Perception vs. Stealth

Best Selling RPGs - Available Now @ DriveThruRPG.com

ffilz

Legendary Pubber
Joined
Dec 17, 2018
Messages
3,494
Reaction score
6,670
I'm thinking specifically about Cold Iron, but I open this for discussion of perception vs. stealth in any game system. There are definitely some things that work well and some things that don't work well in various systems.

So last night we had our first real run of Cold Iron, and we wound up immediately with a Perception vs. Stealth check. This makes it clear to me that I need to look at this more closely to come up with good ratings.

Historically perception has been handled using the Alertness (ALT) attribute as the primary factor with arbitrary target numbers. Now since ALT is potentially improvable with level gain (Expertise Levels), there is a factor of level. So the way I ran the ALT check last night was:

CA = Chance Adjustment produced by rolling on the normal distribution "Chance Adjustment" chart.

ALT + CA vs DEX + Stealth Level.

I don't think that's right. Currently I don't have a Perception skill, and I'm not sure I want to add it. I could just use highest class level.

Also, combat ratings normally use the attribute modifier (adjustment) so ALTadj for example.

On the other hand I also need to consider the perception and stealth affecting spells. Obscurity and Silence allow for up to a -10 modifier. Enhance Vision/Hearing/Smell (there isn't a stealth odor spell) grant a +6 modifier if applicable.

I think given the magnitude of the spell modifiers, I should stick with ALT rather than ALTadj, but do:

ALT + Class Level (depending) + Skill Level (if applicable) vs. DEX, INT, or WIS (depending on type of stealth) + Class Level (depending) + Skill (if applicable)

Fighter Level should apply to ALT checks for combat stuff, an appropriate combat skill might apply
Expertise Level should apply to most ALT checks
Magic Level (Magic User, Cleric, or even Passive Magic) would apply to ALT checks for magical effects and maybe certain searches (and maybe some of those would use INT or even WIS instead of ALT).

Some skills that would apply to perceiving:

Survival (Tracking and general wilderness perception), Shadowing, Research (finding stuff in a library), Devise (finding traps or secret compartments), Engineering (finding secret doors), Naturalist (finding specific plants or even animals).

Some skills that would apply to not being perceived:

Stealth, Disguise, Forgery, Devise (making a trap), Survival (making a deadfall or trip wire), Sleight of Hand.

On the not-being-perceived side, usually Expertise Level but sometimes Magic Level or Fighter Level.

Situational modifiers should be used to adjust things away from the nominal 50-50 chance. Like hiding in a particularly good or bad place.

I also need to think about the numbers game for number of people perceiving vs. number of people trying not to be perceived.
 
Don't know how well it'll format so I'll put it in a spoiler. Sources I used to work up a stealth system for another game. They're mostly military studies of distance, nighttime, jungles, movement, size, sound, etc. It turns out militaries give a shit about hiding & spotting and you don't have to blindly ass-pull numbers for game systems.
Bibliography
(just in case you really want to replicate some
of this for yourself)

AD-753 600 TARGET DETECTION AND
RANGE ESTIMATION
James A. Caviness, et al
Office of the Chief of Research and
Development (Army), November 1972

RESEARCH MEMORANDUM,
MOONLIGHT AND NIGHT VISIBILITY
Thomas F. Nichols and Theodore R. Powers
USAIHRU, January 1964

Jungle Vision II: Effects of Distance,
Horizontal Placement, and Site on Personnel
Detection in an Evergreen Rain forest
Dobbins, D.A. et al.
U.S. Army Tropic Test Center, Fort
Clayton, Canal Zone, March 1965.

The Effects of Observer Location and
Viewing Method on Target Detection with
the 18-inch Tank-Mounted Searchlight
Louis, Nicholas B.
HumRRO Technical Report 91, June 1964.

Report Bibliography on Target Detection
and Range Estimation
ASTIA, Humans, Armed Forces Technical
Information Agency, Arlington, Virginia,
November 19 60.

Scripps Institution of Oceanography.
Predictions of Sighting Range Based Upon
Measurements of Target and Environmental
Properties
Jacqueline I. Gordon.
VisibilityLab/reports/SIO_63-23.pdf

Memorandum RM-6158/1-PR, Target
Detection Through Visual Recognition: A
Quantitative Model
H.H. Bailey, February 1970
research_memoranda/RM6158z1.html

RESEARCH ON VISUAL TARGET
DETECTION PART I DEVELOPMENT
OF AN AIR-TO-GROUND DETECTION/
IDENTIFICATION MODEL
Margaret E. Franklin, John A. Whittenburg
June 1965

Detection of random low-altitude jet aircraft
by ground observers (Tech. Memo. 7-60;
AD 238 341)

AD-758 875: CAPABILITIES OF
GROUND OBSERVERS TO LOCATE,
RECOGNIZE, AND ESTIMATE
DISTANCE OF LOW- FLYING
AIRCRAFT
Robert D. Baldwin
Human Resource's Research Organization,
March 1973
 
Trying to keep this above the stormy seas of the OGL... :smile:

OK, that's quite a list of studies, but has anyone actually distilled that into game mechanics?

