D&D 5E Surprise and Sneak Attack

[Greenstone.Walker]

Which can get downright peculiar if the assassin is acting on their own. If the assassin decides to make an attack on an unaware target, initiative is rolled, and the target is surprised but gets a higher initiative than the rogue, can the rogue on their turn declare "actually, I've changed my mind - I'm just going to stay hidden for now", ending the encounter with no actions having taken place? And can they then declare "okay, this time I'm definitely gonna go for it", initiating a new encounter in the hopes of winning the initiative this time around?

Begin and end with the fiction.

Something has alerted the target. Maybe a slight sound, perhaps a smell, maybe just some premonition of danger. What do they do? If they are a guard, maybe the call out for reinforcements. Maybe they start patrolling. Perhaps they draw their weapons and ready an attack. Maybe they cast a light source into the nearest patch of darkness, hoping to spot an intruder. Certainly they are now on their guard, and probably won't be able to be surprised again this watch/night/whatever.

If the rogue wants to initiatie a new encounter, they can, but it is going to first require resolving this encounter.

 

[FreeTheSlaves]

Heh, a twitchy assassin second guessing themselves. If I saw that 'will I/won't I' behavior I'd rule disadvantage on initiative checks until they get therapy or Greater Restoration.

Or roll NPC initiative behind the screen. No reason to know initiative until round 2.

 

[Paul Farquhar]

Yeah, the gloomstalker is a better assassin than the assassin in most campaigns.

That's because the assassin subclass, more than any other, isn't designed to be a monster slaying adventurer. It's designed for a character who wants to infiltrate the palace disguised as a servant and slip poison in the king's wine. It's not designed for someone who wants to sneak around dungeons backstabbing dragons.

 

[caudor]

I tend to agree with Flamestrike. A video on youtube might be helpful to watch: Top 5 Dungeons and Dragons 5e Rules Everyone Gets Wrong
As for RAW, there is no surprise round in 5e. Surprise is like a condition on a PC or monster which simply means you don't act on your initiative. At times, I still find that I've carried over a rule from a previous editions. I'm certainly no rules master; I have to rely on Crawford and sage advice (and others) when it comes to tricky issues. Rules aside, the most important thing is having fun.

 

[Paul Farquhar]

There is no surmise round, because some participants may be surprised and others not surprised, on either side. But that is tied to awareness - those who are unaware of the threat are surprised, and don't act on their turn, those who are aware are not surprised.

 

[Ovinomancer]

You're ruling 'insta-death' in a situation clearly covered by the rules.

I agree, and my ruling is within those rules. To address your concern -- who decides combat is occurring? The GM does. The GM decides if this action declaration results in combat, and then uses the combat rules to address this because combat is inherently uncertain (as you well note). If the GM decides combat does not result -- that the outcome of the declared action is already certain -- then you just narrate the outcome, per the rules.

This deprives Fighters and other martials of their core class feature of extra Luck (represented by their higher HP and HD) for absolutely no good reason.

This is more argument from jerkdom, not a rules issue. I can deprive any character of anything while strictly applying the combat rules just by using infinite dragons, so the deprivation isn't actually an issue of proper application of rules but a feature of being a jerk. I am not a jerk, so this isn't a valid argument about application of the rules. In point of fact, I point out above how I'd not reach this situation with PCs because I'm not going to use "rules' to have an NPC assassin sneak in and kill PCs in their sleep. Or, doing so would be at the end of such an outrageous set of circumstances that I have difficulty imagining it. I suppose if PC saw the assassin in their room and the player declares the PC goes to sleep anyway we might get here, but... I don't see that happening unless I'm utterly failing as a GM to begin with.

Instead, I flipped this onto ruling for PCs assassinating NPCs, at which point your argument is again mis-aimed -- I'm depriving my infinite set of NPCs their core class features? Nah, not an issue.

If you drop the argument from jerkdom, it would appear your primary issue with the ruling also drops. Or, at least, the primary focus of your counter-argument. I can believe that you might feel that combat is required anytime you get near violence, but that's not required. As I say above, it's the GM that determines if combat is occurring, not anything else. Violence can occur outside of "combat" and lethal or less-than-lethal results can also occur. Should that be normal? Eh, I guess that depends on your tolerance. It would appear it's far more normal in my game than yours, but it's far from common in my games. I allow for it, which it appears you do not, and I think that gives my more flexibility in application without once stepping outside the framework of the rules. That framework being that the GM determines uncertainty and when combat occurs and applies the mechanics as needed. If there's no need, I don't feel locked into using the combat mechanics just because they exist, especially in situations I wouldn't consider "combat".

But, please, please, please, do try to avoid argument from jerkdom if this goes forward. It's impolite to accuse others, even indirectly through example, of either being jerks or being incompetent to the point of jerkiness. It also doesn't actually support your argument, because postulating imaginary jerks as support isn't very strong of a position.

