Quote from: Bloody Stupid Johnson;632968True... Perhaps this is something that is only a possible problem if mechanics are set up explicitly to specify whether a failure is external or internal, then.

If that, and if the explanational for the Story Point is something in-character (like I try really extra hard because like that's important). In those cases would would like to say that the Story Point can't be used against external failure.

Personally I think is to ditch the mechanic completely. It's not real, and highly skilled people choke in reality because they try too hard. Such a system owes more to daddy and mommy telling timmy he can always succeed if he just puts enough effort into it than anything else.

Reality Altering Fluff (like TORG) would be unaffected IMO.

Quote from: Bloody Stupid Johnson;632971EDIT: rereading that I can see it wasn't too clear. Another way to put it may be that external results the GM can describe, are limited to things that wouldn't have provided a modifier to the original roll.

I would agree with that as it just makes sense.

If I was going down this path I'd move the external failure closer to the success chance. For example if a character had a 68% skill on d100, I think an automatic external failure is better as a 69 rather than a 100. That way each time it came up, the player's first thought is "I almost had it!".

Appearances can be important.

This has been some fun food for thought.


It has given me an idea. The failure rate for 1st level B/X thieves is astronomically high. Just for S&G I'm going to come up with a custom excuses table for each skill.

Quote from: Exploderwizard;633228This has been some fun food for thought.


It has given me an idea. The failure rate for 1st level B/X thieves is astronomically high. Just for S&G I'm going to come up with a custom excuses table for each skill.

Those can be fun until the results start to repeat. To keep it fresh however, you could make a different table for each encounter area, or even a subtable that fills in a a section of a general one.

Quote from: gleichman;633231Those can be fun until the results start to repeat.

That's one of the reasons my first inclination is to simply improvise based on the situation, rather than try to set everything down in advance.

Quote from: Black Vulmea;633244That's one of the reasons my first inclination is to simply improvise based on the situation, rather than try to set everything down in advance.

Bingo.

So,  Black Vulmea and Benoist have nothing better to do than jump in and agree with my original post...

...I think I like it better when they threadcapped.

Tables can be more generic and just provide a result the GM has to interpret situationally/get inspiration from, or you can have 2 tables where you have to join the results together so that exact duplicates are rarer.
http://www.theyfightcrime.org/         =P (OK, not a real example)
 
I do have to admit we haven't found any really convincing reasons as to why we would want to determine why something failed, apart from consistency issues (in things that no one would notice anyway), given that they're hidden beneath the abstraction of the system. For me, the blog post & thread has helped me clarify in my mind some ideas as to how to make sense of random dice rolls at least.

Other than that, another thought wrt this...
 

Quote from: beejazz;632797I guess leaving the reasons for failure within the abstraction of dice-rolling doesn't bother me as much as certain other things. For example, if you ran a game of Pulp Fiction, the scene where the gun goes off accidentally could never happen because no one was attempting anything. Basically people rarely if ever choke on potato chips or get T-boned at an intersection in RPGs, but that sort of thing can be interesting and in-genre in certain cases.

Quote from: ewilen's blog postMythic RPG has "interrupts"--basically any die roll can trigger a random event. In either of these games I suppose you could interpret a regular failure as being the character's fault, but special failures are extrinsic. But I don't think this idea has ever been made fully explicit.

I don't know anything about Mythic, but this looked related. Someone gets an 00 (or whatever it is in Mythic) for Drive and the random event goes off?
 
Either that or something non dice-related - Marvin had a Dark Fate flaw. Or it was the start of a new session and the player couldn't make it...=P

No one-size-fits-all solution from me, but there are surely options in between improvising on the spot and having a fixed table.

Like, you could have a table which speaks in generalities or is subject to flexible interpretation--anything from a taxonomy of causes to an oracular card draw (tarot, Everway, etc.) to word-association as in Mythic.

Or you could have a table or tables with fairly specific stuff, which you populate before the game. If something gets used, you cross it off and reroll if it comes up again. In between games, you re-populate. Effectively this is the same as drawing slips of paper from a hat, and if one is so inclined, one could ask players to contribute slips.

* * *

I like the argument that external causes should be restricted to a fixed chance. Not so sure about the idea that multiple characters rolling should independently trigger external events that affect all characters. I'm sure there are ways to engineer things to avoid the perverse incentives that this could generate.

* * *

One important point which is obscured by the question of "what exactly happened?" is that I'm emphasizing the simple benefits of uncertainty and defending them against what I'll uncharitably call "failure of imagination". A game might have what seems to be a ridiculously high failure rate for trained persons. But once you think about external causes, maybe not. And external cause is extremely flexible--part of the plausibility comes from the level of abstraction involved in framing the situation that leads to the dice roll. If your abstraction limits the player's ability to control the exact conditions--and the GM's ability to dictate the exact conditions--then anybody can fail at anything.