I really am interested in having a good game mechanic that both holds up to "realism" sufficiently while also being "fun" for gaming. Realism may be reduced to improve fun.

Hmm, has anyone ever compiled statistics on hikers spotting various critters? That would help understand the probabilities of actually having a meaningful encounter. On the flip side would be statistics on encounters were animals have attacked hikers, and a secondary, but important question on those, what are the statistics on encounters where hikers have been ambushed by an animal?

Of course military statistics on ambushes would help. An owlbear wanting to ambush an adventuring party is going to have encounter statistics closer to a military ambush that a hiker spotting a moose on the trail even if the military ambushes are impacted by technology such as camouflage nets and binoculars or other vision enhancing devices. Of course the impact of night vision goggles on ambushes would be great statistics for infravision/night vision critters vs. humans.
 
Well yeah I worked that into a game system. That was the whole point. Had a game with perceptiin & stealth skills but zero functional rules surrounding them.

I also ended up using some stuff that didn't make it into that list, but it was mostly just reinforcing the graph shape. Stuff like facial recognition by distance, methods & math of correctly & effictively simulating distance with size for vision studies, visual clutter studies, etc.

Here, google drive link to the DtD40k7e pdf of my home rules. The stealth & perception bits are in appendix A, page 360.


Edit: oy, didn't expect that to try to load the doc.
 
Where in the document should I look to understand the various attributes like size? And do you have a quick summary for the spread of the randomizer?
 
Trying to keep this above the stormy seas of the OGL... :smile:
Well, we should stick to non-d20 systems here...:grin:
I really am interested in having a good game mechanic that both holds up to "realism" sufficiently while also being "fun" for gaming. Realism may be reduced to improve fun.
:shock:
No. It may not. If we reduce it, we can just as well stick to opposed rolls.

Now for my actual contribution (hopefully)...
I consider sneaking to be a variation of all those other times when you have to do something that other people might want to prevent, and probably doing that before they realize you're doing it.
So, for example, you have two kinds of guards: alert and searching for intruders, assuming that they are there, and non-alert, not searching for intruders, having pushed the idea of possible intruders in the dark recesses of their mind along with Form 3/11333 that everyone has to fill to get lunch...no, wait, that was the form for breakfast:shade:!
Savage Worlds also classifies them like this, BTW. IIRC, passing by the unalert guards is a simple unooposed test, while passing by the alert ones is an opposed one (and you still have to pass the simple test, because fuck "guards looking for you makes it easier").

OK, and there's the third kind - the guards that are hurrying towards the raised alert; if you encounter those, you fucking failed at stealth already:tongue:!

So basically, the first obstacle is to get them to lower the guard. That might require time, if you can afford it.
Or you can guess that they'll be most alert at the start of their shift or right after a check by a superior, and add appropriate penalties for acting at another time (which might well be necessary, if your PCs are short on time).

And then there's the visibility thing. Is there a path where the PCs can pass without being noticed? Brushes and the like? There's a reason the military cuts those off. If there isn't, and anyone is watching, like a sentry in a crow's nest, you autofail (unless the sentry was there for 12 hours already and fell sound asleep; that's basically a botch on a Discipline check or the like).

So my approach is simply "consider what has been going on in the place where they're trying to sneak, and move from there".

Telok Telok the first link is giving me this notification:

"

This document is not available in digital form.
If you are supporting DoD or U.S. Government research please Sign In using a CAC, PIV or ECA or register with DTIC. Once registered, sign in, search for your document, and click on “Request Scanned Document”.
If you are ineligible to register, you can request this document through FOIA.

DTIC’s public technical reports have migrated to a new cloud environment. The link you used is outdated. Please use the information below to correct the link. Contact 1-800-CAL-DTIC (1-800-225-3842) if you still have issues."
 
Where in the document should I look to understand the various attributes like size? And do you have a quick summary for the spread of the randomizer?
Easy enough. Size is scaled: 1-10 each point is about 0.5 meters, 11-20 is 1m, 21-30 is 2m, 31-40 is 5m, 41-50 is 10m, 51-60 is 20m, 61-70 is 50m... pattern repeats.

