Roleplaying Uncertainty

[Glorelendil]

One thing I’ve never been satisfied with in RPGs are using skill rolls to resolve things that in the real world would generally result in uncertainty, even on a success. Is the chest trapped? Is the captive lying? Which passage is the right one? Are the mushrooms poisonous? Should I cut the red wire or the blue wire?

The problem with resolving these things with a simple skill roll is that it removes the uncertainty. You beat the target number, therefore you succeed. The LM will often say, “You think that…” but really you don’t think, you know. You can see the dice, and they say you know.

Or maybe the dice say the opposite: you don’t know. So now the next player, perhaps with lower skill, says, “I’ll try.” On the one hand this is perfectly fair. If you’re trying to determine if the captive is lying or the chest is trapped, why not get second and third opinions? On the other hand, this kind of breaks the point of the mechanic, in a statistical sense.

Furthermore, failure can’t provide a false positive, because again everybody can see the dice. Imagine the askance reaction if somebody utterly fails an Insight roll and the LM says, “You think he’s lying.”

I’ll digress here for a moment to say that I think this problem gets at the heart of what roleplaying and immersion are all about: inhabiting your character in the sense of experiencing what your character experiences, not merely acting a part for the benefit of others at the table. If my character is quite sure that it’s the blue wire, but in reality not entirely 100.0% sure, then as the player I want to have the same nagging doubt. I want to be worried I’m wrong. I want to be in suspense.

The traditional model of "success means success" leaves no room for suspense.

Many of these questions are not about knowledge but about clues. For example, nobody can know for certain if somebody else is lying (that’s what “Detect Lie” spells are for). At best we non-Wizards might use Insight to notice body language and facial expression and verbal clues, but we still have to interpret them.

So one option is for the LM to describe the clues, but that’s both hard to improvise consistently, and honestly it’s also nearly impossible to deliver ambiguously. "Umm...you notice his eye twitching." For chrissake you just beat the TN by 6 so you know perfectly well that "eye twitching" means he's lying.

Another option is for the LM to roll secretly. I’m opposed to LMs (and DMs and GMs) making rolls for players on principal...it feels like loss of agency...but this solution by itself also doesn’t provide enough nuance: it allows for players to be uncertain why they don't find anything, but not for them to be uncertain about the reliability of their answers when they do. "You don't think he's lying" could either mean he's not lying, or that you failed the roll. But "you think he's lying" can only mean you succeeded, unless the GM is arbitrarily interpreting the results.

Essentially what I want to do is make that interpretation less arbitrary. I’d like to propose another kind of secret roll: an additional roll made after the player succeeds. This approach is based on the premise that success means only that you notice a clue, not that you know with certainty how to interpret it.

After a success, the LM rolls the additional secret die, and that result determines whether the correct or incorrect result is given, with the threshold changing based on the degree of success. For example, let’s use a single d6: On a regular success if the LM secretly rolls 4+ (50%) then the player gets the correct information. On a great success it’s 3+ (66%), and on an extraordinary success it’s 2+ (83%). If this secret roll fails then the player gets the opposite information. “Yeah, you think he’s lying.”

(Note for AiME: this system works best with degrees of success, but with d20 you could treat each 5 points about the DC as another success. So 23 vs. DC 15 would be 2 successes.)

We can easily tweak the odds by using different dice schemes. It could be roll 1d6 for each success, and all 1’s mean false positive (so 83%, 97%, and 99.5%). Or we could use a Feat die, with thresholds of 2, 1, and Eye of Sauron. I would suggest that these odds be based, though, not on the difficulty of the test...that’s what the TN of the original roll is for...but on the consequences for the heroes. Will acting on the wrong information result in a TPK, or just inconvenience/embarrassment?

Note that the other players now have an incentive to participate in the action, instead of watching the guy with the highest skill go, and only jumping in if he fails. And the participation of more players in turn lets us set the TNs higher for what should be difficult tasks. “Ok, three of you failed, two ordinary successes think he’s telling the truth, but the extraordinary success thinks he’s lying. What are you going to do?”

I’ve been thinking about this problem as I continue to struggle through my ideas for underground travel, for which I believe this kind of uncertainty is essential for creating the appropriate atmosphere. As I considered the solution it occurred to me that it also was applicable to lie detection, trap detection, and a variety of other common scenarios.

