D&D 5E A case where the 'can try everything' dogma could be a problem

[pemerton]

because I want some real-time unpredictability in the game.

The unpredictability issue was discussed a bit upthread. [MENTION=22424]delericho[/MENTION] said that he (? I think) likes unpredictability, and that this resides principally in the players' choices as to what to do with and about the backstory they know. I agree, but because unpredictability reside principally in the players' choices, don't feel that there is any great benefit in uncertainty as to access to backstory per se.

The comparison I made upthread is to making a d20 roll to see which module you use, rather than just choosing the one you think will be most fun.

But there have been instances where some elements of a PC's backstory might be randomly generated. It's not generally done, because PCs are usually placed under the player's control, including backstory, but that's not absolute.

<snip>

on the one hand you're talking about the PC's backstory and on the other you're talking about the setting's backstory.

I don't think so. The question of what my PC learned in school is a fact about my PC, not a fact about the broader setting.

If I don't have to roll to learn whether or not my PC was taught cobbling at school (but instead just pick the cobbler background, or cobbler skill, or whatever), why do I have to roll to learn whether or not my PC was taught about such-and-such a secret cult?

Is it about fairness? Eg it would be unfair for a player to declare that his/her PC knows every secret about every cult, but equally it would be unfair for his/her PC to know nothing about any secret cult, so the information is rationed by dice rolls? I think something like this is what is going on with the classic D&D thief's Read Languages ability, but that is in a broader context of information rationing in which the notion of fairness has purchase.

Whereas in non-Gygaxian games (that are also different from some of the non-Gygaxian approaches that I've described upthread, eg BW) the broader context that makes sense of information rationing is what I'm trying to understand.

If I choose not to give the players all the information that is tied to the icon then I'm deliberately choosing to force them to make choices based on bad data. I don't do that - hence the roll.

But how is it better to force a roll that might make them make a choice based on bad data? If making choices based on bad data is bad, how is it less bad if it happens sometime based on rolls, rather than always based on GM mandate?

Or, to flip it round, if it sometimes good for the players to make choices based on bad data, why is it not sometimes good for the GM to mandate that such a situation will occur?

I know the answer to these questions in Gygaxian play: because acquiring the data is itself an element of player skill. I don't think that answer generalises to other playstyles, though.

It seems to me there is a difference between the game coming to an unexpected end because the players declared actions for their PCs, and failed due to bad luck; and the game coming to an unexpected end because the players missed out on relevant backstory due to random dice rolls.

Actually, they're exactly the same. In both cases the players are making a bunch of choices, some or all of those choices result in dice rolls, and if those dice rolls go bad then the game ends.

The only real differences are that in the combat case those dice rolls come in quick succession and lead to a definitive end (TPK), while in the investigation example they're considerably more spread out and probably don't - very likely there is at least something the PCs can still try to get back on track.

I don't see them as being at all the same.

For instance, in the combat case the GM can't stop the TPK (assuming no Deus Ex Machina) other than by suspending the action resolution rules (eg fudging die rolls, or overtly declaring that the NPCs miss rather than hit). Whereas in the case of the religious icon, the GM can just declare that "Yes, your PC did learn about such-and-such a secret cult when studying at Obscure Knowledge College."

The analogue to the TPK in the investigation scenario would be the players not declaring that their PCs search the room, or not declaring that their PCs look at the icon. But once they have searched the room, and have declared that they examine the icon, the players have made all the action declarations that they can - it all turns on what backstory is authored about their PCs (what did they learn at college) and hence what backstory the players become privy to.

It would be possible to construct a combat that had a similar dynamic - eg the demilich puzzle in Tomb of Horrors - but in the typical FRPG combat encounter, the ability of the players to make meaningful and effective action declarations doesn't generally turn on having access to some particular bit of backstory which is rationed like the icon information.

In the investigation case, it's vastly more likely that if the PCs blow all their rolls they'll find themselves with no obvious way to proceed and have to revisit some locations and/or go visit some helpful NPC for more assistance.

If the PCs have to return to locations because the players didn't make sensible action declarations (eg they didn't search the desk, or whatever) that is quite different from the knowledge roll scenario. It's like making a bad choice in a combat, and you wear the consequences.

