I've been thinking alot on how I was planning to implement stealth lately. Personally, I hate the whole "push button -> invisible" way that stealth is done.

However, due to the technical limitations, "true" stealth is pretty difficult to achieve, either because of model edits client side, camera angles, boosting contrast on teh monitor, etc.

I liked the way it was being implemented in Stargate, with you having a stealth 'HP bar' that diminished over time based on the people around you, etc. I was also thinking about eh stealth cover system in WANTED, and came up with a basic design, however it still has some issues.

I thought at first, if you were considered in stealth, you would be completely invisible to everyone. Things like noise, skills, etc. would increase your chance of stealth breaking early, but the basic idea is that you could pass behind someone unnoticed.

If you were in the 180 degree forward arc, you would remain in stealth as if youwere behind them, so long as you remained out of their los (walls, furniture, doors, etc.) however if you were in the 30 degree arc (directly in front more or less), your stealth would diminish rapidly.

However, someone brought up that a user that thinks a stealthed player is nearby can simply rotate slowly, and more or less AoE break stealth. The basic purpose of stealth in game is literally to sneak around, not for combat reasons (aka, no backstab abilities).

Do you think the current system would work alright in that context? or do you have any ideas to tag on in dealing with the system? I'm currently at a loss, so any help would be greatly appreciated.

(Sorry for my writing being sloppy and generally blah, it's 2 am and I have to get up in 3 hours, brain not working so great.)

That's actually doesn't sound too bad if you can figure out how to properly balance it.

First question do parties have shared sight.   For instance one person seeing a stealthed person and the rest of the party not.   There are quite a few fun/frustrating mechanics you could add with that.

Is it possible to hide in plain sight.   IE: insanely easy to not be noticed if you aren't being looked for.

I'm not sure, I'd imagine there would be no automatic way to share sight, forcing them to communicate.. but then again maybe that's asking too much sometimes.

I'm pretty against the hide in plain sight, however since most of the game is in buildings/streets, there's lots of LOS cover. Basically if it were timed right, you could run say.. from table to table, without breaking stealth.
 
I guess I'll just have to playtest it and go from there.  =O

Sounds like an OK consept to me.

As for the group sight. If you make stealth similar to a HP bar, than you have pretty much solved this issue. If one person from the group spots you, your stealth-hp will drain to zero and you will become visible to all. You can than choose to hide behind some LOS object to regenerate you stealth-HP and attemp to re-stealth, but your enemy will have a much better idea by that time about where to look for you.

This kind of situation seems reasonably realistic to me. It means that even tho sneaking past a group will be harder than a single enemy (because of their combined field of vision), it is still possible to get past them with sufficient cover and player-skill.

For example, its probably stupid for a stealthed player to run past 2 enemys looking at him, his stealth-hp would deplete very fast. If he times his movement to a moment when both of the enemies are looking somewhere else, hed be safe, because of the back 180 degrees being safer than the front. This makes it reasonable to add some sort of distraction options: throwing small rocks to create noise in a favourable direction, etc.

As for the "rotating slowly" stealth breaker: you can balance around it. The "rotation-slowly" will be made impossible, if you balance the speed at which the stealth-hp depletes and the speed the stealthed player can move at. If the stealthe player has sufficient movement speed to move from hiding place to hiding place faster than the rotating-player can drain his stealth-hp that the rotating will become a problem for the rotater himself. And anyway, it makes sense (to me atleast) to enforce the stealth players to utilize cover. It does not make sense to me to be able to score stealth attack on an open field. Springitng from behind a cover and attacking your enemy before they have a chanse to turn around - ok, doing same in an open (coverless) field - not ok.

My 2 silver on the subject.

Stealth is best applied in a static first-person viewpoint. You can't see behind you, simple, and problem solved without having to resort to mechanics that defy logic... unless it's a sci-fi and you can get away with it. Stealth, to me, in this fashion would be a way to hunch down a bit while walking, slowing pace but muffling footsteps. Because in first-person the only thing that will give you away is sound.

I also think the horror genre is best suited to first-person, but that is another thread entirely.

 I can't say I thought what I'm about to say through, but what if stealth was a skill that players had to "spec" into?  I always liked the idea of "situational" stealth.  In other words what if stealth was a skill that depended on the enviorment.  Forested areas, urban areas, aquatic, snowy areas, the ability to stealth better during the day or at night.  What if moving in stealth started out painfully slow and players could take skills that allowed them to move faster...Players with access to stealth would have to choose what envorments they wanted to stealth in and play thier game accordingly.  