The one thing this solution doesn’t provide for is the player failing the roll but thinking he succeeds. However, I like to think of failure as “doesn’t have anything to contribute” and success as “noticed a clue”: the secondary, secret roll simulates whether his interpretation of the clue is correct.

[KathD]

I've seen a nice solution for scenarios where the players making a choice is key, and where the GM doesn't know the outcome in advance.

For example, imagine a chest which may or may not be trapped. Whether or not the chest is trapped is determined by rolling a d6; high = trapped, low = not trapped. Before deciding to open the chest, the player makes a Search roll. If the player succeeds, they get to roll the 'trap die' before deciding to open the chest. Otherwise, they roll the trap die after making the decision. The player's success or failure has no effect on whether or not the chest is trapped, and so they might decide to open it even if they failed on the roll.

The precise situation, probability of the bad thing happening, and TN for the player roll can all be adjusted, but this is still limited to situations where the GM is happy for for a random die roll to determine the in-character reality. So probably not that good for social situations, but I could see it working for a variety of physical hazards.

[Glorelendil]

I've seen a nice solution for scenarios where the players making a choice is key, and where the GM doesn't know the outcome in advance.

For example, imagine a chest which may or may not be trapped. Whether or not the chest is trapped is determined by rolling a d6; high = trapped, low = not trapped. Before deciding to open the chest, the player makes a Search roll. If the player succeeds, they get to roll the 'trap die' before deciding to open the chest. Otherwise, they roll the trap die after making the decision. The player's success or failure has no effect on whether or not the chest is trapped, and so they might decide to open it even if they failed on the roll.

The precise situation, probability of the bad thing happening, and TN for the player roll can all be adjusted, but this is still limited to situations where the GM is happy for for a random die roll to determine the in-character reality. So probably not that good for social situations, but I could see it working for a variety of physical hazards.

(Thanks for actually reading my wall of text!)

I've considered the mechanic you are describing and in general I like it because when the LM makes these sorts of choices it can form a pattern that players pick up on. (It reminds me of how writing algorithms to play rock-scissors-paper is a great exercise for computer science students because eventually they realize that it's not possible to 'randomly' pick your next move.)

What the trap-roll doesn't do, however, is leave any doubt after a success Search. If the player succeeds, and then rolls the trap die, he/she knows with 100% certainty whether or not there is a trap before attempting to open the chest. There's no suspense left. But I think that "succeeding" should still leave room for doubt.

I've been thinking more about this since my middle-of-the-night-after-feeding-baby post, and I think I want to backtrack on my suggestion that "success" means a clue is found. The problem with that is that if everybody in the party finds clues they could just pool their knowledge and arrive at a consensus.

I think instead that the common theme across these scenarios is that they rely on some form of intuition or gut feel. What success indicates is that the character has such an intuition; the reason for it is a matter of after-the-roll narration. Finding a clue, albeit one with ambiguous meaning, could be one such narration.

Also, using 3-2-1 from a d6 isn't the right odds, for the simple reason that even a basic success should give a better chance than 50/50, which is the same as just guessing. Perhaps 2-1-Eye of Sauron on a Feat die is the right choice. That's a 3/4 chance of being right on a success, 5/6 on a great success, and 11/12 on an extraordinary success. That leaves enough wiggle room that even on an extraordinary success the heroes will have some doubt. And a +1 or -1 modifier would shift the odds better/worse in a nice way.

To give some examples (in my underground journey rules) that I'm trying to simulate:

1) The players come across some mushrooms. Are they edible? If they eat them and they are good it will give them much-needed sustenance (by increasing their "Supplies" rating). If they leave them, they risk running out of supplies. But if they eat them and they are bad they risk becoming Poisoned.

With a strict binary skill role (Lore? Hunting? Healing? Or auto-success with Herb-lore?) it's really not a very interesting choice. If one character fails, another one tries. And even if somebody succeeds it suddenly becomes uninteresting: "Ok, I passed a skill test and now we know. They are good/bad." The players rolled some dice, but nobody made a hard decision. I want the skill check to help make the decision, not make it.