If the visit to the NPC has a cost, then that becomes like the "stakes" examples we discussed upthread - the reason for the knowledge check is to find out what resources, if any, the players have to forfeit in order to get the information.

If there is no cost other than time spent at the table in visiting the NPC then again the rationale of the whole structure becomes more opaque to me. Couldn't we just short-cut straight to the information - either by declaring that the PC knows it, or describing how the PC visits his/her friendly old mentor who tells him/her all about the nefarious cult of so-and-so?

Would it help if I said it was due to a simulationist, rather than narrativist, approach? That is, we're not rolling to determine how much of a "backstory ration" the PC gets; we're rolling because the PC knows a certain amount about religion which may or may not include the facts tied to this icon. So we roll to determine whether he happens to know these particular facts. That these facts tie into the story being told is incidental to the process - we'd roll just the same if the scene instead just happened to take place in a (completely unrelated) church and the PC chose to look at the stained glass windows.

That helps with the aesthetic, yes. I think it raises some other questions, though, or at least leaves them unanswered, like what is "the story being told", and who is telling it?

I am 'rationing backstory' to let the players collect potentially useful clues.

<snip>

I have never made a certain check-dependent piece of knowledge critical for the successful completion of the whole adventure

<snip>

I want to use Knowledge check rolls for a couple of reasons: because players make decisions about what their PC are proficient at, and such decisions have a cost

This seems to me to be getting close to the heart of things, but some elements are still a bit unclear to me.

I agree that PC build choices, including acquiring Knowledge skills for a PC, have a cost. The question is, then, what is the player buying for that cost?

In Gygaxian D&D, the answer is fairly clear: the ability to discern backstory (secret doors, traps, etc) which will enable a skillful play to engage in even more skillful play, and thereby increase the XP & treasure per session ratio.

In 4e, the answer is fairly clear: the ability to learn monster stats during combat, which can be an advantage in that system; and the ability to declare actions in skill challenges (as was discussed upthread, 4e knowledge skills also include "active" aspects, like using Nature to find the way or to tame an animal, or using Religion to exorcise an evil spirit).

In Burning Wheel, the answer is also fairly clear: knowledge skills enable action declarations which will permit player authorship of backstory (eg the "chink in the armour" example upthread, or "locking in" that the feather the hawker is selling really is an angel feather).

In both 4e and BW, knowledge skills can also be used to gain access to the GM's secretly-authored backstory, but that is not their sole use. When used in this way, the main benefit the player gets is to force the GM to declare, and thereby "lock in", the backstory. It's a sort of "tempo" benefit.

In the style of play which emphasises "collecting potentially useful clues", what is the benefit to the player in collecting those clues? How do they relate to the adventure - and what exactly is the adventure (or, in [MENTION=22424]delericho[/MENTION]'s terminology, the story being told? If a group of players turns up with PCs completely lacking knowledge skills and divination magic, what will they not be able to do, in terms of engaging the game and doing well at it, that they could do if their PCs did have those abilities?

 

[The Crimson Binome]

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").

The system you are describing is one which should allow re-tries for such a check. Many games of D&D include both that narration, and a mechanic for making multiple attempts (often along the line of a critical failure causing a jam or break which prevents subsequent attempts). The system which I describe is one which explains why additional checks cannot be made. If a system doesn't allow multiple attempts, but the GM is explaining the attempt in such a way that this would not logically follow, then that's a problem with the GM.

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."

The real world doesn't exist, as far as the game world is concerned. Nothing in the game world can happen because of something in the real world. I think we're in agreement on that, at least.

Where we seem to be in disagreement is that I purport every action the player declares to correspond strongly with actions that the character takes, and every die roll to have inherent meaning within the game-world. Every stat on your character sheet represents some truth about the character, and every skill check you make represents how that skill allows for your success or failure at a particular task. The reason why you're rolling a Perception check in order to find an item must be because your Perception is the deciding factor on whether or not you can find it, in order for causality within the game-world to be maintained; in any case where that thing doesn't exist, your Perception would not be the deciding factor, and no such check should be allowed.

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).