Now you could vary this however you wanted, if you wanted to say that every player with stealth could do it in every enviorment, but "spec"ing would give you better chances, or longer stealth times.

From my point of view this forces players to make choices about how they play their character in the context of the game and of thier character as a whole.

Having said that, I really like the idea of Stealth HP....

So, how I read it, everyone has this constant limited-ranged cone attack of "stealth eating." I like that idea. It seems like it could work pretty well in a third-person game.

If you were to use a first-person stealth mechanic, a system I find that is pretty well balanced is from the Spy in Team Fortress 2. You can't attack while in stealth, people can hear your footsteps, and if you bump into them you become partially visible to everyone until you're no longer colliding with an enemy's bounding box.

If you uncloak, people can hear it, and you can't attack / activate anything until your cloak is fully deactivated (takes about a second and a half).

Pretty neat stuff, stealth systems can be.

I prefer the ones that enforce some amount of strategy and timing (moving from one hiding place to another before the cloak runs out / an enemy sees you). Also stealth systems that aren't always about getting the first strike - like the ones where you can go about poising yourself for an advantage based on terrain or sabotage an enemy outpost.

Like the recon class from BF2142. You can sneak around and find a good sniping spot or get into the enemy base... but you always emit a high-pitched screech (can be heard 8ft away) while cloaking, and it takes time to switch from the cloak to a weapon, so it's not the best thing to use for a first strike.

It really depends on how you want your game played. Do your sneaky people have the ability to go toe-to-toe with other classes, and stealth is just a way to get the first hit in? Or do they rely on other strategies, such as long range, deception, and/or sabotage to vanquish foes?

You know, I've never seen stealth applied as something that has to be "built up" to use. I mean it's always a given on a timer that cooldown and goes back to the static course, I've never seen one that needs to have a bar filled to use it properly.

All you need is a proper route to filling that bar; killing people, standing in a shadow/light to recharge it, absorbing it from vincinity of teammates to promote close-group play. I can sit here all day and figure ways to tie it in, but I think the idea itself works well. You would use stealh far less often, and usually in clinch situations to flee or gain an early advantage, and from there you better figure out how to get it back. I think I am gonna look at a couple of my GDDs and see which ones suit a method like this being employed.

A method such as this requires effects in stages, though for fast paced-FPS types I would recommend having it be full to use so that it's used far less often and treated as a trump card. If the game has a little reaction leniency, or combat relies on server timing, it suits having a tiered setup where it needs minimum 20% of the bar filled to use, but works poorly, and the effects are better and last longer the fuller it gets once used.

If you were in the 180 degree forward arc, you would remain in stealth as if youwere behind them, so long as you remained out of their los (walls, furniture, doors, etc.) however if you were in the 30 degree arc (directly in front more or less), your stealth would diminish rapidly.

This is exactly the implementation I've been exploring.

I imagine the following:

For me. Stealth is a binary condition its on or its off.  

The stealthed actor is only visible when directly in front of the observer.

Movement rate and surroundings affect sound but do not affect visibility.  

Weapons use affects stealth from sound and muzzle flash, but the observer would still need to "center his sights" to see the stealthed actor.

For example the sound of a player running past may warn someone of a pressence and cause an observer to begin scanning.  As a counter measure the stealthed player can adjust his movements to remain out of the LOS.

I'd been exploring the use of projectors with the unreal engine.  Having a tagged projector that projects only on a stealthed player with a set FOV. Otherwise, the stealthed actor is invisible.

Sort of like having every player walking around with an infrared flashlight that only lit up on a stealthed actor that entered its field of view. You could adjust for appropriate effective distance or widen the angle as a temporary toggle when "stealth aware".

This would allow for losing a pursuer by removing oneself from his FOV and escaping his scan... a bit of a skill element on both sides.  

For balancing issues:  Armor might make more noise and impede speed.  The combination of speed and silence becomes important for evading someone's scan. Weapon noise and muzzle flash would contribute as well. Awareness range and angle might be increased with an energy expenditure.  "Lit" actors might be visible to other players.

Zone design can have stealth hot spot and cold spots.  Long narrow hallways would force the stealthed into the narrow FOV. Wide open spaces would make avoiding scans more difficult. Bright lighting between actors might obscure the projection and cover allows for the stealthed actor to "lose" his pursuer.

.