Target number/DC chart is exactly the same as D&D 5e despite being written 2 years earlier than 5e.

Average human stat is 2, max 5. Skills 0 to 5. Untrained skills rolled at stat-1. Stat 2 & skill 1 hits 15 about half the time, stat 3 & skill 3 hits 25 about half the time, stat 4 & skill 4 hits 35 about half the time.

About the links, well that's interesting. I still have the docs around somewhere. I'll throw them on google and link tonight if at all possible. Considering so many are 50+ years old, sheesh.

My having written it has made me really salty about cover & stealth & perceptiin rules that are all just a bunch of text and ass pulls. The pics helped so damn much. And of course you still need to hammer home if it's a roll-everything, roll-important, or roll-only-if-absolutely-no-other-choice game system.
 
One approach that I think could work is the one adopted by stealth-based video games like Thief, Dishonored, or The Dark Mod where it's not just a binary choice between being undetected and getting busted. These games have 3-5 suspicion levels and only the highest of these is an outright bust. If a guard hears a tiny noise or catches sight of some movement in the corner of his eye he's not going to immediately sound the alarm. He'll mutter something like, "Did I see something?" and remain a little bit more alert for a while, but then settle back down to a lower suspicion level. At higher suspicion he might abandon his patrol route to search for a bit, but if he doesn't find you he'll eventually go "Must have been a rat" and calm down again.

You could do something similar here. Roll your stealth against his alertness- on a success nothing happens, on a fail the guard gets more suspicious and might get a bonus on subsequent alertness checks, but you shouldn't get outright busted unless he hits the highest suspicion level.

This, I think, is better than rolling stealth all the time with a single failure spelling disaster. A little bit more bookkeeping, but I think it's worth it.
 
One approach that I think could work is the one adopted by stealth-based video games like Thief, Dishonored, or The Dark Mod where it's not just a binary choice between being undetected and getting busted. These games have 3-5 suspicion levels and only the highest of these is an outright bust. If a guard hears a tiny noise or catches sight of some movement in the corner of his eye he's not going to immediately sound the alarm. He'll mutter something like, "Did I see something?" and remain a little bit more alert for a while, but then settle back down to a lower suspicion level. At higher suspicion he might abandon his patrol route to search for a bit, but if he doesn't find you he'll eventually go "Must have been a rat" and calm down again.

You could do something similar here. Roll your stealth against his alertness- on a success nothing happens, on a fail the guard gets more suspicious and might get a bonus on subsequent alertness checks, but you shouldn't get outright busted unless he hits the highest suspicion level.

This, I think, is better than rolling stealth all the time with a single failure spelling disaster. A little bit more bookkeeping, but I think it's worth it.
That's more or less the approach taken in Heaven's Shadow, one of my Favourite Games That I Haven't Really Played...:grin:
Namely, it has a "fighting for advantage" rules which it also uses for grappling as well* as stealth. If you get in a situation noted as "fighting for advantage", you initiate an opposed roll. If you win, you can move to the next step and fight for decisive advantage.
To make it, you have to basically get two consecutive wins (or one Double Win, rolling twice more than the opponent). If you lose, well, the NPC gets to break off (in the example in the book, it becomes "the goon the Agent is following slips off"). If you win, however, you can use the effects listed under "Double Win". And those are simply brutal.


*Funny enough, while writing my first post in the thread, I was thinking of grappling/fencing, but ultimately decided to stick to sneaking and instead wrote the kinda-clumsy sentence "a variation of all those other times when you have to do something that other people might want to prevent, and probably doing that before they realize you're doing it". Then I remembered Heaven's Shadow had that, and just checked it...:shade:
 
I feel like the Perception vs. Stealth match up should always be treated as a contested roll rather than as one or the other party rolling vs. a task number or skill rating. For my money, the best contested roll mechanic anyone ever dreamt up is the one from Pendragon. So, that's how I think all such situations should be handled.

The worst possible way to do it is traditional (pre3E) D&D. And that is coming from someone who loves and is currently running traditional D&D. But if you are in that OSR-y part of the gaming metaverse and not using house rules to handle perception and stealth, then you need to get your life right.
 
Thanks for all that. I'll have to digest it.

