Problem with Stealth?

[Andreas]

Originally Posted by Verjigorm

You've never knocked over something loud while walking through your house at night to the bathroom?

Never stubbed your toe and yelled an obscenity?

Never accidentally kicked a glass bottle in an alleyway and it rattles, clatters and crashes?

No. I have bumped into stuff, but not anything that loud.

Not if I'm trying to be stealty. Unless they have very poor self control, I have a hard time seeing someone yelling out like that without even holding back the volume of their voice while seriously trying to be stealty (sure, it could happen if you don't really care much about being quiet in the first place and abandon that attempt due to the greater concern of moderate pain, but that does not really fit most adventuring situations).

Not while trying to be stealthy.

[David Johnston2]

Originally Posted by kdtipa

That is an excellent point. But I'm not likely to roll for everyone in the camp. Just people who have line of sight or even the potential of hearing. .

Which still boils down to at least one stealth contest for every hex you move making success functionally impossible.

If you stick with no one getting to make a perception roll unless they're specifically looking, then Ninja-San suddenly has nothing to worry about once he's past guards on a perimeter. It feels wrong.

And is wrong because it isn't just guards assigned to the perimeter who are on edge in a war zone. Ninja-San might have to make a couple of contests because the camp has an NCO who is on the alert for misbehaving soldiers, or a soldier with a bit of PTSD keeping him on edge. And it's likely that Ninja-San's specific objective will be valuable enough that it ill have a guard or two.

But I've played in games where the GM just threw stealth roll after stealth roll at me until I was inevitably caught and shot. It wasn't fun and it wasn't reasonable.

[whswhs]

Originally Posted by Andreas

They aren't extremely uncommon in GURPS. Unless your effective skill is very high, they happen one in 72 times (and otherwise one third as likely as that), so it makes sense for them to represent misstakes that could plausibly happen with such frequency rather than extreme outliers.

There is no way it can be one in 72. If your effective skill is 15 or less, it's one in 54; if 16 or more, it's one in 216.

And that method of estimation produces absurd results. Consider, for example, vehicle operation. Any critical failure is not merely a crash but a bad crash. A control roll is required for a landing in bad weather. One bad crash out of 216 such landings would cause so many deaths per year that air travel would be shut down on any such day. The 1/216 odds are unrealistically high. The probabilities you have in GURPS play are not probabilities in the real world, but more like probabilities in movies or novels where exciting things happen.

[Stormcrow]

Originally Posted by Brandy

A lot of things that I've seen people point out as "problems" through the years go away if you stop thinking of GURPS as a simulation engine and start thinking of it as a game.

I would like everyone to please tattoo this on themselves.

[Gumby Bush]

There are several resolutions of Stealth:
1. I sneak up on so-and-so.
2. I sneak past the guards.
3. I sneak through the base into the inner sanctum.

1 and 2 are pretty well handled by Stealth as it stands, although we occasionally see the "how close can I get without being noticed?" question. A higher-resolution handling of Stealth for type-3 should probably be a single roll handled in much the same way as the "how close" question gets answered, with modifiers based on guard density.

Off the top of my head: let's say you take a penalty to Stealth based on your speed (Size & Speed/Range Table) and how much distance you are covering (Also Size & Speed/Range Table). If you fail or are seen, it is at the first point where the distance penalty would be too great for you.

When sneaking past guards, it is contested by the highest skill held after ordinary modifiers would apply (so, if they put a guard on a hilltop with a telescope, see if he has a higher skill that way), +1 for each alert/wary guard, and -1 for troops who are supposed to be alert but are out of it (in other words, gaps in security) and with a penalty for the size of the area, which should be a flat value: -1 per 100yds? (this is where a +10 for spotting something in plain sight balances out the -10 for distance).

You may want to modify the resulting number for gross discrepancies (Eagle-Eyed Erik is on watch, but everyone else on alert is half-deaf), but it should work for most situations.

