D&D 5E Assassinate

[Arial Black]

Thanks for the opportunity to correct some mis-conceptions about 'version 1'.

Arial's surprise (what we're calling surprise version 1) adds words to RAW and ignores RAW in several places,

It kind of adds words, in that both of us are extrapolating things from what little is written, but version 1 doesn't actually break any of the written rules.

but I will focus first on one example. Specifically, it adds the words "surprise ends on noticing a threat" to RAW. Those words do not appear in RAW.

Agreed. Just as much as version 2 adds the words "surprise ends for a creature when it can take reaction". Neither version is written, so that's not a valid way to choose between the two versions since they both add words.

And in adding those words, surprise version 1 ignores other RAW that makes it clear that surprise confers to individual combatants. It instead necessitates that surprise is tracked to pairs of combatants. (Working through multi-way combats demonstrates that. Multiple assassins can lose surprise at various times to various targets.)

This is a mis-conception. A creature is surprised if it doesn't notice any threat, and is not surprised if it notices any threat!

Imagine there are five threats. A creature is only surprised if he fails to notice every threat. As soon as he notices any single threat, then he is not surprised, considers himself to be under threat, is on the lookout for danger just as much as anyone who is not surprised.

So you don't have to track every pair of assassins/victim possibilities. As soon as a creature notices a threat, any threat, it is not surprised.

Surprise version 1 benefits ranged-assassins and increases the power of the Skulker feat, while penalising melee-assassins who will need to close without being noticed. The streamlined base rules of 5th edition don't include facing, so a DM might decide to adopt the optional Facing rules from the DMG to mitigate that. Albeit any melee-using creature closing from behind will then get advantage, making the first part of Assassinate meaningless for melee-assassins.

It's the other way round. Version 1 gives the correct weighting to the Skulker feat and ranged assassins, while version 2 lets melee assassins run 30-feet across open ground toward watching opponents and still treats them as oblivious to danger.

A possible defence for adding "surprise ends on noticing a threat" to RAW is if we decide that the first-turn effects of surprise expressed on PHB189 are not synonymous with being surprised.

Exactly! They never were synonymous. One was the cause, the other was one of the effects.

If so, then RAW does not tell us when surprise ends so we require additional words.

Again, exactly this! The RAW does not tell us when surprise ends, so we must extrapolate from the words that are written.

Whatever words we choose can't be justified by our wanting surprise to continue past its first turn effects in order to serve as a handshake for features such as Rogue Assassinate and Kenku Ambusher. Because that commits the fallacy of begging the question (our premises directly or indirectly include the claim that the conclusion is true). Additionally, ending surprise on noticing a threat can result in surprise ending prior to its first turn effects completing, necessitating that we launch those effects before participants start taking turns! A baroque addition to the mechanic that falls afoul of Occam's Razor.

With 'version 1', creatures are surprised until they are not. Since surprise = 'not noticing a threat', creatures don't notice a threat until they do notice a threat. See how that makes sense?

What a creatures reaction speed may be has no relevance before they even start reacting! A sprinter may have exceptional reaction to the starting pistol, but this is of no help if he never even hears the pistol fire!

One way to notice a threat is to take damage from an attack. This means that being hit by that attack ends surprise for that creature, but only after that attack is resolved, so Assassinate auto-crit still works for that attack.

There are many other ways to become aware of a threat. The DM's job is to make adjudications like this. Sometimes the creatures become aware of a threat early, and sometimes the never do. But the timing of their awareness of a threat is tied to their ability to sense things, as it should be. One of the problems with version 2 is that the awareness of danger is tied to a mechanic that has no relation to sensing things.

 

[clearstream]

This is a mis-conception. A creature is surprised if it doesn't notice any threat, and is not surprised if it notices any threat!

Exactly. That was the clarification I expected.

It's the other way round. Version 1 gives the correct weighting to the Skulker feat and ranged assassins, while version 2 lets melee assassins run 30-feet across open ground toward watching opponents and still treats them as oblivious to danger.

That remains wedded to perception ending surprise and thus begs the question. If surprise is indeed flat-footedness, then yes - an assassin who wins initiative is able to close without losing the benefits of surprise because noticing doesn't matter after it imposes the condition. That in turn explains the need to win initiative very nicely. You have to get the jump on your target.

