Quote from: Silverlion;363502I like failure, but I like it to be much as Monte Cook suggests--get worse before it get better. The trick is not to focus too much on "worse" or "better" but allow both to occur and NOT TOTALLY screw the the players either way.

Yes "better" for PC's, can hamper the players long term fun.

Yeah, I definitely think that there's two types of failure: constructive failure and stonewalling failure. Constructive failure results in consequences which inspire further actions - so even though the PCs didn't get what they want, they still have something to do. That's what Cook is talking about in his post. Stonewalling failure is what happens when you fail a roll... and nothing happens. The PCs don't make any progress, but at the same time there aren't any real consequences of their failure beyond that. The status quo is maintained, the situation doesn't evolve. The PCs end up precisely where they were before they failed the roll, except they're now mildly frustrated.

To be honest, I think occasional stonewalling failure is inevitable. Even the greatest of improvisational GMs can't dream up fabulous consequences whenever someone fails a spot hidden roll. But when it comes to crucial rolls - the ones which could well define which direction the campaign is going to in next - then I reckon GMs should strive for constructive failure. "You didn't beat the wizard - guess the campaign's wrecked" is the response of a lazy and unimaginative GM who isn't willing to do the improv or put in the work to come up with what happens next. You want to be saying "You didn't beat the wizard, and now all *this* has happened - *now* what are you gonna do?"

Oh, and as far as fumbles go - if a GM uses such things to make PCs look like bumbling morons, then they've rolled a fumble on their GMing roll.

When I run games with fumbles (generally with groups that enjoy such things), I try to make sure that fumbles are due to factors external to the PCs. It's not that your character's too dumb to read the ancient manuscript, it's just that the handwriting is horrible and the scroll is badly preserved. Your 1 on that Smooth-Talking roll doesn't mean you just walked up to the Princess and called her a stuck-up bitch, it just means that she's having a really, really bad day and isn't inclined to listen to anyone at the moment. Your botch on that combat roll doesn't mean you planted your sword in your skull, it just means you've been shoved over by your opponent and have his sword at your throat.

In general, I don't regard dice roll as being a measure of how competent a PC happens to be at that particular moment - that's what stats and skills measure. A dice roll represents the external factors beyond the PC's control - whether dumb luck or enemy action is going to step in and cause a problem for the PC or not. Of course, the more skilled a PC is, the more likely they are to rise above the external factors and succeed more often. But failure doesn't mean a PC is incompetent, they just aren't quite competent enough to counteract the external factors.

Quote from: Warthur;364288Oh, and as far as fumbles go - if a GM uses such things to make PCs look like bumbling morons, then they've rolled a fumble on their GMing roll.

When I run games with fumbles (generally with groups that enjoy such things), I try to make sure that fumbles are due to factors external to the PCs. It's not that your character's too dumb to read the ancient manuscript, it's just that the handwriting is horrible and the scroll is badly preserved. Your 1 on that Smooth-Talking roll doesn't mean you just walked up to the Princess and called her a stuck-up bitch, it just means that she's having a really, really bad day and isn't inclined to listen to anyone at the moment. Your botch on that combat roll doesn't mean you planted your sword in your skull, it just means you've been shoved over by your opponent and have his sword at your throat.

In general, I don't regard dice roll as being a measure of how competent a PC happens to be at that particular moment - that's what stats and skills measure. A dice roll represents the external factors beyond the PC's control - whether dumb luck or enemy action is going to step in and cause a problem for the PC or not. Of course, the more skilled a PC is, the more likely they are to rise above the external factors and succeed more often. But failure doesn't mean a PC is incompetent, they just aren't quite competent enough to counteract the external factors.

Bottle this and sell it!  Good post.

Quote from: Peregrin;363739If your GM isn't giving out conditional bonuses like candy, they fucking suck.  It's pretty easy to get a +20% on nearly any roll if you play smart and your GM knows when to apply bonuses.

It's still not my favorite system, but you cannot play it without taking conditional modifiers into account.  It won't work.

I love DH, but I do agree that you need to milk conditional modifiers for all they're worth. In DH, if you go into a "fair fight", you'll likely lose. The trick is to be cunning and make sure it's not a fair fight.

Hell, I don't know any game where I wouldn't want to encourage people to go for circumstantial modifiers on a regular basis. (Amber, as far as I can tell, only works at all because of 'em.)

Quote from: Imperator;364256You know, it quite doesn't work this way.

Mind explaining how it does work? Because "Gumshoe is the investigative game where the GM makes sure the players definitely get all the clues" seems to be a very common misconception. It's certainly the impression I had of the game; if I'm wrong about this, I would welcome being corrected.