The broad strokes of the fiction are governed by the uncertainties of the situation (in the die roll) and the abilities of the participants. The GM isn't compelled to narrate the orc's death based on some agreed-upon narrative power-control arrangement; rather, the fiction of the orc's death follows directly from whatever fictional reality corresponds to damage interacting with whatever fictional reality corresponds to HP.

In any sort of Simulationist game, there is a strong correlation between each game-level element and its in-world representation. (Story-based games require the sacrifice of such correlation in order to further other goals.)

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.

You are mistaken on this point. The action economy has strong correlation to the fiction it models. The reason why the character can't try again (until the next round) is that any action requires time to perform, and there's not enough time to make a second attempt before the orc gets a chance to retaliate. A very small part of this comes down to necessary concessions in modeling real-time events with mechanics that human players can implement at the table, but the bulk of it is comes down to the reality of the game-world that the character can only inflict so much injury upon an opponent in a given period of time (while simultaneously avoiding other attacks, moving, and doing whatever other background tasks are going on).

 

[The Crimson Binome]

No edition of D&D has ever had any sort of algorithmic process for resolving the action declaration "I manipulate the forces of entropy so as to seal off the Abyss from the rest of the multiverse". That is always going to require the GM to exercise discretion in its adjudication (eg, at a bare minimum, what skill is to be tested?).

I know that 2E and 3E had explicit rules for this. Namely, you would spend months in your tower in order to invent a new spell that was capable of doing this. It just wasn't the sort of thing that you could do on the fly. D&D wasn't that sort of game.

My point is that Gygaxian D&D is roleplaying - given that it was the first instance of an RPG - and it doesn't abide by the strictures you set.

Words have meaning, and definitions can change over time. He may have popularized that term, in referring to that game, but that doesn't mean the description necessarily still applies today. From what I've gathered, based on his quotes, he was a hard-core Gamist who looked down upon role-players as wannabe play-actors. In fact, I wouldn't be surprised if his use of the term was meant more in the 4E sense, as the role a character fills within the party.

It's hard to reconcile that style of gameplay with any modern definition of "role-playing". (I could imagine including that type of game under a broad RPG tent that also included story-telling games, but that sort of Gamist approach to the game would not qualify as role-playing.)

 

[Andor]

What would have happened if the players decided to claim their rights to all try a roll of any knowledge each time?

1) With four characters, they would have succeeded almost always. Unless I artificially increased the DC.

2) The proficient character would only marginally succeed more often than the others, at least at low level when the proficiency bonus is only +2.

3) Some unproficient PC (the Rogue) had a much higher Int than the Cleric. So the Cleric player would be actually penalized by choosing the Religion skill, rather than just stick to Wis-based skills.

Nevertheless I wanted to share this, because IMHO it is something to be aware about.

I also want to point out that this is a problem only with skills that benefit the party as a whole + have no penalty for failure + have no better effects if more PC do the same. Knowledge skills are the main case, but also Thieves' Tools and Investigation are others, and some charisma skills might also be (but it depends how you adjudicate a simultaneous success by someone and failure by someone else).

I'm coming late to this discussion and I haven't read through 8 pages of discussion to make sure I'm not repeating already offered thoughts, so apologies if I'm stepping on any toes.

I would question the "have no penalty for failure" examples you offer, many GMs I've seen will give bad information on a failed knowledge roll, in fact I often see players invent their own wild goose chase ideas when they get a bad roll. Similarly a failed investigation check might lead to bad data or tip off the person you are investigating.

The problem is that with public rolls everyone knows who succeeded and who failed so only the right answer is heeded, in spite of the fact that this is technically meta-gaming. But you don't want to take agency/dice karma away from the players.

My solution to this is that when I GM I hand out 3x5 cards to all the players and have them write down their characters name and the stats of interest to me, as well as a series of 10 d20 rolls. Then through the gaming session when a hidden roll is called for I simply use the next number on the list and cross it off.

So if everyone wanted to make a Knowledge Plot test to ID the Macguffin I would give them all answers based on how well they failed or succeeded, what I would not do is tell them who failed or succeeded. Same thing for perception tests. Occasionally if the game is getting slow I'll go through the motions of crossing off a number from each card while looking around the table and making "Mmm-Hmm" noises, then say nothing.

