If I don't know how to build an airplane, then I just don't know it. Maybe I have a 4% chance to build one, right now, but after a sufficient period of evaluation it will come to light that I either do or do not. When I learn more about building airplanes, then perhaps it will be time for another evaluation.
As I posted upthread, in my experience most failed open lock or remove trap checks in D&D are not narrated by reference to "Schroedinger's backstory" - does or doesn't my PC have training in this particular device - but rather are narrated by reference to the quality of the performance on this particular occasion (eg "You try to pick the lock, but you just can't get the tumblers to budge").
pemerton said:
The reason you can't try again is because the game system doesn't permit retries. Once you're in the pit, you have to proceed from there.
Which is not a reason that derives from the fiction, and is thus meaningless from an immersion/RP standpoint.
You seem to be running together imaginary, in-fiction causation and actual, real-world, game rules causation.
The difference between the two can be seen in the context of resolving a D&D combat. A player declares that his/her PC attacks an orc. The orc has 3 hp left. The player rolls to hit, succeeds; rolls damage, gets a result of 7; and the GM then narrates "You run the orc through; it falls to the ground, dying."
In the fiction, why did the orc die? Because it got run through by the PC. In the real world, why did the GM narrate that the PC runs the orc through and kills it? Because the damage roll is a 7, and hence reduces the orc's hp below zero. If the damage roll had been a 2, then the GM would instead have to narrate something like "The orc parries your blow at the last minute, but suffers a vicious cut across the arm as it does so." And then go on to declare and resolve the orc's action.
This is a pretty simple case where the rules of the game set the parameters for narration of the fiction. But the fiction itself is not generating "reasons" for that narration. There is no fiction of the orc being run through
until the GM narrates it (s/he could equally have narrated that the orc is decapitated, or has a limb cut severed, or any other sort of fatal wound).
Notice also that, in the event that the player rolls a 2 for damage and hence doesn't kill the orc,
there are no retries until after the GM gets to declare an action for the orc. This is not determined by any reason derived from the fiction, either - it is determined by the action economy. The GM might (or might not) narrate some appropriate fiction - eg if the orc attacks and rolls a hit, the GM might narrate "Though you cut the orc across the arm, but the orc's parry leaves you exposed to its counter-attack - take 3 hp of damage". That is fiction that explains the results dictated by the action economy, but it is not the reason that the game was resolved as it was - that reason was provided by the game rules for taking turns in combat.
In a game that takes the "no retries" idea out of combat and into non-combat, as Burning Wheel does, the GM is similarly expected to narrate fiction that explains why retries
by the character do not permit retries
by the player - eg "You try to climb out of the pit, but its walls are too slick for you" or "You haggle with the merchant, but she won't lower her price". But as with the D&D combat example, the in-fiction reason is not the gameplay reason.
I think it's pretty widely recognised that one of the most important skills a GM can develop is managing the relation and interface between the reasons that arise out of a game's rules, and the in-fiction situations that make sense, within the fiction, of the outcomes that the rules mandate. Lazy GMing in this domain produces unsatisfactory play experiences, like combat that is simply "bingo" rather than actually providing any sense of a life-and death struggle between protagonists (think OttS's notorious "duel of clerics"), and players who have no sense of how they might engage the fiction via their PCs and simply wait to be led by the GM.
pemerton said:
Whether a failed knowledge or perception check in BW determines a metaphysical state of affairs ("no chink") or merely an epistemic one ("no chink observable and hence exploitable by the PC") depends on context.
Except one of those results would be ridiculous. If such a thing does not exist, then that state
cannot possibly depend upon your ability to perceive it, such that it would be guaranteed to exist
if only you looked hard enough.
This is like saying "It's ridiculous that rolling a 2 vs rolling a 7 for damage should determine whether or not the orc parries".
The damage roll - which is an event that happens in the real world - isn't exercising causal force in the imagined world. Rather, as an event that occurs within the context of a rule-governed game, it constrains the permissible narrations of the fiction, including in-fiction causal process, that are open to the GM.
Likewise the Perception check is an event that happens in the real world. The result of that roll, in conjunction with other contextual factors (eg the fictional framing of the Perception check), determines what the GM will narrate about the fiction, including any salient in-fiction causal processes.
If the GM permits the action declaration, and it is successful, then the GM is obliged to narrate the chink in the armour. But it's not that the PC looking hard
guaranteed that the chink should exist. The chink has some ingame causal explanation. However, the
player spending action economy on declaring an Assess action (and the GM permitting the action within the current fictional context) means that, if the roll succeeds, the GM is obliged to narrate the fiction, including its ingame causal processes, in a certain way, namely as including armour with a chink in it.
This is somewhat like the roll of a 7 for damage, which obliges the GM to narrate the fiction as including the orc being killed at the hands of the PC. Rolling seven didn't cause the orc to die; being run through did that. But rolling a 7 for damage does not itself "model" running anyone through (eg there are many rolls of 7 for damage that don't kill their targets, because they have more than 7 hit points remaining).
what if someone else wanted to attempt the same task? Would he or she be bound by the same result? Or could they only possibly succeed if they had a better Perception check than the first person?
That would depend on framing. If the failed roll establishes "no chink in the armour", then there is no chink. If the failed roll establishes "you don't find a chink", then perhaps another player can declare an action to enable his/her PC to find a chink the first player's PC missed.
