavern needs to guess what their aim is and so sometimes get it wrong and have them fail their aim when by the roll they should have moved towards it.

Agreed. And not only that, but the GM should always say whats the stakes are before the roll is made.

Player: "I wanna sabotage this engine and also make it look like it was Adam fault".

GM: Ok, but if you miss the roll, the shit can blow up in your face and you may let evidence that it was your fault all over the place. Wll you do it ?

Say the stakes and ask.

Which means it isn't possible to have many forms of inter-party conflict or have hidden aims. Some of the best stories and best experiences are based on players with opposing aims which doesn't work when they tell each other their aims.

Disagreed. Both Apocalypse World and Mountain Witch are conflict-resolution based and both promote inter-player conflict. The later even includes Dark Fates that each character has occultly against each other. And both has a Trust mechanic in place, which means treason may come out of nowhere:

Player A: "I will sabotage this engine and make it look like it was Adam fault"

Player B: "Good, you do that while I cover you".

GM say the stakes and ask the player to confirm the intention.

Player A confirms, but before it rolls, player B interferes..

Player B: "I will sabotage his attempt on sabotaging the engine, to make it so it explodes in his face".

GM say the stakes and ask: "Ok, roll his Trust score. If you suceed you impose a -2 penalty to his roll. If you fail you suffer the same xonsequences of his roll - including getting blowed up together with him. Will you do it ?"

Player B:"Sure".

Player A: "Motherfucker"

[Foxwarrior]

But at this point, Player A knows that Player B is a traitor. The coolest treason scenes are when a player reveals that they've been messing with the other players for a long period of time, not when a player attacks another one spontaneously.

[silva]

Indeed, Fox.

Its important to point out though, that for a player to betray another in an efficient way its recommended that he have gained the targets´ trust. This is made in game by receiving “trust points” (“Hx pts” in the case of AW), at the end of each session. And a player only deposits trust pts to another one that he judges a partner/wont betray him, or whose support he needs ( since a trusted player can give you bonuses to actions happening nearby ).

This creates an interesting dynamic where each player consider carefully who to deposit trust points to. “Did Wilson help in last session was sincere ? Or he just wanted me to deposit him trust pts for betraying me later ?” or “I really need Claire´s help for breaking through the camp tomorrow.. should I deposit trust on her ?”

So, while the mechanics don’t support actual behind the back acts of betrayal, it does support a kind of behind the back scheming and paranoia that can lead to Reservoir Dogs-style betrayal-fests. (The Mountain Witch is specially good at this, because of the Dark Fates and how its resolution is solely Trust-based ).

[CCarter]

A thing I've heard complaints about with regard to some storygames is that the level of abstraction means that player ability to influence game events is limited to what inputs the system can handle.
So its not just the GMs ability to meddle with stuff at the task level is curtailed, its also the players/PCs ability to mess with stuff at the task level.

For instance sometimes, a task resolution may actually be described in fictional terms after the dice roll -the task resolution roll gets made, the PC passes or fails, and the player's role is limited to interpreting the dice roll. In Dogs In the Vineyard say, you would choose dice rolls that would have a mechanical effect in the social conflict (to Raise or See or Take the
Blow or whatever) then would have to roleplay in accordance with that, instead of having role-playing have a mechanical effect at the outset. Compare that with a traditional RPG where a good player speech might give a bonus to a PCs seduction check, etc. You can imagine similar problems with other highly abstract situations.

There may also be a tendency to abandon working explanations into the fiction at all, since its unnecessary to use whatever bonuses.

(I'm getting this partly from an interview with Vincent Baker here
http://redmoosegames.blogspot.com.au/20 ... ncent.html that touches on these problems with the storygame approach here; if you wanna waste your life its basically blather for the first 15 minutes but it starts explaining Baker's clouds-and-arrows diagram - http://www.lumpley.com/archive/156.html - at 16:20 with some good stuff between 24 and 34 minutes, which is about where I stopped. He doesn't tie the problems back to the idea of 'conflict resolution' though, that's just my take on it).