2) The players come across a 3 way junction, and they don't which way to go. A Riddle check can provide guidance. The wrong choice might become obvious (e.g. a dead-end) but it might just lead to rougher terrain or a longer route or a danger. To really simulate the feeling of being there, ideally the heroes would never know if it was the right choice. "Man, this is tough...we are in trouble." "I wonder if we made the right choice back there?" "No way to know..."

If the heroes just roll a great success and the LM says, "You want the left passage" it is, again, an example of the dice making the decision, not the players.

[Deadmanwalking]

Another option is for the LM to roll secretly. I’m opposed to LMs (and DMs and GMs) making rolls for players on principal...it feels like loss of agency...but this solution by itself also doesn’t provide enough nuance: it allows for players to be uncertain why they don't find anything, but not for them to be uncertain about the reliability of their answers when they do. "You don't think he's lying" could either mean he's not lying, or that you failed the roll. But "you think he's lying" can only mean you succeeded, unless the GM is arbitrarily interpreting the results.

Huh? This doesn't follow logically at all. I've played in and run games where failing a roll can give you a false answer, thus 'He's lying.' might easily mean you failed the roll and he's telling the truth.

However, that solution, or the one you suggest, seem to me extremely ill-suited to The One Ring specifically for the good and simple reason that it kinda screws over the basic principles of the way Hope works. I mean, either way you'll need to know whether it matters (which messes with the lack of transparency) and inj your suggested variation even if you do spend it and it makes your roll a 'success' that only gives you a 50% chance of it actually mattering (in most cases). It also makes PCs vastly less effective unless you drop TNs a lot. I mean, assuming you have an 75% chance of success, and assuming you roll a 6...you only get the right answer 50% of the time, total. That's...a rather large problem, really.

I do agree that this is a potential problem, mind you, I just think this solution results in a variety of unfortunate things. I'm trying to think of something more workable, but it's a tricky issue to solve in TOR specifically (in other games, IME, just not telling PCs the TN for things and sometimes rolling for them works pretty well, but that really doesn't work with Hope and its use in TOR).

[Glorelendil]

Huh? This doesn't follow logically at all.

Not trying to be a jerk here, but you really have a knack for coming across as a jerk.

But regardless...

I've played in and run games where failing a roll can give you a false answer, thus 'He's lying.' might easily mean you failed the roll and he's telling the truth.

Yes, that's what I mean by arbitrarily interpreting results. There's nothing in the mechanics to differentiate between "you can't tell" and "you interpret wrong". And let's say the roll is a great roll; I think the players should have that information so that they trust their intuition more. But if you have to keep the actual roll secret, you can't let them know.

In any event, I'm still opposed to DM rolling for the players. Throwing the dice is a big part of the fun.

However, that solution, or the one you suggest, seem to me extremely ill-suited to The One Ring specifically for the good and simple reason that it kinda screws over the basic principles of the way Hope works. I mean, either way you'll need to know whether it matters (which messes with the lack of transparency) and inj your suggested variation even if you do spend it and it makes your roll a 'success' that only gives you a 50% chance of it actually mattering (in most cases). It also makes PCs vastly less effective unless you drop TNs a lot. I mean, assuming you have an 75% chance of success, and assuming you roll a 6...you only get the right answer 50% of the time, total. That's...a rather large problem, really.

One could equally argue that spending Hope on an attack against an enemy with Great Size should do more than cause a measly 5 endurance loss.

As noted in my second post, 50% is the same as guessing randomly and therefore is meaningless; thus the Feat die is a better option than a Success die. Also, spending Hope could potentially turn a failure into a Great or Extraordinary Success, in which case the benefit would be much more. Finally, players have the option of spending Hope: they don't have to do so if they don't think the benefit is great enough.

And, yes, TNs should be adjusted under my proposal. If only a single player can roll then the TN should in general be lower, as even a success doesn't guarantee success, as it were. And if the whole party is going to roll then the TNs can be much higher.

I do agree that this is a potential problem, mind you, I just think this solution results in a variety of unfortunate things. I'm trying to think of something more workable, but it's a tricky issue to solve in TOR specifically (in other games, IME, just not telling PCs the TN for things and sometimes rolling for them works pretty well, but that really doesn't work with Hope and its use in TOR).

[Deadmanwalking]

Not trying to be a jerk here, but you really have a knack for coming across as a jerk.

But regardless...

I'm sorry if I came across that way. That wasn't my intent at all.