Exactly! They never were synonymous. One was the cause, the other was one of the effects....

Again, exactly this! The RAW does not tell us when surprise ends, so we must extrapolate from the words that are written.

Once you decide they never were synonymous, as you have, you need extra words. But if you instead decide that they are synonymous, you do not. To see how that works, assume for the sake of argument that surprise and the effects of surprise actually are synonymous. You cannot be surprised and not suffer its effects, and you cannot be suffering its effects and not be surprised. When does surprise end, given that assumption? Still, your points show that we can improve the clarity of surprise version 2, and added to your earlier clarification about the delays, we could do with restating both versions.

Surprise N (for ends on noticing)
Applied when "not noticing a threat" and ended upon "noticing any threat". Upon first being applied, surprise N launches an effect that delays actions and reactions until the end of a combatant's first turn. Until it ends, it acts as a handshake for other effects. For avoidance of doubt, ending surprise N does not end the delay-effect even if it comes sooner.

Surprise F (for ends with its first-turn effects)
Applied when "not noticing a threat" and runs until the end of a combatant's first turn. While active, surprise F prevents actions and reactions and acts as a handshake for other effects.

Interestingly, the consequences of the delay-effect and the prevent have converged (neither does anything mechanically after the end of a combatant's first turn). Launching the delay-effect is only required by surprise N because of the possibility of surprise N ending before its first turn effects end (which should be a clue that somethings up!) Working from RAW, surprise is about being able to act and react. For instance, the Basilisk text calls out that creatures that aren't surprised can avert their eyes. And Shield demonstrates that a last-second reaction can be fast enough to undo a hit. A target capable of reacting can curl around the blade, twitch aside from the arrow, or whatever - just enough to avoid the auto crit. The Paralyzed condition gives us further insight - "any attack that hits the creature is a critical hit"... Being surprised has similarities with being briefly paralyzed. And Assassinate grants rogues the ability to exploit that through subtlety (unnoticed before the encounter starts) and speed (winning initiative).

 

[Arial Black]

Conceptually, the game assumes that participants in combat are constantly on the lookout for danger. There are no rules for facing because is assumed that combatants are constantly looking around to be able to respond to the events around them.

So the surprise mechanic is the crunch expression of the fluff that surprised creatures don't realise that they are in combat, are not constantly looking out for danger, and are just living their lives, walking around, eating snacks, reading books, whatever. The consequences of this lack of awareness of any threat include that they are slow out of the blocks, even when they finally notice a threat. It is also the basis for the concept of the Assassin's auto-crit: it works because the target is unaware, not looking round for danger, and still eating snacks and reading books, totally oblivious to danger and therefore not throwing themselves to one side at the last minute and not making themselves a difficult target, which they do start to do as soon as they realise that they are under attack!

Example:- A wizard is engrossed in his books, totally oblivious to the assassin and his bow and poisoned arrow. The assassin draws the bowstring while trying to remain as quiet as possible, and the wizard doesn't see or hear him. The wizard is defenceless; the perfect Assassination target.

In both versions, because the assassin beat the wizard in every contested Perception/Stealth roll, the wizard starts combat 'surprised'. Why? Because he doesn't notice a threat.

Initiative is rolled, measuring reaction speed not 'awareness of a threat'. The wizard has a higher initiative result than the assassin. In both versions, the wizard cannot act/react until after his first turn.

In version 1, the fact that the wizard has faster reactions has no impact on whether or not he knows that there is anything to react to. The assassin looses the arrow, and if it hits it will auto-crit. Why? Because the wizard is defenceless, oblivious to danger, still engrossed in his book.

In version 2, after the initiative count goes past the wizard's initiative total....nothing at all is different about the situation, the wizard is still oblivious, still engrossed in his book, but somehow he is as hard a target as a fully-aware combatant, and immune to auto-crits from Assassinate. Why? He still doesn't know about the still hidden assassin.

Version 2 makes the oblivious wizard somehow no longer oblivious despite not detecting the assassin in any way, because an ability check totally unrelated to awareness was used to end 'surprised'. It makes as much sense as ending surprised if you roll well on a Religion check.

 

[sithramir]

That wizard clearly had one level of ranger from the latest UA article!

