For what it's worth, the probability an undetected invisible assailant (advantage on initiative) wins initiative against a single surprised opponent (disadvantage on initiative) with the same bonus to initiative is only 81.58%. So almost 1 in 5 times, a single ambushed enemy goes first.
This gets worse the more surprised enemies there are. P(ambusher with advantage goes first | all combatants have the same initiative modifier):
2 surprised opponents: 66.56%
3 surprised opponents: 54.30%
4 surprised opponents: 44.30%
5 surprised opponents: 36.14%
6 surprised opponents: 29.5%
So even with just 3 opponents, you've only got about a 1 in 2 chance of going first as the ambusher. Against typical party sizes, you expect at least one of them to act before you.
If any of the surprised opponents has a higher initiative bonus than the ambusher, these odds drop even further.
Given a party of adventurers is generally 4-6 players, it isn't a rare occurrence for an ambushing enemy to be beaten on initiative. It's actually the expected result.
(Method: generated the probability of any given result for both 2d20 drop highest and 2d20 drop lowest. Computed the probability of each die pairing by multiplying the vectors and forming a matrix. multiplied all elements of this matrix by the corresponding elements of a matrix of the same size that was 1 wherever 2d20 drop the lowest > 2d20 drop the highest (which is a triangular matrix with a diagonal of 0s, and 1s above it), then summed all squares of that final matrix).
There's no time machine. When action initiate combat its not resolved before Initiative but after in turn order, as explained in the DMG which can still be anticipated. Wether you're Invisible or surprised by combat starting doesn't make any mention changing that fact, all it does it say is affecting Initiative with Advvantage & Disadvantage respectively.
So when the narratives transition to violence everything pause to follow combat steps by steps how combat unfold; 1. Establish Positions 2. Roll Initiative 3. Take Turns
At this point you know you're surprised by combat starting with Disadvantaged to Initiative, while the Invisible creature have Advantage to Initiative.
You still can't see anyone Invisible who's action initiate combat which you anticipated somehow. How much is revealed at this point is up to DM.
There's many ways this can be described the rules don't elaborate on this. Did you see, heard, smelled or otherwise sensed something moments before the attack? Or suspect something due to lack thereof from unusual silence. Perhaps birds took off right before the attack following a noise. Ambusher jump out of the woods about to attack etc The more you fill the blank, the more narrative support to rules mechanic there will be.
That is essentially what the Devs were saying in the Dragontalk: Sage Advice when suggesting to come up with narrative ways to describe it when the D20 throw you a curve ball in unexpected ways.
There's no time machine. When action initiate combat its not resolved before Initiative but after in turn order, as explained in the DMG which can still be anticipated. Wether you're Invisible or surprised by combat starting doesn't make any mention changing that fact, all it does it say is affecting Initiative with Advvantage & Disadvantage respectively.
So when the narratives transition to violence everything pause to follow combat steps by steps how combat unfold; 1. Establish Positions 2. Roll Initiative 3. Take Turns
At this point you know you're surprised by combat starting with Disadvantaged to Initiative, while the Invisible creature have Advantage to Initiative.
You still can't see anyone Invisible who's action initiate combat which you anticipated somehow. How much is revealed at this point is up to DM.
There's many ways this can be described the rules don't elaborate on this. Did you see, heard, smelled or otherwise sensed something moments before the attack? Or suspect something due to lack thereof from unusual silence. Perhaps birds took off right before the attack following a noise. Ambusher jump out of the woods about to attack etc The more you fill the blank, the more narrative support to rules mechanic there will be.
That is essentially what the Devs were saying in the Dragontalk: Sage Advice when suggesting to come up with narrative ways to describe it when the D20 throw you a curve ball in unexpected ways.
Yeah, except the dragontalk assumes 2014 surprise rules, where the surprised characters don't get to act in round 1. Totally different, because at least if they're not acting, they don't have a time machine and it doesn't disturb causality, and not at all applicable to the 2024 rules. (Nor does it help that there's no narrative way to handle the unobserved invisible attacker or the subtle spell sorceror - if they're delayed for some reason, combat simply doesn't happen until they do manage to act). Even in the context of the 2014 rules, their discussion about narrative ways to explain weirdness only makes sense in the context of 'i draw my dagger in the middle of the masquerade ball' and similar situations, where people can react to things they can see (the dagger visibly being drawn) before the attack is actually made. If the invisible assassin gets distracted by seeing his aunt sally at the ball or a massive pile of almond pastries, no one does anything in the meantime, because nothing is happening as far as everyone but the assassin is concerned.