Quote from: Warthur;364294Mind explaining how it does work? Because "Gumshoe is the investigative game where the GM makes sure the players definitely get all the clues" seems to be a very common misconception. It's certainly the impression I had of the game; if I'm wrong about this, I would welcome being corrected.

That is the impression I have of Gumshoe as well.

Quote from: One Horse Town;363657How about tedious repititious failure?

Hmm, depends on perspective I guess. Two examples of games I like but my gang doesn't because they view them as nothing-but-failure settings:

Big Picture - Cthulhu. Win the small battles, always lose the war.
Small Picture - Paranoia. Lose even when you win.

Seems like both settings had/have plenty of followers, including me. I just don't get to play 'em.

As far as how often to check for failure/success based on mechanics, usually only when it seems like something that'd make a big difference one way or the other. Fumbles during combat, for example, rarely result in disaster - at most, a jam or the weapon breaks. Searching for traps, now, is a different matter, since a single faliure could conceivably kill a PC.

Quote from: CRKrueger;364247Or...you could learn how to fucking GM.

Now that you've fucking learned how to GM, if you were running a CoC game and the PCs missed a vital clue, how would you rectify the situation?

Seanchai

Quote from: Seanchai;364322Now that you've fucking learned how to GM, if you were running a CoC game and the PCs missed a vital clue, how would you rectify the situation?

Seanchai

Full Disclosure/Disclaimer: I like Trail of Cthulhu, but at the same time, I see Krueger's point. I have no dog in the fight of CoC vs. ToC. Whatever rocks your boat.

Now, as to your question, it'll really depend on the sort of clue we're talking about. If the clue is SO vital as to make the whole adventure crash and burn if the PCs miss it, I would REALLY have to question the adventure's design in the first place. Reading through it the first time, I would come up with contingency scenarios in case the players miss it before the game ever happens. If the designers didn't include these contingencies, shame on them! Learn how to design a fucking RPG adventure, God dammit!

It seems to me that the big problem with highly skilled characters failing at stuff because of the dice-rolls is that many GMs, many players, even many game designers have a piss-poor grasp of very basic probabilities and shy away from even beginning to think about the probabalistic implications of their approach to task resolution.

My own game full-fat Omnifray makes the percentage chances of success very explicit and makes things fit together with a certain logc then comes under heavy fire (partly) because the percentages are there to be seen on a table. I'm experimenting with a d12-based system in Omnifray Lite now which closely mimics the full-fat Omnifray percentages but is hopefully more user friendly.

But it seems to me that a lot of people approach gaming on the basis that they intuitively like rolling dice a certain way, and their concerns are as simple as - do I understand which dice I have to roll, do I understand what I have to roll to succeed and can I do it without thinking too hard? As a GM do I have some easy-to-follow basic principles for what difficulty numbers to set?

These same gamers simply don't engage with the question of how likely it is that a certain result will come up, and whether that fits what they want for the game, whether it be in terms of making it a reasonable challenge, making it feel realistic/credible or making the game interesting in other ways.

But I think that you have to engage with that question (how likely is it that a certain result will come up) unless you want to face a very real risk that at least in certain contexts the probabilities will be so skewed, so different to what you would want for the game, that the game won't deliver what you want, perhaps because it interrupts your suspension of disbelief, perhaps because it's just too damned hard to make any headway or too easy to win constantly, perhaps because the distribution of results doesn't make the game interesting in the way you'd like, whatever.

So, I think game designers should be lauded for engaging directly with these questions, not criticised! OK this is turning into a bit of a rant.

Should a top surgeon be capable of botching a heart operation? Of course.

Should he have a 5% chance of doing so? No fricking way.

Should Legolas have a chance of missing an aimed bow-shot targetted at a fat, slothful orc some 20 yards away who is unaware of Legolas's presence and taking no particular action? Of course.

Should he have a 5% chance of doing so? No fricking way. Not if you want him to feel like an actual Legolas. But a 1 in 144 chance? (2d12, roll two 1s) - it's not statistically true to the character, but it would be so rare in play that I doubt it would break suspension of disbelief for anyone.

Easy solution - when you have a task which the PCs really should fail or really should succeed at, GM's discretion whether to bother calling for a roll - but the GM should handle this inconsistently, sometimes calling for a roll, sometimes not, depending on which way the wind blows.

Well that's my take anyway.

Quote from: Omnifray;364330It seems to me that the big problem with highly skilled characters failing at stuff because of the dice-rolls is that many GMs, many players, even many game designers have a piss-poor grasp of very basic probabilities and shy away from even beginning to think about the probabalistic implications of their approach to task resolution.

In practice, many DMs largely don't give a damn whether they actually understand probability or not.

A very useful tool for understanding how many dice rolls are required to achieve something, is the geometric distribution.  I've only met maybe 1 or 2 DMs over the years who understand its implications.