For comparison, Size in Cold Iron is 1 point of Size is 10 lbs mass. In the 4-20 range, modifiers are linear, beyond 20 the breakpoints go geometric (23, 26, 30, 35, 41, 48...). The randomizer uses the cumulative normal distribution to convert a random number between 0 and 1 into a modifier. It is set up so +0 requires a 0.5 or better, with a +20 being 3 standard deviations, so getting a +3 requires a 0.67 or higher, +7 0.85 or higher, +15 0.988 or higher, +25 0.999912 or higher. Getting a +N requires =NORMDIST(N,0,20/3,TRUE) or higher (we didn't have that handy Excel formula back in 1982...). The randomizer is basically rolling 1d10 for each decimal place, and the way the normal distribution works, basically good enough resolution is had if you roll 2 more decimal places beyond a string of leading 0s or a string of leading 9s.

A bog average human is going to have stats of 10-12 for +0 to +1 modifier. A really good human is going to have 18-20 for +4 to +5 with an exceptional human having a 23 for +6 and a super-human having a 26 for +7 and I GUESS you COULD manage to roll a 30 for +8... (roll an 18 with a 6 potential, and then on rolling an 18, you roll all 3d6 and each 6 is another point, and roll that die again, so the 3d6 have to generate 6 more 6s on reroll - maybe a bit easier with 4d6k3 attributes where you might even roll 4 6s initially - which to be super nice to the player could be a 19, and then you get 4 dice to re-roll...).

Very few things have Dexterity or Altertness much higher than 30ish.

And as mentioned magic can improve stealth or altertness by as much as +10.
 
I feel like the Perception vs. Stealth match up should always be treated as a contested roll rather than as one or the other party rolling vs. a task number or skill rating. For my money, the best contested roll mechanic anyone ever dreamt up is the one from Pendragon. So, that's how I think all such situations should be handled.
D20 blackjack roll-under? It's used in many d100 games as well, BTW, Mythras included:grin:!
And for the Perception vs Stealth being always contested, I would agree as long as the one you're sneaking on/by is actively searching. If not...well, I think a normal skill check to avoid giving yourself out is appropriate. After all, you just have to pass without attracting attention (making noise or a movement when you're in the general area guards can observe).
Which would then lead to the guards searching actively, i.e. a contested roll.


The worst possible way to do it is traditional (pre3E) D&D. And that is coming from someone who loves and is currently running traditional D&D. But if you are in that OSR-y part of the gaming metaverse and not using house rules to handle perception and stealth, then you need to get your life right.
Tentatively agreeing on both accounts...though I'm not sure how pre-3e D&D was handling stuff, other than "it wasn't by contested skill checks":thumbsup:.
Let me guess, a 2-in-6 check?
 
Telok Telok the first link is giving me this notification: "This document is not available in digital...
I could only find a couple, but I did successfully get all the ones with links. I think. Here's a google drive folder with the docs and a couple other bits I've found useful.

drive.google.com/drive/folders/1-rcnN9InvOMwQTU48ceVX-7kpdgfoTZ6
 
Here's that sighting table from Delta Force by Task Force Games. I'm not sure whether we left these untweaked - I remember re-writing a lot of the numbers, but this was in the '80s so memories of details are sketchy. We did use it a lot (in my Traveller conversion) because characters avoided firefights like the plague in favour of well-prepared sniping - always try to take down the opposition before they know you're on the same planet.

Sighting-Combat Tables.png
 
Bump since another similar conversation started...
 
I thought I'd bump this again. I'm working on my perception and stealth rules for Cold Iron Blackmarsh Adventures and I could really use some guidelines on likely encounter distances based on terrain type. I like what I read in Dungeons the Dragoning mentioned above, but while it helps with DCs for perception checks, it isn't easy to suss out encounter distances.

Also, in more open terrain, vision is going to be the major factor, while in dense terrain, sound may be more important. And smell should come into play (mostly for creatures).
 
I thought I'd bump this again. I'm working on my perception and stealth rules for Cold Iron Blackmarsh Adventures and I could really use some guidelines on likely encounter distances based on terrain type. I like what I read in Dungeons the Dragoning mentioned above, but while it helps with DCs for perception checks, it isn't easy to suss out encounter distances.

Also, in more open terrain, vision is going to be the major factor, while in dense terrain, sound may be more important. And smell should come into play (mostly for creatures).
Do you have a systematic treatment of the density of cover in terrain types? "Terrain value" or anything like that?

In ForeSight you set up encounters by calling for a Scan (perception) roll from the character with the best scan (in the environment) and the worse Stealth (in the environment) on each side, add them for a sum for each side, and then modify it for behaviour (e.g. attempting to set an ambush). You compare the totals and look up the difference between them on a table, which gives you results of "surprised" or contact at contact range or at observation range. Then the values of contact range and observation range as functions of terrain value (which is tabulated by terrain types).

Sorry, that's not as clear as it could be.