When the Assess action was used in a combat in my BW game, the player of the short-sword wielding sorcerer-assassin was declaring it in order to get a bonus to hit that was needed in order to generate sufficient damage to get through some zombies' damage resistance. In the fiction, she was lining up her strikes so that she struck at their vulnerable bits. In this context, Let it Ride applied to the lining up of each strike, but after resolving each Assess and then each Strike she was allowed to declare a new Assess - the cost to the player was action economy.
the idea of stakes does solve the conundrum of retries by a bit; it's not that you can't succeed, so much as you can't succeed before something happens to make the check irrelevant. The die roll, in that case, would indicate the aggregate of all attempts during that period.
What you describe might be one way of handling "no retries" - that's roughly how the D&D combat action economy handles it (ie you get one roll to represent all your attempts to bring down your foe before that foe then gets a roll).
But it needn't be the only way. If a character is stuck in the pit, it may be that nothing happens to make climbing out irrelevant - there is nothing in the
game mechanics of BW that prevents the game ending because the player doesn't declare any new action declarations for his/her PC and hence that character dies of thirst in the pit.
Suppose, in BW, the PC is pursuing an enemy, and the GM declares "You come to a pit blocking your way. You're pretty sure you can cross it if you want to, but your quarry might get away." After a bit of discussion getting clear on the details of the fiction, the PC's capabilities, etc, the framing might end up this way: "You can cross the pit, but your quarry will have escaped - if you want to track him down, you'll have to make a Tracking check. Or you can leap the pit to keep up the pursuit, but if you fail you're stuck in a pit!"
The player might then say "I'm a trained climber, so even if were to fall in the pit, or barely make it and grab the edge, I could scramble out - so can I FoRK in Climbing?" (In BW, using a related skill to get a bonus is called FoRKing, for "Fields of Related Knowledge".) The GM will probably agree, as BW encourages the GM to be generous in allowing FoRKs and also situational advantages, and there are other elements of the system which discourage players from always using every bonus that they might be able to generate. So the player now rolls Speed (BW's stat for jumping) with +1 die for the FoRKed Climbing. Succeed, and s/he can continue the chase; fail, and s/he's in the pit.
The player can't now declare a Climb check for his/her PC to escape - that s/he stuck in the pit is already settled via the prior check. In the fiction, the narration can be whatever seems appropriate - "The sides are too slick", "Your sore ankle won't bear your weight", etc. At the table, the rules of the game say that the player must declare something else - say a Perception check to find a secret door at the bottom of the pit, or a Circles check to have a friendly person wander buy and find him/her in the pit.
The point of the "no retries" rule - it's official name is
Let it Ride - is to drive gameplay. Not to manage the ingame time that correlates to making a skill check.
I must say I find the burning wheel examples (and the previous bash down a door example) to be weird.
<snip>
Making the penalty for failure be that you make noise is fine. Making the penalty be that the door does not fall, but you have made some level of noise and you can decide on a future course of action, fine. Being told up front that I'll repeatedly hit the door noisily if it doesn't fall on the first hit... weird.
Similarly for the "because you failed to jump between two ships, nothing you do for the next few minutes matters". Surely the consequence is simply that you're in the water and we proceed from there?
It's not about "being told up front" - the GM doesn't tell the player what his/her PC does. The player is the one who makes the action declaration.
If the player declares "I'll keep going until it breaks down" then in BW that is called "working carefully" and grants a bonus - but if the check fails then the GM is entitled to impose a particularly serious time-based complication (eg "Someone turns up because of all the noise you've been making"). If the player wants to avoid the risk of a particularly serious time-based complication then s/he can opt not to work carefully, and thereby forego the bonus - but there are no retries. So if the door doesn't budge, something else will have to be tried.
What the player can't do is just keep rolling and rolling. That's the point of no-retries.
In the case of the ship example, it's not "Nothing you do for the next few minutes matters." That doesn't make any sense. It's "for the next few minutes you do nothing but try and stay afloat while making your way to the side of the friendly ship, where the crew help you scramble back aboard - once aboard you are able to cough the water out of your lungs and regain some of your composure!"
In the session in question, that character's player was absent and so he was being played by another player. Hence having him out of the action actually simplified things for both me as GM and for the players, although it also deprived the players of a useful resource. Had the PC's player been there, I may well have narrated the consequences of failure differently - eg narrating some sort of confrontation or conflict between that character and the rescuing NPCs, which would then keep him occupied while the other characters explored the ship they had jumped onto.
Burning Wheel is intended to be played on the assumption that non-narrated backstory is flexible and will be developed by the GM in order to frame the action and keep the game moving. This applies both to relatively trivial backstory (eg is the crew of the PCs' ship inclined to make fun of, or pick fights of, landlubbers that they have to pull out of the water after they fail to jump between ships?) as well as to more substantial backstory (eg, and to allude to the situation that obtained at the end of my last BW session, what exactly do the orcs intend to do with the Elven princess - one of the PCs - whom they have captured?).
Here is an actual play report of a 4e session that I GMed that was intended deliberately as an exploration-style scenario, where I handled backstory along similar lines. At the end of that post I conclude with the observation that "while sandboxing might rely heavily upon exploration, exploration can be done without sandboxing. Most of the interesting details of the exploration were worked out by me on the fly, whether as needed or even in response to player actions".