 

[KarinsDad]

Finally, we got to play the real 5e game also on tabletop (before last week it has been only PbP or tabletop with playtest packets). I was the DM and we had both experienced and beginners players at the table.

With relation to skill proficiencies, I made sure everybody knew that in 5e you don't need to be proficient in order to do or use something. Rather, it is convenient for you to use/do what you're proficient at, instead of using/doing things you are not. The simplest obvious example is with weapons: with an emergency or a specific circumstance you might pick up and swing a weapon you're unproficient, but on the long term (i.e. normally) you'd better use weapons you are proficient.

So far so good... but not with skills!

The players spontaneously gravitated towards using skills they were proficient at. The Rogue had Stealth and Thieves' Tools so naturally she went scouting and searching for traps, while the others waited and watched. This felt just right!

But the players also spontaneously did the same for Knowledge skills, and here is the crux of the matter... I was very happy that they did this on their own volition, but I dread what would have happened if they exploited the system. Because in theory, all of them could have rolled knowledge checks each time it was potentially useful.

So what they did was simply, the Cleric (and only he) rolled Religion & History checks, while the Wizard (and only she) rolled Arcana checks. Whenever they needed a clue, players asked "is it worth rolling an X knowledge check here?" and they everybody looked with hope at the player rolling the check. In other words, (s)he got the spotlight, and it felt great when succeeded, and hilarious when failed (I admit that was in part thanks to me making up ridiculous results on a failure instead of just saying "you don't know").

What would have happened if the players decided to claim their rights to all try a roll of any knowledge each time?

Players do not have that right. As DM, I control when rolls are made, not the players. And typically, I have the trained PCs make knowledge checks. If the PC trained in religion does not know, the group does not know. Ditto for using the Help action for skill rolls. Sometimes I allow knowledge rolls from players not proficient in a skill, but that tends to be rare.

 

[pemerton]

The real world doesn't exist, as far as the game world is concerned.

Obviously (except in some self-referential games like Over the Edge).

Nothing in the game world can happen because of something in the real world. I think we're in agreement on that, at least.

From the point of view of the gameworld, it has its own internal causal processes.

But considered as the fictional thing that it is, everything in it is a consequence of a choice of authorship made in the real world. It doesn't write itself.

The GM isn't compelled to narrate the orc's death based on some agreed-upon narrative power-control arrangement; rather, the fiction of the orc's death follows directly from whatever fictional reality corresponds to damage interacting with whatever fictional reality corresponds to HP.

But your formula here contradicts your own rule about no crossover. In the fiction, the orc's death is not a fiction - it is a fact. And in the fiction, the fact of the orc's death follows from the fact of him/her being stabbed, or decapitated, or whatever.

But how is that fiction authored? In D&D, the mechanics of a damage roll in and of themselves don't dictate anything - nothing at all about the fiction follows from a damage roll of "7" rather than "2". (Note the contrast with, say, Rolemaster, where a damage result of "foe decapitated" does directly tell us what the damage is; and also where the loss of (say) 10 concussion hits does correspond directly to a particular degree of bruising and blood loss.)

Even when you compare the damage roll to the NPC's hit points remaining, all the rules tell you is that the NPC is now up, or now down. They don't tell you anything about the sort of blow that was struck, nor the cause of death. (Again, games like RQ, RM, and even BW are quite different in this respect.)

The GM of the D&D combat has to make an authorship decision. (Or else no one at the table will know what happened in the fiction.) You don't have to call it an "agreed upon narrative power control arrangement" if you don't want to - I don't. I just call it playing the game within the parameters set by the rules.

Where we seem to be in disagreement is that I purport every action the player declares to correspond strongly with actions that the character takes, and every die roll to have inherent meaning within the game-world. Every stat on your character sheet represents some truth about the character, and every skill check you make represents how that skill allows for your success or failure at a particular task.

There is no disagreement here. In BW, for instance, having a high Perception score expresses a truth about the PC - namely, s/he is perceptive. The action declaration "I look for a chink in the armour", and the ensuing dice roll, corresponds to an attempt by the PC to find a chink in the armour.