ETA

In ForeSight there is a system for classifying terrains depending mostly on how rough they are and how much vegetation they have, but with suburban, urban, water surface, and submerged sprouting off to the sides of the table.

The Environment modifier E appears in BEFs of skills dependent on perception. since a good idea of what to expect in a place allows a person to concentrate on worthwhile subjects for his/her attention. Environment familiarities are expressed on a table to allow the great variety of possible environments to be compactly expressed. In general the environment is expressed as a contour (BR: broken, UN: uneven, FL: flat) and a surface feature (BN: barren, LV: light vegetation, MV: medium vegetation, HV: heavy vegetation. MA: marsh, IC: ice); there are also miscellaneous environ- ments tacked onto the (in my opinion on the day) most appropriate places (SF: surface of water, SB: submerged in water. SU: suburban, ie. scattered build- ings, UR: urban, ie. built up). Wherever a value for E is called for, the most appropriate is chosen.

Each character has a degree of familiarity with each terrain, for none to 3, which are set out in a little table on the character sheet.

Screenshot 2024-02-28 at 13.05.47.png

The character's familiarity with the terrain they are in enters into the ease factor for each of the Scan skill and the Stealth skill. Each terrain type also has a Terrain Value, set out in a table in Ch. 5. And there is a table in the section on "Resolving Wilderness Encounters" in Ch. 3 that gives a contact range and an observation range for each terrain value.

To resolve the circumstances of a wilderness encounter, the character with the worst Stealth (lowest PCS times modified Ease Factor) and the character with the highest Scan (highest PCS times modified Ease Factor) on each side rolls the skill, getting a Quality Rating 1 to 4 (or 7 on a fail, 10 on a botch).

[3.10] Resolving Wilderness Encounters
Procedure: Both the sides involved make a Scan roll (using the best SC - EFxPCS in the group) and a Stealth roll (using the worst) and add the QRs together (this sum is reduced by four if the party was attempting an ambush, by two if the characters were proceeding cautiously, ie. at half speed, and is unadjusted otherwise). If one side's sum is four or more lower than any other side's then the former is said to have surprised the latter (if an ambush was attempted, then the latter walked right into it); if one side's total was three lower than the other's, that side may surprise the other at observation range, or start with the initiative at contact range; if one side's total was two lower, it may start combat at either contact or observation range (its choice) and determine initiative normally, or it may have the initiative and leave the choice of range to the other side; otherwise, the initial range is determined randomly (25% chance of contact range, 75% chance of observation range). See Combat Set-Up [4.1].

Screenshot 2024-02-28 at 13.08.49.png
 
Last edited:
Off the D&D5e gm screen

Encounter distance
Arctic, desert, farmland or grassland 6d6x10 feet
Forest, swamp or woodland 2d8x10 feet
Hills or wastelands 2d10x10 feet
Jungle 2d6x10 feet
Mountains 4d10x10 feet

Audible distance
Trying to be quiet 2d6x5 feet
Normal noise level 2d6x30 feet
Very loud 2d6x50 feet

Visibility outdoors
Clear day, no obstructions 2 miles
Rain 1 mile
Fog 100 to 300 feet
From a height x20
 
Do you have a systematic treatment of the density of cover in terrain types? "Terrain value" or anything like that?

In ForeSight you set up encounters by calling for a Scan (perception) roll from the character with the best scan (in the environment) and the worse Stealth (in the environment) on each side, add them for a sum for each side, and then modify it for behaviour (e.g. attempting to set an ambush). You compare the totals and look up the difference between them on a table, which gives you results of "surprised" or contact at contact range or at observation range. Then the values of contact range and observation range as functions of terrain value (which is tabulated by terrain types).

Sorry, that's not as clear as it could be.
Would love some details.
 
Off the D&D5e gm screen

Encounter distance
Arctic, desert, farmland or grassland 6d6x10 feet
Forest, swamp or woodland 2d8x10 feet
Hills or wastelands 2d10x10 feet
Jungle 2d6x10 feet
Mountains 4d10x10 feet

Audible distance
Trying to be quiet 2d6x5 feet
Normal noise level 2d6x30 feet
Very loud 2d6x50 feet

Visibility outdoors
Clear day, no obstructions 2 miles
Rain 1 mile
Fog 100 to 300 feet
From a height x20
The first encounter distance seems small. Max of 360 feet in grasslands?
 
The distances are based on video game play. You can't really see anything until it "appears" on screen 360' away.
 

Also, in more open terrain, vision is going to be the major factor, while in dense terrain, sound may be more important. And smell should come into play (mostly for creatures).
I have some stuff worked up that applies to sight, sound, smell.