 

[clearstream]

...the surprise mechanic is the crunch expression of the fluff that surprised creatures don't realise that they are in combat, are not constantly looking out for danger, and are just living their lives, walking around, eating snacks, reading books, whatever. The consequences of this lack of awareness of any threat include that they are slow out of the blocks, even when they finally notice a threat. It is also the basis for the concept of the Assassin's auto-crit: it works because the target is unaware, not looking round for danger, and still eating snacks and reading books, totally oblivious to danger and therefore not throwing themselves to one side at the last minute and not making themselves a difficult target, which they do start to do as soon as they realise that they are under attack!

...after the initiative count goes past the wizard's initiative total....nothing at all is different about the situation, the wizard is still oblivious, still engrossed in his book, but somehow he is as hard a target as a fully-aware combatant, and immune to auto-crits from Assassinate. Why? He still doesn't know about the still hidden assassin.

Unfortunately that is a retreat to a preconceived fluff-based objection, rather than resisting any of the arguments about the mechanics arising from RAW. Let's instead - for the sake of argument - say that being surprised by an expert assassin who beats your initiative means she is able to treat you as being partly paralysed. Conversely, if you beat her initiative then that indicates that you are able to react to her hit sufficiently well to mitigate an auto-crit. You fold over her knife, jerk away from her arrow, whatever. The Basilisk creature description, and the Incapacitated and Paralyzed conditions support that narrative. And the Shield spell demonstrates that given the ability to react a hit can be turned into a miss. Given the ability to react, an auto-crit becomes a normal hit. With that narrative we're done, right? Not at all - we need to address any substantive points about the RAW.

Note BTW that "he is as hard a target as a fully-aware combatant, and immune to auto-crits" is a red herring. You might as well say "why does my roll of 8 miss this hide-armoured combatant just as if he were plate-armoured!?"..."why is he immune to my attack!?". The wizard isn't immune. He just happens not to be susceptible on this occasion. In particular the example commits the fallacy of finishing the narrative before rolling the dice. You've decided that no matter what he rolls, the wizard is just as fully oblivious and engrossed. But if we adjust our narrative to our dice rolls as we should, then we can describe that if his initiative roll is low he's engrossed, nodding off, clearly not going to react, while if it is high he looks fidgety, alert, less engrossed, clearly capable of reacting.

Example:- A wizard is engrossed in his books, totally oblivious to the assassin and his bow and poisoned arrow. The assassin draws the bowstring while trying to remain as quiet as possible, and the wizard doesn't see or hear him. The wizard is defenceless; the perfect Assassination target.

This contains a convenient omission. "the wizard doesn't see or hear him AND is caught utterly flat-footed" i.e. the rogue beat his passive Perception and beat his initiative.

In both versions, because the assassin beat the wizard in every contested Perception/Stealth roll, the wizard starts combat 'surprised'. Why? Because he doesn't notice a threat.

This contains another omission. It should go on to say that - according to RAW - once in combat even if the wizard notices the assassin prior to his first turn, he remains "surprised". The only time noticing is checked is before the combat begins. That is unambiguously stated in the call-out box on PHB189. RAW does not say that being surprised can be alleviated by noticing the threat once in combat.

Initiative is rolled, measuring reaction speed not 'awareness of a threat'. The wizard has a higher initiative result than the assassin.

According to RAW winning initiative is enough. Being on this occasion reactive enough compared with the assassin mitigates the auto-crit. But for the sake of argument let's say your version of surprise is correct. Why is winning initiative good enough to get rid of Assassinate's advantage effect but not the auto-crit? How is it that our perfectly unnoticed assassin, beaten by the wizard's initiative roll, somehow loses advantage? Our totally oblivious wizard is somehow just sufficiently aware to avoid giving up advantage, yet just sufficiently unaware to avoid the auto-crit!?

 

[clearstream]

UPDATE - I found this tweet from Mearls from 2014!

Doc Oc [MENTION=6801278]grazel[/MENTION]
[MENTION=32417]MikeM[/MENTION]earls Seeing a discussion and want an official WotC answer. How long does the suprised status last (due to an assassin feature)

Mike Mearls [MENTION=32417]MikeM[/MENTION]earls
[MENTION=6801278]grazel[/MENTION] end of creature's first turn - roll initiative as normal, and once surprised creature first turn has passed it is not surprised

 

[Paraxis]

UPDATE - I found this tweet from Mearls from 2014!