Though, for the record, I've seldom gotten similar comments from people other than you (not only on this forum, but more generally). Which is seriously weird since you, likewise, don't seem to have this kind of issue with other people. Maybe the two of us just have really incompatible communication styles or something?

Yes, that's what I mean by arbitrarily interpreting results. There's nothing in the mechanics to differentiate between "you can't tell" and "you interpret wrong". And let's say the roll is a great roll; I think the players should have that information so that they trust their intuition more. But if you have to keep the actual roll secret, you can't let them know.

Sure, but we're talking House Rules, you can fix this if you like in several ways.

In any event, I'm still opposed to DM rolling for the players. Throwing the dice is a big part of the fun.

True in some cases. Though in fairness, this is only even useful on a small subset of rolls. I mean, doing it on combat rolls or others with obvious effects is obviously unnecessary. Stuff that involves noticing things or figuring out things that can't be immediately checked is only a small portion of the rolls made in the vast majority of games.

But yeah, rolling yourself is more fun, and the whole thing is a bad idea in TOR anyway due to, as I mentioned, the way Hope works.

One could equally argue that spending Hope on an attack against an enemy with Great Size should do more than cause a measly 5 endurance loss.

Well, in that case you still got something. If you're making it so people don't know on a failure, but get either right or wrong info on a success (which is what you seem to be suggesting based on the above), that means that sometimes you're getting something worse than nothing at all with a Hope expenditure. That's more like the enemy healing sometimes when you hit them.

As noted in my second post, 50% is the same as guessing randomly and therefore is meaningless; thus the Feat die is a better option than a Success die. Also, spending Hope could potentially turn a failure into a Great or Extraordinary Success, in which case the benefit would be much more. Finally, players have the option of spending Hope: they don't have to do so if they don't think the benefit is great enough.

Sure, but I don't think disincentivizing spending Hope because 1/4 or 1/6 of the time it makes the outcome actively worse is the best way to modify the existing game system.

And, yes, TNs should be adjusted under my proposal. If only a single player can roll then the TN should in general be lower, as even a success doesn't guarantee success, as it were. And if the whole party is going to roll then the TNs can be much higher.

I'm not sure I quite follow on the second part of that. The whole party can already roll for most things that fall under this category, and thus be quite sure of success. Raising the TN adds uncertainty, sure, but it adds even more under this rule, which seems a bit unfair to the players in many ways. Besides which, something that's TN 12 when you're alone suddenly being TN 16+ when there are several people trying is extremely weird and immersion breaking.

I'm still trying to figure out a way to arrange this sort of thing that doesn't devalue success...maybe go with something involving Eyes rolled? For example, the LM being able to give the PCs one wrong answer on a successful roll instead of a Revelation Episode. Or just reduce the Hunt number by one per false answer. That version's not random, though, which might be a mistake. I'll keep thinking.

Some reward (probably at the end of the session so the players don't know what it was tied to) when the PCs lose a correct answer due to this does seem warranted, though. That really expands the design space without making the PCs feel as helpless or making things harder unnecessarily.

[Glorelendil]

Not trying to be a jerk here, but you really have a knack for coming across as a jerk.

But regardless...

I'm sorry if I came across that way. That wasn't my intent at all.

Though, for the record, I've seldom gotten similar comments from people other than you (not only on this forum, but more generally). Which is seriously weird since you, likewise, don't seem to have this kind of issue with other people. Maybe the two of us just have really incompatible communication styles or something?

Based on history, I have a habit of calling out what other people notice but let slide. Where others overlook poor manners in the interest of civility, I'm willing to cause awkwardness at the banquet table. And, no, you're not the only one here who does this, nor the only one I've called out for it. It's just been a while since I've gone to the mats over this; possibly since before you joined.

In any event, if it doesn't occur to you that telling somebody their idea "doesn't follow logically at all" might be taken as dismissive/disrespectful then there might not be grounds for discussion.

Sure, but we're talking House Rules, you can fix this if you like in several ways.

I don't think the paradox can be solved in a single dice roll; I think it requires one public and one private roll to cover all bases.

But yeah, rolling yourself is more fun, and the whole thing is a bad idea in TOR anyway due to, as I mentioned, the way Hope works.

/sigh

Sure, but I don't think disincentivizing spending Hope because 1/4 or 1/6 of the time it makes the outcome actively worse is the best way to modify the existing game system.