Likewise a Circles check - having a high Circles score expresses a truth about the PC, namely, that s/he has many social contacts (having moved in many social circles - hence the label). The action declaration "I check Circles to see if someone has come to rescue me", and the resulting roll, corresponds to the PC hoping and looking for a rescuer.

In other words, the point of difference is not about whether the PC sheet tells us something about the PC, nor about whether action declarations relate to events in the fiction. It is about who gets to author the fiction - both backstory and current events - and under what sorts of constraints. This also plays into related issues about framing.

To focus just on the "chink in the armour" example, for instance, you take the view that only the GM can author this, and that the GM must do so (perhaps secretly) before the action declaration to look for it can be declared. (And if the GM authors it as "no chink", then that action declaration will result in failure even if the GM lets the player roll the dice.)

This is why, in my reply not far above to [MENTION=22424]delericho[/MENTION] and [MENTION=1465]Li Shenron[/MENTION], I concluded by asking some questions about what is meant by "the story being told" or "the adventure". The key issue here seems to me to be about who enjoys what sorts of privileges of authoring the fiction, and related issues like secrecy vs transparency of backstory.

 

[pemerton]

I know that 2E and 3E had explicit rules for this. Namely, you would spend months in your tower in order to invent a new spell that was capable of doing this. It just wasn't the sort of thing that you could do on the fly. D&D wasn't that sort of game.

"Creative use of spells" has always been an element of D&D. And I seem to recall that the 3E rulebooks talked about various sorts of ad hoc use of Turn Undead in order to channel positive energy so as to achieve magical effects like closing Abyssal portals.

Moldvay Basic had guidelines for the adjudication of action declarations that the rules don't cover.

There may have been some D&Ders - and it seems that you are one - who treated the formal action resolution algorithms as exhausting the space of permissible action declarations, but the rulebooks that I am familiar with have never asserted this, nor even implied it.

The action economy has strong correlation to the fiction it models. The reason why the character can't try again (until the next round) is that any action requires time to perform, and there's not enough time to make a second attempt before the orc gets a chance to retaliate. A very small part of this comes down to necessary concessions in modeling real-time events with mechanics that human players can implement at the table, but the bulk of it is comes down to the reality of the game-world that the character can only inflict so much injury upon an opponent in a given period of time

This is implausible even for a 6-second combat round - is human motion in the D&D world quantised into discrete 6-second periods? (Although even that doesn't quite work, because the stop motion aspect of resolution in 3E and onward makes things like peasant railguns possible.)

For a 1 minute round, which was the AD&D standard, it's completely implausible. That's why Gygax explained it in these terms (DMG, p 61):

One-minute rounds are devised to offer the maximum of choice with a minimum of complication. This allows the DM and the players the best of both worlds. The system assumes much activity during the course of each round. Envision, if you will, a fencing, boxing, or karate match. During the course of one minute of such competition there are numerous attacks which are unsuccessful, feints, maneuvering, and so forth. During a one minute melee round many attacks are made, but some are mere feints, while some are blocked or parried. One, or possibly several, have the chance to actually score damage. For such chances, the dice are rolled, and if the"to hit" number is equalled or exceeded, the attack was successful, but otherwise it too was avoided, blocked, parried, or whatever. Damage scored to characters or certain monsters is actually not substantially physical - a mere nick or scratch until the lost handful of hit points are considered - it is a matter of wearing away the endurance, the luck, the magical protections. With respect to most monsters such damage is, in fact, more physically substantial although as with adjustments in armor class rating for speed and agility, there are also similar additions in hit points. So while a round of combat is not a continuous series of attacks, it is neither just a single blow and counter-blow affair. The opponents spar and move, seeking the opportunity to engage when on opening in the enemy's guard presents itself.​

Notice particularly how the abstraction of the action economy and the abstraction of hit points interact - part of the reason why we don't need to decompose the events of the one minute round into the detail of actual manoeuvres, strikes, parries etc performed is because we are assuming that what is being achieved is the wearing down of one's enemy, until the final decisive blow is struck.