Distance based on some hunting type data (note most animals smell and hear you much further away than you think it seems, thank modern trail cams for that data)

Overall it is a total stimuli versus senses, stealth reducing the level of the stimuli. This is for detection. It can be ported to a stealth v. perception system with an assumed stimuli (e.g. noise level). It’s all linear and interchangeable, so a +1 noise is countered by +1 stealth.

For encounter distance, I make some base assumptions to determine a range…deviations from base assumptions modify that range. In a simple way, like +1 or +1 die.

I categorize the ranges, close, near, sort, long, etc then different environments get different ranges. Dense rain forest gets close, open desert long, for example.

I’m more about fantasy adventure verisimilitude than simulation, but feel it gives results not too off from reality.

Now my stuff is based on a count success approach but could give you the statistics (%) for certain assumptions and could align with your preferred mechanic.
 
I have some stuff worked up that applies to sight, sound, smell.

Distance based on some hunting type data (note most animals smell and hear you much further away than you think it seems, thank modern trail cams for that data)

Overall it is a total stimuli versus senses, stealth reducing the level of the stimuli. This is for detection. It can be ported to a stealth v. perception system with an assumed stimuli (e.g. noise level). It’s all linear and interchangeable, so a +1 noise is countered by +1 stealth.

For encounter distance, I make some base assumptions to determine a range…deviations from base assumptions modify that range. In a simple way, like +1 or +1 die.

I categorize the ranges, close, near, sort, long, etc then different environments get different ranges. Dense rain forest gets close, open desert long, for example.

I’m more about fantasy adventure verisimilitude than simulation, but feel it gives results not too off from reality.

Now my stuff is based on a count success approach but could give you the statistics (%) for certain assumptions and could align with your preferred mechanic.
That sounds good. Thanks for anything you can offer. Also pointers to any raw data would be cool too.
 
The basic problem with having a perception type skill is that your players will know there is something up whether they roll well or not. It is the GM saying " OK LISTEN UP! SOMETHING IS SNEAKING AROUND NEARBY. ROLL TO SEE IF YOU NOTICE IT." :errr: Roll or no roll, everyone knows that something is afoot.
For a campaign I am planning ( modified OSE rules), I am creating a perception stat based on INT+WIS/2. The key part is that stat will not be on the character sheet and the players will not be aware that it exists. I can compare things to possibly notice to their PER stat which functions as a passive check. Specific actions by the players that would give a bonus to PER score for a given instance are noted. The players simply state what they are doing ( moving ahead quietly and taking time to listen, looking very carefully at something specific, etc) This way the players can just naturally play and not go into high alert mode because a check is called for.
 
One approach that I think could work is the one adopted by stealth-based video games like Thief, Dishonored, or The Dark Mod where it's not just a binary choice between being undetected and getting busted. These games have 3-5 suspicion levels and only the highest of these is an outright bust. If a guard hears a tiny noise or catches sight of some movement in the corner of his eye he's not going to immediately sound the alarm. He'll mutter something like, "Did I see something?" and remain a little bit more alert for a while, but then settle back down to a lower suspicion level. At higher suspicion he might abandon his patrol route to search for a bit, but if he doesn't find you he'll eventually go "Must have been a rat" and calm down again.

You could do something similar here. Roll your stealth against his alertness- on a success nothing happens, on a fail the guard gets more suspicious and might get a bonus on subsequent alertness checks, but you shouldn't get outright busted unless he hits the highest suspicion level.

This, I think, is better than rolling stealth all the time with a single failure spelling disaster. A little bit more bookkeeping, but I think it's worth it.
Black Seven is the game that does this. An rpg based on stealth games. It's highly portable.

I'm goin to start a discussion with heresy - stealth isn't a skill. Stealth is a lack of perception. I'm wondering if it should even be there. Is there any skill in walking slow? In being behind cover?

Note, I also dislike perception because I think that it is graphically overused, but vs stealth is the one place I'd consider it.
 
Not really. You don't roll for stealth until it matters. If the sneaking creature never attempts to interact, neither it nor whoever it was sneaking around rolls. It is only when it attempts to do something, such as attack, use a power, commit theft, or some other interaction that might reveal it, that a stealth check is needed. A successful stealth check will make its action more effective in some way, while a poor check will hamper it.
 
I'm goin to start a discussion with heresy - stealth isn't a skill. Stealth is a lack of perception. I'm wondering if it should even be there. Is there any skill in walking slow? In being behind cover?
Yes, there is, even without considering moving quietly. Infantry in armies that expect to spend time in close country (bush, jungle, etc.) train in this. It's also why camouflage is a thing, and why snipers practice camouflage, concealment, and stealthy movement so much.