 

[Ovinomancer]

If I was playing a game and fighting an enemy, then I cast a spell to out that enemy to sleep, do I get to end combat early?

If I cast the sleep spell on a BBEG (no save) and use subtle metamagic to prevent counterspelling, do I get to win since we can kill him in one strike? If I can, I will. Not even trying to metagame, in real life, I would absolutely slit people's throats by forcing them to sleep if I was mortally opposed to them.

I agree we can skip the whole combat part if there isn't a chance they'd live within the rules. If you have a commoner sleeping vs a Level 20 PC, no need to roll initiative. It's basically impossible for them to survive.

If we're talking about a sleeping ancient red dragon, they get their autocrits, their deadly strike damage if the dragon fails, their sneak attack. But they do not instantly kill the ancient dragon.

I think this is fair.

Of course. The GM should be considering the totality of the situation and making the call. They shouldn't, however, feel that because there are combat rules that they've lost that authority to consider the totality of the situation. This isn't a replacement "always" rule. A sleeping target MAY be killed instantly, if the fiction leads to that outcome. It's not a "a sleeping target is ALWAYS killed instantly." This is a false dichotomy -- there's a huge range of middle grounds being instantly killed in their sleep here. For instance, I'm pretty sure that my application of the combat rules would be indistinguishable from @Flamestrike's most if not almost all of the time. Just because I allow for an outcome within the rules does not mean it's a go to or very common at all. ON the other hand, it does sometimes happen.

When? Well, let's take your top example -- sleep is cast, the target fails the save. The target is then tied up and secured. If the other side wants to kill the target at this point, I'm not going to invoke the combat rules and roll for them until it's done -- it happens, the target is dead. This has come up in my games a few times, where the PCs have captured an enemy and been unable to convince the enemy to offer their parole and so have decided to kill the captive. I did not invoke the combat rules in this case, as this wasn't a combat at all. On the other hand, in a fairly recent game, all of the PCs were incapacitated in a combat and were captured, tied to posts, and were being left as sacrifices to a monster their foes worshipped. The PCs tried to escape, with some being still tied, but since some had escaped and combat was in place, I used the combat rules even for attacks on the still bound PCs -- because I was in that tight resolution with a melee ranging around, there was no need to do otherwise. However, if, for some reason I cannot imagine, I decided that the NPCs would just outright kill the tied and incapacitated PCs, I wouldn't have gone to the combat rules -- I used those already when they were captured to begin with. The fiction was already in a place that the answer was pretty straightforward and obvious. And also very uncommon for how the game's fiction is positioned. Follow the fiction, rule uncertainty on the totality of the situation, and use your authority as GM (according to the rules, no less) to make the call as for success, failure, or uncertainty. If uncertain, apply the mechanical test most suited for the situation. This really shouldn't be controversial or require the postulation of imaginary jerks to dismiss.

 

[Maxperson]

I agree, and my ruling is within those rules. To address your concern -- who decides combat is occurring? The GM does. The GM decides if this action declaration results in combat, and then uses the combat rules to address this because combat is inherently uncertain (as you well note). If the GM decides combat does not result -- that the outcome of the declared action is already certain -- then you just narrate the outcome, per the rules.

The uncertainty rule is for ability checks, not combat. It appears nowhere in the combat section, but does appear in the ability check section. If you want to apply it to combat, you have to engage a house rule(rule 0).

 

[Sir Brennen]

Scenario 1) The Guard goes first, he's surprised (cant move or act on his turn). He then stops being surprised and the Assassin makes his attack roll (no critical hit).
Scenario 2) The Assassin goes first, the Guard is surprised (critical hit)

Exactly this. Just because the assassin lost initiative on the first round doesn't mean he doesn't get to act first against surprised targets.

This also covers the sucker-punch, though it may require a bit more GM adjudication. If all parties are aware of each other, but combat does not seem imminent, I'd probably make the PC wanting to initiate a "surprise" attack, like the fighter suddenly lunging across the dinner table, or rogue with a knife behind his back, make a Deception check (opposed by Insight/Wis) instead of the regular Stealth to determine who's surprised. Just like Stealth, the opposed roll would be passive if the target has no particular reason to be wary of an attack at that moment. (Attack at dinner, passive. Enemy who's called for parley, active.)

 

[billd91]

Someone seeking to do harm on someone else is combat. It's structured for a reason. By arbitrarily ignoring those rules, and that structure you're deliberately punishing classes and characters that are good in combat, and have made choices based on that structure.

Sounds like an overly broad concept of combat to me. Appropriately adjudicating whether or not the intended action should be an auto-success, auto-fail, or should be rolled doesn't punish classes and characters that are good in combat.

The section on Using Ability Scores starting on p237 of the DMG is reasonably illuminating in this context since it goes into when it might be appropriate to roll ability checks, attack rolls, and saving throws and when it might be better to just rule success/failure.