The six-second round is, in my understanding of it, no different. The reason for breaking down to 6 rather than 60 seconds is rather to allow for tactical movement in a context where distances are meaningful but feasible for adjudication on a table-top grid, and to allow for turn-by-turn rather than simultaneous and/or continuous resolution without the stop motion effect getting wildly out of control.

When I think of RPG systems that try to actually model the striking of each blow, the making of each parry, etc, they have an inherent tendency to shorten and shorten the round so that what you claim - namely, that the action economy models causal possibilities - becomes plausible. Thus, many highly simulationist combat resolution systems have 1 or 2 second "rounds" (GURPS, HARP, Burning Wheel, various Rolemaster options, RQ's "strike ranks", etc).

In any sort of Simulationist game, there is a strong correlation between each game-level element and its in-world representation.

This may well be so. But if so, it follows that AD&D combat is not simulationist, and nor is d20 stop-motion combat. You need to look at systems that aim for simultaneous and/or (near-)continuous resolution, like some of those I mentioned above.

The system you are describing is one which should allow re-tries for such a check. Many games of D&D include both that narration, and a mechanic for making multiple attempts (often along the line of a critical failure causing a jam or break which prevents subsequent attempts). The system which I describe is one which explains why additional checks cannot be made. If a system doesn't allow multiple attempts, but the GM is explaining the attempt in such a way that this would not logically follow, then that's a problem with the GM.

You are describing some possibilities, but not all of them. Classic D&D does not permit retries on various thief skills, but there is no implication that this is because of a lack of knowledge on the part of the thief. (Similarly, there are no retries for detecting traps or secret doors, no matter how hard the players say their PCs are looking.)

That is because, in these respect, classic D&D is not using it's skill checks solely to model ingame effort. It is also using them to ration various outcomes - opening of doors, disabling of traps, learning backstory, etc.

The point becomes even clearer if we focus on an instance of "no retries" from AD&D that is not a thief skill and not about cognition at all: the reason why only one attempt ever is permitted to force a magically held door, or to bend bars or lift gates, isn't because failure indicates a lack of knowledge! Failure just indicates that the PC is not up to the task.

In all these cases, the player can describe his/her PC as continuing to try, but the GM doesn't have to grant another roll - the action declaration has already been resolved. The GM can narrate the failure of retries pretty simply: "The door still doesn't budge", "You still don't find anything," "You still can't open the lock," etc. It's understood that, in such cases, the PC has simply failed and can't succeed until the relevant condition is met (eg gaining a level).

There are some actions where retries are expressly permitted: opening (non-magically held or locked) doors by brute strength, and listening at doors, are the two main ones I can think of. In these cases, the only resource cost is time (and hence wandering monster rolls).

From what I've gathered, based on his quotes, he [ie Gygax] was a hard-core Gamist who looked down upon role-players as wannabe play-actors. In fact, I wouldn't be surprised if his use of the term was meant more in the 4E sense, as the role a character fills within the party.

It's hard to reconcile that style of gameplay with any modern definition of "role-playing".

I don't find it hard at all.

When my students engage in roleplay activities - say, mooting, or mediation exercises - they are not trying to achieve some particular inner state of "immersion". They are pretending to occupy a certain professional role - barrister, mediator - and they are practising their exercise of the skills that are germane to that role, while also accepting the constraints upon action that come with occupying that role (eg appropriate court manner). The point is to develop skills.

Likewise in Gygaxian D&D, where occupying a role - defined, as per my quotes upthread, primarily by class - both confers a permission, in the game, to make certain moves, and also imposes certain limits on permissible moves. Player skill consists in identifying those moves, becoming familiar with them, deploying them cleverly, while avoiding or negating the adverse consequences of the limits.

In both the professional roleplays that students undertake, and in a RPG of the sort Gygax invented, certain experiences - eg immersion - may be byproducts or side-effects, but they are not the aim of the activity.