Being stealthy in a non-tactical situation is in some ways even harder, especially in a group - people like to chatter, they joke and gorse around when bored (or grouse and complain if they're too tired for horseplay), they careless bang tins together, stomp around, leave litter behind, and so on. Learning not to do this stuff, and having the discipline, is hard for many people. Naturally laconic loners will find the not-talking bit easier, but everyone needs to learn the little details.

And that's not getting into avoiding smelling like you don't belong.
 
I'd argue that stealth is a skill. Being able to walk quickly without making noise, knowing which floorboards are likely to creak if you tread on them and which aren't, being able to climb over things silently, and being able to hold completely still to avoid giving your position away by moving. It's more than just staying out of sight and tiptoeing around.
 
Nobody who's ever watched a Louis de Funes movie has ever doubted that stealth is a skill... one the main characters are lacking, usually:shade:.
 
The first encounter distance seems small. Max of 360 feet in grasslands?
Yeah, I don't agree with some of those myself. For example how could you see 20 times further in a fog when you are up high... :shock: I thought that that chart might be a useful starting place for you to build on.
 
Those rules for Twilight 2000 are pretty nice. I would argue that bad weather and night DO reduce how far you can see in woods. Maybe a simple rule is that unless the bad weather is something that truly reduces visibility to less than 25' or so, the minimum visibility is 1d10x5m (the rules as is allow better visibility in woods under poor conditions than in a swamp... the only one that might make sense there is fog over a swamp is likely thicker than the neighboring woods...).
 
A possible starting point

duckduckgo.com/?q=military+research+on+engagement+distance+in+terrain&t=fpas&ia=web

Found a couple likely looking pdfs. Also the overview type publications make a good start point for further reading and searches.
 
Yes, there is, even without considering moving quietly. Infantry in armies that expect to spend time in close country (bush, jungle, etc.) train in this. It's also why camouflage is a thing, and why snipers practice camouflage, concealment, and stealthy movement so much.
I'd be interested to hear what the training involves. How do you practice camouflage and concealment, for example.
Being stealthy in a non-tactical situation is in some ways even harder, especially in a group - people like to chatter, they joke and gorse around when bored (or grouse and complain if they're too tired for horseplay), they careless bang tins together, stomp around, leave litter behind, and so on. Learning not to do this stuff, and having the discipline, is hard for many people. Naturally laconic loners will find the not-talking bit easier, but everyone needs to learn the little details.
I might also argue this is a different skill entirely - social stealth vs non-social stealth.
I'd argue that stealth is a skill. Being able to walk quickly without making noise, knowing which floorboards are likely to creak if you tread on them and which aren't, being able to climb over things silently, and being able to hold completely still to avoid giving your position away by moving. It's more than just staying out of sight and tiptoeing around.
seems to me like most of those things can really be another skill entirely. I'm not sure I've ever heard of rolling stealth to hold completely still. it is mostly staying out of sight and tip toeing.
Nobody who's ever watched a Louis de Funes movie has ever doubted that stealth is a skill... one the main characters are lacking, usually:shade:.
I'm not familiar with these movies.

mind you, I'm asking and committing heresy on purpose here. I think it needs to be defined pretty well.
 
I've cured my players of saying "I go into Stealth" by telling them, "No, you do something stealthily. What are you stealthily doing?"

That's just for descriptive purposes, though, because, as far as Savage Worlds rules are concerned, Stealth is a skill.
 
Those rules for Twilight 2000 are pretty nice. I would argue that bad weather and night DO reduce how far you can see in woods. Maybe a simple rule is that unless the bad weather is something that truly reduces visibility to less than 25' or so, the minimum visibility is 1d10x5m (the rules as is allow better visibility in woods under poor conditions than in a swamp... the only one that might make sense there is fog over a swamp is likely thicker than the neighboring woods...).
I would consider halving ranges in woods and swamps in very poor weather because heavy rain or snow, high winds, etc., mess up hearing as well as vision. Otherwise, ruling that things never get worse than the 'woods' distances would a simple fix (and assume that anything detected beyond the distance the vision rules would've given is heard, not seen).
 
