Stealth System Concept
Intro
Stealth sections in DND have never worked well for me. They always seem to fall apart the moment they start, and you lose that power fantasy of "super stealthy rogue" that the whole class is based around. While nobody at my table has played a rogue for a while, and I doubt we will any time soon, I got thinking about how we might be able to solve this problem. So I came up with a possible solution: Prep Time.
Setting the scene
For starters lets outline the problem. For the sake of this thought experiment we're going to imagine that your rogue is the only one sneaking, not the whole party. We'll cover that later. Additionally, this take is about personal preference. If you like how stealth works in DND, that's great! This is just a hypothetical solution to a problem I personally have. Your mileage may vary.
So your rogue, a master thief or assassin or what have you, want's to sneak into a building. Maybe they want to steal the scepter of a local lord, or have a... private conversation with a gang leader. So they sneak into a building, you pull out a battle map, and now they have to avoid a whole hurdle of guards, constantly rolling stealth and sleight of hand to get inside.
Why is this a problem? For me it's two fold. For starters, having the rogue roll multiple times throughout the engagement means that the risk of failure goes way up, punishing them for engaging with something that their class is supposed to be good at. It turns security into a numbers game: It doesn't matter how good or bad your guards are at their job, have enough of them and someone will notice. It makes it incredibly difficult to make stealth sections distinct from one another, as it becomes almost purely chance. If the rogue rolls high but the guard rolls a naturally 20? Shit out of luck.
And honestly? It takes a long time, it's not particularly engaging, and it's very repetitive. I also think it turns rogues into a bit of a one trick pony.
Stealth Mission
My solution to this problem makes a core assumption: As a rogue, you don't need to prove you can sneak into a place. So instead of playing out each individual moment of stealth, we break it into a single roll.
In this system, a bastion or stronghold, or whatever the rogue is trying to break into, is given a DC. If it's a large military compound for example, it could be DC 15. A small tented area? DC 10. The DC here represents how difficult it would be for the rogue to slip in and get to a pre-determined location unnoticed. Want to get to the tent of the leader of a small merchant guild? DC10 Stealth check and you're in unnoticed. What do you want to ask them?
This serves two purposes: Reduce the number of rolls a rogue can fail (or number of perception rolls a guard can succeed), and decrease time spent in stealth sections. By setting a DC, we also change how different areas feel to break into.
Now, a flat roll might speed up the game, but it's a little boring, and potentially very swingy. If the rogue rolls low once that's it, they're caught?
Prep Time
By setting a flat DC, we open up the ability to modify the difficulty of breaking into or reaching certain locations. Maybe you don't know the layout of the location you're breaking into, or where a specific room can be found. Maybe you're unfamiliar with the patrol of the guards, or what you need is hidden behind a locked door. This opens us up to many more possibilities.
Let's say you're trying to break into a DC20 military compound. If you have +5 stealth, that's maybe a 25% chance of success. Not great odds, all things considered. Of course, your party might be able to bless you, or cast enhance ability. All well and good, unless you need to be in there for more than an hour, as you also need to be able to get out afterwards.
That's where prep time comes in. The rogue can spend time before hand researching and tracking down details about the location they wish to enter before hand. Maybe they bribe a former guard to draw them a map. Maybe they watch the outside of the location, tracking guard routes and shift changes. Maybe they ask another thief who has already broken in once before what kind of difficulties they might find.
This information doesn't have to be correct, mind you, (which can come up later in session) but for each day the rogue spends researching their target, maybe you reduce the DC by 2-3. Now you as the DM decide how much information is available, and by how much the DC can be reduced, but if the rogue spends three days, it could be the difference between a DC20 check and a DC14, or even DC11.
Different bits of information might offer different odds, that's sort of up to the DM to decide. Additionally what this does is encourage the Rogue to think outside the box and interact with NPC's in the local area, which can in itself lead to more roleplay opportunities. Of course if the goal is to reduce the amount of time used on stealth sections, you can always hand wave and simply allow them to spend the prep time, but it provides with more options.
Adding to the system
Now, some places in a compound might be unreachable by normal means. Maybe there's a door protected by an Arcane lock, or after getting past a certain point, the way becomes unknown. In these sections it could be worth breaking out the battle map, or handling stealth on a moment by moment basis. Of course, there are fun scenarios that can happen by playing every moment out. We don't want to lose those entirely. But I think they'll feel even better when they're not sandwiched between minutes of monotony on end.
Another bonus of the system is that the difficulty changes based on how much available time the players have. Given a week, the rogue could probably break into any reasonably large "normal" prison or garrison for example. Under tight time constraints, suddenly that task becomes significantly more daunting. This allows you, the DM, to control how difficult a task may be overall, in a way that feels natural given the circumstances of the story.
Involving the rest of the party
It should be said that this system is not intended to replace a dungeon crawl or full party encounter. Stealth tends to fall apart completely when the whole party is involved anyways. Instead, it's meant for those moments where the players need to retrieve some information, break into a simple vault, or otherwise utilize the talents of a rogue not unlike how they might utilize the research-abilities of a wizard.
If you want to sneak a whole party in somewhere, then you're probably better letting that play out in its entirety, or at the very least running it as a group stealth section instead, with optional moments of heroism here and there to provide interest/challenge.
Conclusion
In short, if you want to give your rogue the opportunity to feel like a master of stealth and subterfuge, consider giving them stealth missions with to locations with variable DC's, and allow them to modify the difficulty by spending in-game-time researching the location they're trying to infiltrate.