It is worth noting that even though "roleplaying" in this sense is no synonymous with "playing a role" in the thespian sense. An actor is not trying to identify and deploy certain moves made possible by the role s/he pretends to instantiate. It is probably also worth noting that even an actor's main aim it not to immerse, but to give a performance. Some schools of acting use a type of immersion in the role as a means to that end, but they don't achieve the immersion by playing the role - rather, they cultivate the immersion as a precursor to playing the role. I don't think that an actor who gave no conscious thought to how the fiction should be performed, and who played his/her role purely in response to in-fiction cues generated by the other performers - that is to say, who approached the task within the constraints that you say are exhaustive of roleplaying -would be doing his/her job.

 

[Capn Charlie]

The solution we found at the table is letting players collaborate on skill checks. Any players proficient in a skill contributes a d20 to the player leading the check, usually the one with the highest bonus, who rolls them all and keeps the highest. Players with no proficiency can attempt a check, but cannot collaborate.

 

[delericho]

delericho said that he (? I think)

Yep.

If I don't have to roll to learn whether or not my PC was taught cobbling at school (but instead just pick the cobbler background, or cobbler skill, or whatever), why do I have to roll to learn whether or not my PC was taught about such-and-such a secret cult?

It's about the focus of the game. If I were running a game with frequent interactions with an evil society of cobblers, then I would make sure to use a "Cobbler" skill in exactly the same way as I use the "Religion" skill currently. But, of course, that would be a pretty unusual game! For a more 'standard' game, issues of whether the character knows shoes are likely peripheral (if they come up at all), while evil cults are much more common.

Is it about fairness? Eg it would be unfair for a player to declare that his/her PC knows every secret about every cult, but equally it would be unfair for his/her PC to know nothing about any secret cult, so the information is rationed by dice rolls?

Ideally, I would prefer something a bit more granular than 5e's binary Proficiency system - at least three different levels of "I know about Religion". So the player gets to choose to just what extend his PC knows about such things, and gets to choose between a ~50%, ~70% and ~90% chance of knowing (with the trade-off of lower skills elsewhere).

But how is it better to force a roll that might make them make a choice based on bad data? If making choices based on bad data is bad, how is it less bad if it happens sometime based on rolls, rather than always based on GM mandate?

One of my self-imposed rules is that I don't mislead players and I don't deny them information they 'should' have (whether because they looked for the secret door in exactly the right place, or because they rolled high enough on their Perception check). Likewise, although an individual NPC may lie (or otherwise deceive) them, when acting in my role as impartial DM I won't do so.

So, no, I won't choose to give the PCs partial information. If they're getting partial information it's because the dice have declared such.

As for whether that's better, I can't really say. I prefer to play that way, and I've had some bad experiences with GMs who do things the other way, but I'd certainly file it under YMMV.

For instance, in the combat case the GM can't stop the TPK (assuming no Deus Ex Machina) other than by suspending the action resolution rules (eg fudging die rolls, or overtly declaring that the NPCs miss rather than hit).

The DM couldn't decide that the bad guys decide to take the PCs prisoner instead of killing them outright?

Whereas in the case of the religious icon, the GM can just declare that "Yes, your PC did learn about such-and-such a secret cult when studying at Obscure Knowledge College."

Thus overriding the action resolution rules I'm using.

The analogue to the TPK in the investigation scenario would be the players not declaring that their PCs search the room, or not declaring that their PCs look at the icon. But once they have searched the room, and have declared that they examine the icon, the players have made all the action declarations that they can - it all turns on what backstory is authored about their PCs (what did they learn at college) and hence what backstory the players become privy to.

The emphasis is mine, because there's possibly a key distinction I make here: it's not about what they learned in college, it's about what they recall at this time. After all, if they fail the roll then they have the option to go consult with their old lecturer, revisit their notes, head to the library, or otherwise run down the details.

If the visit to the NPC has a cost, then that becomes like the "stakes" examples we discussed upthread - the reason for the knowledge check is to find out what resources, if any, the players have to forfeit in order to get the information.

If there is no cost other than time spent at the table in visiting the NPC then again the rationale of the whole structure becomes more opaque to me. Couldn't we just short-cut straight to the information - either by declaring that the PC knows it, or describing how the PC visits his/her friendly old mentor who tells him/her all about the nefarious cult of so-and-so?