I'd be interested to hear what the training involves. How do you practice camouflage and concealment, for example.
You learn and practice putting on that lovely camouflage paint (makes your face come out in zits), and adding foliage to your hat, rifle, etc. You earn and practice muffling all the noise-making kits, and keeping it that way. You learn about avoiding sky-lining yourself, and to move with irregular spacing (especially when in plantations with regular rows of trees). Stay off the tracks if you can, move half-way up a hillside rather than along the top or in the valley (all this also means 'move slowly and burn a ton of energy doing so' - there's a reason people and animals put the paths on the ridge lines and by the streams). Learn how to move through thick bush without getting tangled up and making noise, and how to place your feet quietly and to recognise what you can put your feet on that won't make noise (nothing like standing on a log only to find it's rotten pine, and you make a massive CRUNCH sound as your foot goes right through it - and then you make more noise as you stumble round trying to get your balance back).

For dealing with known sentries, etc. - learn and practice moving when they're looking away. Learn to use 'dead ground' (areas that are out of line of sight). Practice not looking at sentries when you don't absolutely have to (people are very good at spotting faces, and also exposed skin is shiny). One exercise to practice this is to have an experienced NCO stand on a small platform or the back of a truck and direct another to where they think they've potted someone. Meanwhile all the someones try and sneak up to a point right by the SNCO through rough ground with bushes or tussock clumps. Those caught get some kind of incentive to improve (a bunch of press-ups, or a bit of a run, probably). Those who win get some minor reward. Those who just don't move fast enough, but don't get caught, probably just get noted for incentivisation if they don't do better next time.

I might also argue this is a different skill entirely - social stealth vs non-social stealth.
That's fine. One could also argue that being stealthy is part of some other skill, like say 'Fieldcraft' for soldiers and outdoorsmen (or lump into something like World Without Numbers' Sneak, which is all things stealthy and dodgy). I'm just saying it's definitely a learned skill.
seems to me like most of those things can really be another skill entirely. I'm not sure I've ever heard of rolling stealth to hold completely still. it is mostly staying out of sight and tip toeing.
And not moving when someone's looking right at you. Not making a sound when someone's quietly and suspiciously listening out for you.

It's quite possibly part of several skills, and looks different for each. Sneaking around alleys in a city is a bit different from hiding in scrub - more about being quiet and exploiting hard cover, less about spotting little bits of dead ground that'll let you crawl in safety.
 
I've thought about Stealth v. Perception a lot, and one conclusion I've come to is that it's not, or should not be, a directly opposed roll. Stealth determines the quality of your hiding, but being really good at hiding doesn't preclude someone who is really perceptive from making an exhaustive search and finding you. It's okay for something like DC Heroes where almost all actions are about the hierarchy of who is better, but it's a problem in something like a d20 game or BRP where characters have ratings based on probabilities. Something like Pathfinder 2 is kind of in the middle. It's also questionable how trainable innate perception is, but you can definitely train searching. Stealth is trainable, but ultimately good Stealth is based on when and how to Stealth, and using the environment and knowledge of the enemy to maximize your odds.

So in a system that allows it, I think it's better if Stealth operates by being successful or not, and then giving you grades of success. Innate perception should be fairly "flat" but should make some allowance for trained observation and search skills that involve some kind of routine and effort.
 
A possible starting point

duckduckgo.com/?q=military+research+on+engagement+distance+in+terrain&t=fpas&ia=web

Found a couple likely looking pdfs. Also the overview type publications make a good start point for further reading and searches.
I did a google search:


And found this really good one immediately:

 
I've thought about Stealth v. Perception a lot, and one conclusion I've come to is that it's not, or should not be, a directly opposed roll. Stealth determines the quality of your hiding, but being really good at hiding doesn't preclude someone who is really perceptive from making an exhaustive search and finding you. It's okay for something like DC Heroes where almost all actions are about the hierarchy of who is better, but it's a problem in something like a d20 game or BRP where characters have ratings based on probabilities. Something like Pathfinder 2 is kind of in the middle. It's also questionable how trainable innate perception is, but you can definitely train searching. Stealth is trainable, but ultimately good Stealth is based on when and how to Stealth, and using the environment and knowledge of the enemy to maximize your odds.

So in a system that allows it, I think it's better if Stealth operates by being successful or not, and then giving you grades of success. Innate perception should be fairly "flat" but should make some allowance for trained observation and search skills that involve some kind of routine and effort.
In Cold Iron, I do an opposed roll perception + chance adjustment roll vs. stealth. A chance adjustment roll is a roll on a cumulative normal distribution table than produces a value from -25 to +25 (the table COULD be extended past 25, I have a table that goes out to 40...). So it isn't each person rolls for success or not, it's simply, THIS TIME, who did a better job, observer or person trying to be stealthy? The ratings can be modified by situational things. Also, the margin of success can be used to modify the range at which detection occurs.
 
Banner: The best cosmic horror & Cthulhu Mythos @ DriveThruRPG.com
Back
Top