As long as it's understood that the result is itself a probability...no different from spending a point of Hope to get a Pierce, with no guarantee that a Wound will result...then the players can make the choice whether to spend the point. It's only if the mechanics are misunderstood that there's a problem, i.e. that the outcome is "actively worse".

I'm not sure I quite follow on the second part of that. The whole party can already roll for most things that fall under this category, and thus be quite sure of success.

Yes, and I think that breaks the story rather badly, far too often. How many times has something come up to find/understand/solve something, and one after another each players says, "I guess I'll try." What's the point of even requiring a Lore check if every player can take a shot at it until somebody gets lucky?

Raising the TN adds uncertainty, sure, but it adds even more under this rule, which seems a bit unfair to the players in many ways. Besides which, something that's TN 12 when you're alone suddenly being TN 16+ when there are several people trying is extremely weird and immersion breaking.

[/quote]

If you somehow ran an adventure in which the exact same challenge happened twice, once with one person and once with a group. In that one-in-a-thousand edge case, go ahead and keep the TNs the same. Otherwise adjust the TN dynamically to fit the situation.

[Deadmanwalking]

Based on history, I have a habit of calling out what other people notice but let slide. Where others overlook poor manners in the interest of civility, I'm willing to cause awkwardness at the banquet table. And, no, you're not the only one here who does this, nor the only one I've called out for it. It's just been a while since I've gone to the mats over this; possibly since before you joined.

This would be more likely as an explanation if I hadn't gotten a lot of comments on several forums regarding me being overly polite...

Still, that might come down to different standards of politeness and discussion on different forums, pr something like that.

In any event, if it doesn't occur to you that telling somebody their idea "doesn't follow logically at all" might be taken as dismissive/disrespectful then there might not be grounds for discussion.

Now that you've mentioned it, yeah, I can see where you might take offense, and am legitimately sorry for that. But in my previous experience, it's more usually taken as I intended it: A statement that I do not follow your argument and you might wish to rephrase (or possibly rethink).

I don't think the paradox can be solved in a single dice roll; I think it requires one public and one private roll to cover all bases.

Very possibly.

/sigh

Uh...that was a specific reference to the idea of the GM rolling, something I rather thought we were in agreement on in regards to TOR.

And, for the record, that "/sigh" comes off as every bit as dismissive and disrespectful as anything I said could. I wouldn't even mention it (as I try and assume people don't intend that sort of thing), but since we're discussing the nature of civility, it seems relevant.

As long as it's understood that the result is itself a probability...no different from spending a point of Hope to get a Pierce, with no guarantee that a Wound will result...then the players can make the choice whether to spend the point. It's only if the mechanics are misunderstood that there's a problem, i.e. that the outcome is "actively worse".

Well, barring a radical redesign of some sort, it makes Hope expenditures on such rolls worth less. That's not inherently bad, it's true, but it's something that probably needs to be accounted for. Hence my suggestion that PCs get some reward when this comes up.

Yes, and I think that breaks the story rather badly, far too often. How many times has something come up to find/understand/solve something, and one after another each players says, "I guess I'll try." What's the point of even requiring a Lore check if every player can take a shot at it until somebody gets lucky?

Well, for one thing, that tends to make good logical sense, and for another, in most groups, relatively few people have meaningful amounts of most skills used this way. I really just don't think it's as much of a problem as you seem to.

If you somehow ran an adventure in which the exact same challenge happened twice, once with one person and once with a group. In that one-in-a-thousand edge case, go ahead and keep the TNs the same. Otherwise adjust the TN dynamically to fit the situation.

Sure, but it's gonna seem weird to people whenever an equivalent situation has a much higher TN just because there are multiple people making it. It's also a huge breach in realism and thus suspension of disbelief for some people, which is less than ideal.

[Glorelendil]

My bad (in the sense of being quixotic) for even venturing off-topic.

Well, barring a radical redesign of some sort, it makes Hope expenditures on such rolls worth less. That's not inherently bad, it's true, but it's something that probably needs to be accounted for. Hence my suggestion that PCs get some reward when this comes up.

Well, if the secret roll is using the Feat die then it could get a +1 modifier when Hope is used, which would shift the odds to 5/6, 11/12, 1/1. I don't think it's necessary, but that solution does have the benefit of creating an incentive to use Hope even on a successful roll.