[AlexanderHowl]

Originally Posted by whswhs

There is no way it can be one in 72. If your effective skill is 15 or less, it's one in 54; if 16 or more, it's one in 216.

And that method of estimation produces absurd results. Consider, for example, vehicle operation. Any critical failure is not merely a crash but a bad crash. A control roll is required for a landing in bad weather. One bad crash out of 216 such landings would cause so many deaths per year that air travel would be shut down on any such day. The 1/216 odds are unrealistically high. The probabilities you have in GURPS play are not probabilities in the real world, but more like probabilities in movies or novels where exciting things happen.

It is more like the probability involved for non-routine activities. In the case of Driving, I think that Driving rolls should not happen during routine driving (going to work, picking up groceries, driving in light snow, etc.) as long as the effective Driving skill is 16 or higher (12 plus 4 for routine activities). Instead, I think that Driving rolls should only happen during stressful or unusual circumstances that reduce effective Driving skill below 16 (going to an interview for a new job, picking up a beautiful date, driving during a blizzard, etc.). Of course, life happens, and GMs should call for Driving rolls when appropriate to the story (or once per day with a character with Bad Luck or every time with a character with Cursed).

[Flyndaran]

IMO, it's more that in driving, if you fail, other drivers may succeed reducing a potential accident to merely a close call and middle finger flipping.
Or as a pedestrian, I jump back hard and fast enough, which I have to do on a near weekly basis, no hyperbole.

[Icelander]

There is an implicit assumption that 'normal' rolls are made during adventures, with mundane activities being condensed into a Job Roll over the whole month. And the fact is, under conditions of stress and confusion, like any combat situation ot potential combat situation, people are a lot more likely to fail dramatically and catastrophically.

Everything is very simple in war, but the simplest thing is difficult. These difficulties accumulate and produce a friction, which no man can imagine exactly who has not seen war.
- On War, Carl von Clausewitz.

[whswhs]

Originally Posted by AlexanderHowl

It is more like the probability involved for non-routine activities. In the case of Driving, I think that Driving rolls should not happen during routine driving (going to work, picking up groceries, driving in light snow, etc.) as long as the effective Driving skill is 16 or higher (12 plus 4 for routine activities). Instead, I think that Driving rolls should only happen during stressful or unusual circumstances that reduce effective Driving skill below 16 (going to an interview for a new job, picking up a beautiful date, driving during a blizzard, etc.). Of course, life happens, and GMs should call for Driving rolls when appropriate to the story (or once per day with a character with Bad Luck or every time with a character with Cursed).

Even in landing in bad weather, which is a nonroutine activity, you have at least 1/216 chance of a bad crash, and likely 1/54 for pilots who aren't at the top end of the profession. That's substantially too high for a realistic treatment of air travel. But it's probably accurate for pilots in movies!

The thing is, there are things that people do every day, and that adds up to a huge number of rolls. But RPG sessions don't cover every day (in most cases) and don't have rolls for every activity. The number of rolls a GM can reasonably call for is vastly less than the number of tests of skill people face. (For example, Megan McArdle has written columns about the recipes of 1920-1960 largely being intended to work for cooks, largely female, who weren't highly talented or skilled but had to cook something for dinner every day. In GURPS terms, they were hardly rolling at skill 16 or better!) So RPGs commonly do probability compression, making the risk of dramatic success or failure far higher than it is in real life, to correspond to that smaller number of rolls—and perhaps more than that, because RPGs are supposed to be dramatic.

[whswhs]

Originally Posted by Icelander

There is an implicit assumption that 'normal' rolls are made during adventures, with mundane activities being condensed into a Job Roll over the whole month. And the fact is, under conditions of stress and confusion, like any combat situation ot potential combat situation, people are a lot more likely to fail dramatically and catastrophically.

Or, conversely, the times when freakishly improbable things happen are the times people tell stories about, and therefore play a disportionately large part in narrative.