Quote from: ggroy;364348In practice, many DMs largely don't give a damn whether they actually understand probability or not.

A very useful tool for understanding how many dice rolls are required to achieve something, is the geometric distribution.  I've only met maybe 1 or 2 DMs over the years who understand its implications.

I agree, many DMs largely don't give a damn whether they actually understand probability - or even the specific probabilities that the game mechanics give rise to in critical situations. But it just seems to me that many of the problems gamers often tend to have with some or many game systems could be largely solved by paying a small amount of attention to those details. And if only those same DMs understood that and had some basic arithmetical competence and a bit of common sense, they would find their games went smoother, with fewer unexpected hitches in the mechanics, and they would find the whole exercise of statting up encounters etc. far less mystifying/intimidating. I find it frustrating that people back away from even looking at the obvious. Obviously, one man's meat is another's poison, and most people are less comfortable with numbers than I am - but I wonder whether many people don't realise how easy it can be to work this stuff out. You don't need even advanced high school statistics / A-level statistics to do it.

I don't think you need graphs or mathematical notation to do this.

I'm not sure I quite understand geometric distribution in mathematical terms as explained in its Wikipedia entry (I've got the basic gist of what the graphs illustrate though), and trying to use formulae and obscure/formal mathematical notation in the middle of a game would be a step too far for me. But if the game mechanics make probabilities explicit, it's a lot easier to figure out how things are likely to go than if the game mechanics use complicated dice pool systems with all sorts of funny permutations e.g. for exploding dice, how many rolls you can keep, etc. And to me that is one of the reasons why dice pools are so awful, the other being that they tend to limit the power variance that the game can comfortably accommodate.

Quote from: Benoist;364325Reading through it the first time, I would come up with contingency scenarios in case the players miss it before the game ever happens.

When I read the statement, I wondered how I'd handle it. I decided that I'd rework the clue in somewhere or create contingencies.

But then I thought, isn't that basically the same thing as handing the players a clue? In each case, you're changing the meta material of the game so that the players will get what they need.

Hence my question to CRKrueger. I wonder if he would come up with a solution, being someone who has fucking learned to GM, that was substantively different than what Trail of Cthulhu does...

Seanchai

Quote from: Seanchai;364367When I read the statement, I wondered how I'd handle it. I decided that I'd rework the clue in somewhere or create contingencies.

But then I thought, isn't that basically the same thing as handing the players a clue? In each case, you're changing the meta material of the game so that the players will get what they need.

Well, if you get down to the idea that the players HAVE to have the clue in their hands to conclude the adventure, then yes, you're basically handing it to them, with the caveat that what matters then is HOW they are getting access to it. If you provide different paths to get to the same point, basically, you've got to make sure that 1/ there are multiple ways to reach this point, and 2/ the players don't know that they "have" to reach that point, otherwise that wrecks the game - the choices won't matter to them anymore. What matters then is the manner in which they reach the adventure's bottleneck, i.e. the gameplay itself, not the bottleneck itself.

But really, again, I have something against bottlenecks in adventures in the first place. Whether you have that clue that the players HAVE to get, or that single door in the dungeon that the players HAVE to go through, it's the same thing - you're setting up a bottleneck in the game play, and if that part fails, then the whole adventure fails. That's a game play disaster waiting to happen, IMO.

Quote from: Omnifray;364364I'm not sure I quite understand geometric distribution in mathematical terms as explained in its Wikipedia entry (I've got the basic gist of what the graphs illustrate though), and trying to use formulae and obscure/formal mathematical notation in the middle of a game would be a step too far for me.

The geometric distribution is like a waiting time.  The expectation value of the geometric distribution gives you on average how many times something doesn't happen, until it actually happens.  For a probability p of something happening and 1-p probability of something not happening, the expectation value is 1/p.

For example, an obvious case would be how many attacks it takes to hit a monster given a probability p of hitting the monster.  In the case of requiring a minimum 19 or 20 (on a d20) to hit a monster (ie. 10% probability), it will take on average 1/0.10 = 10 attacks to hit such a monster once.  In the case of rolling a 16 or over (on a d20) to hit a monster (ie. 25% probability), it will take on average around 1/0.25 = 4 attacks to hit such a monster once.

To extend this, for a monster (with a 25% probability of being hit by a player) that dies after being hit 6 times, it will take on average around 4(6) = 24 attacks to eventually kill it.

I've used this framework in the past to estimate approximately how long a encounter will last, and how many hit points particular monsters should have such that they die after a particular number of hits.  I did this for my 3.5E games, for the reason that the encounter construction guidelines in the 3.5E DMG doesn't work particularly well.

This same framework can also be used to analyze other stuff, such as how often a bonus matters (ie. +2 to attacks, etc ...) on average.