Doc Oc [MENTION=6801278]grazel[/MENTION]
[MENTION=32417]MikeM[/MENTION]earls Seeing a discussion and want an official WotC answer. How long does the suprised status last (due to an assassin feature)

Mike Mearls [MENTION=32417]MikeM[/MENTION]earls
[MENTION=6801278]grazel[/MENTION] end of creature's first turn - roll initiative as normal, and once surprised creature first turn has passed it is not surprised

Yeah I have quoted him in this thread multiple times and linked to the sage advice page.

There are people who just don't care and dismiss his views on the matter because he is not Crawford, and in a way they are right Mearls himself has said he is not the rules guy and that his tweets are about how he would handle things at his table not necessarily how the RAW is in the books.

Surprise doesn't have a clear end time, but to many people including myself and Mr Mearls the best time to say a creature isn't surprised is after their first turn is over and they can take reactions.

 

[Arial Black]

I've seen this tweet before, and others from Mearls on various subjects. They nearly all illustrate why Mearls in not the rules guy.

He frequently gives the opposite answer to the same questions asked to both him and Crawford, and sometimes even contradicts himself!

I have no confidence in his rulings matching RAW; they are, as he says himself (usually when it's pointed out where he's gone wrong), just the rulings that he'd make at his table.

A tweet from Crawford would be more influential. It must be said that some of his tweets have been unintentionally ambiguous, with both sides of a debate claiming that the same tweet supports their case! I think the 140 character limit is the main culprit here.

Also, our single thread on this subject has over 500 replies, many with lengthy posts laying out detailed arguments with citations and examples, and even we don't agree!

Even if the designers read through the whole thread (never mind other threads on the same subject), it would be hard to sum up in 140 characters. But that is not what happens. What happens is that someone tweets them 140 characters, without 500 pages of debate attached, or even one post attached which would lay out the argument.

I understand the urge to get the immediate response that Twitter offers, but find it a limited medium to get solid answers to difficult or subtle questions.

As to why I'm not satisfied with the 'until his first turn is over' solution: it doesn't make sense that you notice something hidden just because time has passed, it is not written in the rules that surprise ends just when one of its effects has occurred, and a better solution which doesn't have any drawbacks follows right on from the written rule of what causes surprise, 'not noticing a threat'.

Thanks for the debate. The rules become clearer for me as I think about how to respond to the case. There have been occasions where my mind has been changed because the case I was answering led me to an understanding that was new to me. The way actions are used in 5E was one of these occasions, which led to me changing my stance of the timing of Shield Master bonus actions because of that new understanding.

This hasn't happened on the subject of when surprise ends. I don't expect it will. If Crawford himself publishes a Sage Advice article on the subject, then I'll have to go along with it whether or not I think the ruling is a good one.

 

[Noctem]

Surprise only affects a creature until the end of their first turn. After which they can take a reaction as normal. The assassin feature let's you gain a benefit against creature's that have not acted yet during an encounter essentially.

 

[Arial Black]

Surprise only affects a creature until the end of their first turn.

Can you quote where the book actually says that? If it did, this thread would've ended over 500 posts ago.

After which they can take a reaction as normal.

Being surprised at the start of combat means that your first action/reaction will be delayed. There is nothing tying 'the ability to take reactions' as synonymous with 'not surprised', nor 'the prevention of reactions' being identical to 'surprised'.

The cause is 'surprise'. One of the effects is to delay your first action/reaction. The effect doesn't determine the cause, nor when the cause begins and ends.

For example, getting caught speeding is the cause, paying a fine is an effect. But the moment I pay the fine has absolutely no impact on when/if I stop speeding.

The assassin feature let's you gain a benefit against creature's that have not acted yet during an encounter essentially.

The feature in question grants two abilities, each with it's own timing.

One ability works until the target's first turn.

The other ability works while the target is surprised.

These are two different timings. If they were identical, they wouldn't be written in two different ways and it would save space to have one timing governing both abilities. The reason they have two different wordings is because they are two different timings.