And the possibility of failure may be more fun than a sure thing, as long as the consequences of failure don't suck. Don't have a result on your table be "Life on Earth obliterated by asteroid, end of game." At least not unless you assign it the actual probability, which is infinitesimal. But a foolproof plan to burrow into the bank vault? Somewhere there's a chance of cocking it up or having someone stumble over your activities. So roll for that.

The caveat I would make here is that rolling for a perfect plan or perfectly-laid sniper shot is only absolutely necessary for the GM's entertainment. If the GM really has things detailed well enough that the players can say what they do, not knowing with absolute certainty if their plan will succeed, but with confidence that the cause of failure won't just be "because the GM thought it would be cool", then you don't need to roll the dice. The purpose of a stochastic model is to [objectively] abstract the unknown; if things are actually there but literally unknown, you can use a deterministic model with limited information.

Quote from: Bloody Stupid Johnson;633329I do have to admit we haven't found any really convincing reasons as to why we would want to determine why something failed, apart from consistency issues (in things that no one would notice anyway), given that they're hidden beneath the abstraction of the system. For me, the blog post & thread has helped me clarify in my mind some ideas as to how to make sense of random dice rolls at least.

For me it's been a fun thing to play with. Taking an concept, turning it over a few times to find the best possible way to do it and then deciding if it's worth doing is just plain good excercise.

In this case, I likely won't change my methods of at all. However my narrative of the failure will be done with a bit more awareness of possible causes, and externals ones will become more common on 'near misses' with the dice.

Quote from: Elliot Wilen;633337I like the argument that external causes should be restricted to a fixed chance. Not so sure about the idea that multiple characters rolling should independently trigger external events that affect all characters. I'm sure there are ways to engineer things to avoid the perverse incentives that this could generate.

IMO the best solution is to never allow multiple character rolls like that.

But assuming I did, what perverse incentives are you speaking of?

Say there's a fixed value for "external failure". On an 00 result, e.g.

Or suppose it constitutes 5% of the base chance of failure. E.g. if you're success chance is 80%, you have a 20% chance of failure, ergo a 1% chance of external failure--again, a 00. But if you have a 60%, then it's 99-00.

Not saying whether either of these better for the individual case; the point is it's always the same whether you have one person or multiple people rolling.

Then say that an external failure is defined as something that affects the whole group.

In that case, if multiple people make the attempt, the chance of the canonical truck crossing the intersection goes up. Perverse incentive is to not have multiple people chase the bad guy or whatever.

As I said, there are solutions. One is to use just one roll for the best skill, with bonuses for helpers. Another is to only have a "group external" on the first roll. I'm sure there are others.

Quote from: Elliot Wilen;633407In that case, if multiple people make the attempt, the chance of the canonical truck crossing the intersection goes up. Perverse incentive is to not have multiple people chase the bad guy or whatever.
 
As I said, there are solutions. One is to use just one roll for the best skill, with bonuses for helpers. Another is to only have a "group external" on the first roll. I'm sure there are others.

Yep, tricky. I noticed this as well (alluded to it in post #4). Its problematic both from a game incentive POV, and from world consistency.
 
The idea of a fixed percentage of external failure works well I think for comparing unskilled vs. skilled characters, but becomes perverse when the number of rolls increases. (Using multiple characters rolling low to judge an external event has occurred is something that I would use with best judgment of how unlikely an event is, rather than a fixed percentage on each roll).
 
A third approach (to go with those two listed - rolling only once, & externals on first roll only) would be to have a second 'confirmation roll' for external results. If any character rolls a 'possible external event', then another dice is rolled to see which character's roll applies.
i.e.  a system where d20s are used and a '1' is possible external cause.
4 characters are in pursuit with d20 rolls of 1 (possible external) for character A, 1 (possible external) for character B, 5 for character C and 19 for character D...the GM then rolls a d4 for the four characters to see which roll applies (A, B, C or D); in this case a roll of 1-2 on d4 would mean an external event.
This keeps odds at more-or-less a fixed 5% chance regardless of how many characters are rolling. A bit clumsy though. (Something I made, that never made it off my desktop, used a similar idea to do fumbles for characters with multiple attacks).
 
That system would let characters with high skills roll badly (i.e. 1s in the above example) and not have an external cause to explain their failure however. Possibly the same problem for the '1st roll is group external roll' approach;  if James Bond is the second person to roll he doesn't get an excuse for cocking it up.
 
Actually, I think that's a general thing; if multiple characters are rolling either the odds of external failure will rise, or low results become more inexplicable except through incompetence.

Quote from: Elliot Wilen;633407In that case, if multiple people make the attempt, the chance of the canonical truck crossing the intersection goes up. Perverse incentive is to not have multiple people chase the bad guy or whatever.

Which is why I never have multiple people roll, as I consider the perversion to work both ways (i.e. it's better for multiple people to have low or middling skills in a lot things than it is to specialize in a few when every one gets to roll against a task).

You already had a Perverse Incentive problem before you ever reached the question of external failure.