Well, for one thing, that tends to make good logical sense, and for another, in most groups, relatively few people have meaningful amounts of most skills used this way. I really just don't think it's as much of a problem as you seem to.

Our mileage varies, then. I see this come up all the time: one at a time, working from high skill to low, everybody tries "Search" or "Lore" or "Craft" or whatever it was until somebody scores a success. (Note that my proposed system wouldn't address all these cases, for example it wouldn't apply to a Craft test to pick a lock.)

Sure, but it's gonna seem weird to people whenever an equivalent situation has a much higher TN just because there are multiple people making it. It's also a huge breach in realism and thus suspension of disbelief for some people, which is less than ideal.

I'm not seeing it. TNs vary all the time. And this would apply to a fairly narrow sub-set of skill tests, when the result of a success is to think you know something, without being sure. Such as the examples I gave at the top of the page. In the vast majority of cases normal rules would apply.

Anyway, using the same TNs whether or not it's a group is completely optional: the approach still works better than the default rules (where I'm defining "better" to mean "with an element of uncertainty that will heighten narrative suspense".)

[cuthalion]

I think this is essentially a good and interesting addition. And I like the applications--to journeys (underground, unmapped) and encounters. Maybe there are more.

Couple of thoughts:

With a strict binary skill role (Lore? Hunting? Healing? Or auto-success with Herb-lore?) it's really not a very interesting choice. If one character fails, another one tries. And even if somebody succeeds it suddenly becomes uninteresting: "Ok, I passed a skill test and now we know. They are good/bad." The players rolled some dice, but nobody made a hard decision. I want the skill check to help make the decision, not make it.

One big issue here--not really sure this is the way the rules work? Normally repeated tries are not allowed. Forget in the rules where it says this, but it's pretty explicit. So, I'm not necessarily disagreeing with the theme of what you're proposing, but I'm not sure you're characterizing the current system accurately.

I think, RAW, if someone fails a roll, the characters are left not knowing about the mushrooms (nothing is there to stop them from trying them anyway). If someone succeeds, they might know various levels of information about the mushrooms, thanks to degrees of success. An ordinary success might let players know they aren't poisonous, but a great success might also suggest that the character knows proper preparation is key to avoiding stomach aches all round.

Additionally, if somebody fails, that's it, there shouldn't be more attempts, except for specific circumstances. The players will be left, in fact, not knowing.

If the heroes just roll a great success and the LM says, "You want the left passage" it is, again, an example of the dice making the decision, not the players.

And this, in fact, is the system as designed. Per RAW, the dice emulate everything that goes into making that decision, the circumstances, the character's life/background, etc. We are consulting the oracle, as it were, to see how the character acts/reacts. We're not actually needed, as players, to make decisions at that level for our characters.

I know you know all this, but thought it bore reiterating. I still think what you're suggesting is smart.

If I understand you right, you're saying that for circumstances where uncertainty is to be emphasized, we can use this new mechanic to zoom in on and simulate that uncertainty in more detail, much like we do for journeys or combat, and there I'm totally with you.

I like the mechanic you propose, I'm not totally sure which die/combo of dice is best, as stats aren't my strong suit. Personally, I don't much like rolling a feat die and interpreting a chart. I think I like the use of success die more. Rolling one per level of success seems most in line with the theme/existing rules for me, and it's a novel mechanic, which doesn't hurt. The odds there look a bit too tight tho--how do you make a extraordinary success much more reliable, as compared to an ordinary?

Lastly, the only other thought I have is on the multiple participants. Again, I wouldn't say multiple people participating raising the TN as you describe it fits the RAW. My suggestion would be to have the effects of multiple participants adjust the Uncertainty die/ce, not the action roll. That can still produce the kind of group uncertainty/conflicts you're going for, and also then eliminates Deadmanwalking's issue with the TNs. Not sure if it gets around the hope spending issue.

Thanks for sharing your thoughts! Would be interested to see some played out examples of this working out. I might try some myself if something appropriate comes up.

*P.s. One last thought--maybe traits could be invoked when rolling Uncertainy die to somehow affect the results. This would seem to make thematic/logical sense, that a character with a relevant trait is less likely to choose the wrong option.