Because when the PC fails the roll there are many options for what they do next, each of which potentially has a different cost and consequences. If they choose the sage, that probably costs money. If they instead revisit an old lecturer, well oops, it turns out Professor Teabing was the BBEG all along. If they consult their underworld contacts they now owe a favour. And so on. Or if they decide not to chase it down at all, they have to work without that information - but that may be fine too, since there will be other clues out there.

(And note that some or all of these chosen options may very well be dealt with quickly - for example "consult the sage" may be a matter of "he charges 500gp. Once you've paid he tells you..." Others may be more involved if, for example, the evil cult has infiltrated the library and you have to deal with their goons.)

But if the PC makes the roll, they skip all of that - the information is a freebie.

That helps with the aesthetic, yes. I think it raises some other questions, though, or at least leaves them unanswered, like what is "the story being told", and who is telling it?

Ultimately, it's the story of "what the PCs do", and it's being told collectively by the group. It's likely that that story will turn out to be "investigate the cult" and/or "stopping the cult", but that's by no means guaranteed.

I agree that PC build choices, including acquiring Knowledge skills for a PC, have a cost. The question is, then, what is the player buying for that cost?

A better, but by no means perfect, chance to access information. Just as the sword-master has a better chance to hit with his sword but not 100%.

In both 4e and BW, knowledge skills can also be used to gain access to the GM's secretly-authored backstory, but that is not their sole use. When used in this way, the main benefit the player gets is to force the GM to declare, and thereby "lock in", the backstory.

It's worth noting that almost* all of my backstory is locked in from the outset. A knowledge roll causes me to pull back the curtain, but the picture behind it is unchanged.

* "Almost all" because I'm not going to pretend to do all of this with absolute purity. I've found that rigidly sticking to just about any playstyle leads to disaster.

In the style of play which emphasises "collecting potentially useful clues", what is the benefit to the player in collecting those clues? How do they relate to the adventure - and what exactly is the adventure (or, in delericho's terminology, the story being told?

Hopefully, it changes it from "the story of how we failed to defeat the cult" into "the story of how we defeated the cult". Both can be enjoyable stories, but let's be honest - people are going to enjoy a story where they are the all-conquering hero rather more than the one in which they all-too-often fail.

(But, at the same time, although they'd prefer to win, they'll also feel much more satisfaction if they solve the puzzle if it's difficult than they will if the DM just gives them the answer. Personally, I would rather lose fairly than win because the DM declared it so. YMMV, of course... and of course there are a lot of variations in difficulty level between "murderously hard" and "trivially easy" - finding the right balance, now that's the trick. )

If a group of players turns up with PCs completely lacking knowledge skills and divination magic, what will they not be able to do, in terms of engaging the game and doing well at it, that they could do if their PCs did have those abilities?

Ah, now that's about campaign design. If I'm going to be running an investigation-heavy game (which, incidentally, I'm about to be), I'll explain this to the players ahead of time, and encourage them to build PCs with appropriate skills. If they produce a party that's utterly unsuited to the campaign then that's probably an indication that they're not really interested in the game I was envisaging, so we'll do something else instead.

But there may come some cases where a PC, or indeed the party, simply don't have the skills to solve a given puzzle in the 'standard' way - to use a 3e example, perhaps nobody is trained in the relevant Knowledge skill and you can't roll untrained (5e obviously doesn't have this feature). What that means is that the PCs have to find some other way to solve the problem (sage, library, underworld contacts; or this clue is useless so they need to find some others).

That last case is obviously not ideal, but provided it's just one part out of several it's not a huge issue. It becomes an issue if too many of the paths through the adventure get closed off because the PCs lack the tools to interact with them. (And I've been lucky enough that this has only ever been a theoretical concern - the PCs have always had enough skills, and made enough of the relevant rolls, that they've been able to do what's required. But, as I said, maybe that is just luck.)

 

[Ristamar]

Ideally, I would prefer something a bit more granular than 5e's binary Proficiency system - at least three different levels of "I know about Religion". So the player gets to choose to just what extend his PC knows about such things, and gets to choose between a ~50%, ~70% and ~90% chance of knowing (with the trade-off of lower skills elsewhere).

Technically, there are three tiers (Unskilled, Proficient, Expert), but third tier access is rather limited.