Here's the problem: Some characters are surprised but act before anything happens that would have surprised them - that's a time machine in action. The cause is in the future.
And the initiating character does not even expect to act first most of the time. Nor are they committed to do the action that would have initiated combat. So combat just starts for no reason, and could potentially not involve any combat at all (the attacker chooses to withdraw unobserved because the enemies act first). That's non-linear causality, quite possibly resulting in no causality at all (the initiating character gets killed before they get a chance to do the thing that was theoretically initiating combat, and is thus killed for no reason at all).
Both of these are huge problems, and no one could reasonably think things should work like that.
And they also lead to narrative problems. The evil chancellor orders his guards to kill the PCs, and the guards catch the PCs by surprise somewhere in the castle. The PCs win initiative and kill the guards before they act. Now instead of acting in self defense, the PCs are murderers under your interpretation - the guards never actually did anything! The PCs just suddenly slaughtered them, and that's how any eyewitness would describe the events that actually happened. The chancellor comes off as the good guy when he advises the king to imprison the PCs. It's an absolute disaster from a narrative perspective, because the PCs only attacked the guards because they were told they were being attacked!
There's no time machine. When action initiate combat its not resolved before Initiative but after in turn order, as explained in the DMG which can still be anticipated. Wether you're Invisible or surprised by combat starting doesn't make any mention changing that fact, all it does it say is affecting Initiative with Advvantage & Disadvantage respectively.
So when the narratives transition to violence everything pause to follow combat steps by steps how combat unfold; 1. Establish Positions 2. Roll Initiative 3. Take Turns
At this point you know you're surprised by combat starting with Disadvantaged to Initiative, while the Invisible creature have Advantage to Initiative.
You still can't see anyone Invisible who's action initiate combat which you anticipated somehow. How much is revealed at this point is up to DM.
There's many ways this can be described the rules don't elaborate on this. Did you see, heard, smelled or otherwise sensed something moments before the attack? Or suspect something due to lack thereof from unusual silence. Perhaps birds took off right before the attack following a noise. Ambusher jump out of the woods about to attack etc The more you fill the blank, the more narrative support to rules mechanic there will be.
That is essentially what the Devs were saying in the Dragontalk: Sage Advice when suggesting to come up with narrative ways to describe it when the D20 throw you a curve ball in unexpected ways.
Initiative is an mechanic used to make narration of actions happening in quick succession easier to manage. Seeing, hearing or otherwise sensing something is covered by Perception, not Initiative. If someone did not have a high enough active or passive Perception to notice an ambusher, having a high Initiative does not mechanically provide a benefit to Perception. A gut feeling due to an unusual silence, or animal reactions might be Insight. The same issue arises. There's no mechanical benefit to Insight from a high Initiative.
An attacker jumping out of the woods and being spotted doesn't apply to the OP's scenario, which has an ambusher already in hiding using an ability which makes them magically transparent and their footsteps silent.
Both of these are huge problems, and no one could reasonably think things should work like that.
And they also lead to narrative problems. The evil chancellor orders his guards to kill the PCs, and the guards catch the PCs by surprise somewhere in the castle. The PCs win initiative and kill the guards before they act. Now instead of acting in self defense, the PCs are murderers under your interpretation - the guards never actually did anything! The PCs just suddenly slaughtered them, and that's how any eyewitness would describe the events that actually happened. The chancellor comes off as the good guy when he advises the king to imprison the PCs. It's an absolute disaster from a narrative perspective, because the PCs only attacked the guards because they were told they were being attacked!
Guards attacking PC's in the castle is not a good example to support your case, because there could be any number of indications, expressions, body stance, pace of movement, and obviously the sound of swords being drawn.
The surprise system only really has problems for an unseen attacker who kicks combat off, but their initiative puts unaware characters ahead of them and lets them react to actions which haven't happened yet.
While action initiating combat aren't resolved before Initiative, something is iniatiated to make others surprised by combat starting. Something is told to DM to start combat.
DM preferring to limit Initiative randomness can always use Initiatice Score;
So what I actually do is a system that amounts to "if you successfully surprise the enemy, you win initiative". To cut down on die rolls I just combine the check for whether you successfully surprise the enemy (stealth vs perception most of the time, sometimes deception vs insight) with the initiative roll. This does mean the person who kicked off the combat might not go first, but at that point there's a narrative explanation: you botched your attempt to surprise them and they saw you coming.
OK, it's becoming quite clear that this is a you problem. You have a particularly idiosyncratic idea of how combat in D&D should behave, and it's incompatible with the 5e rules.
The 5e rules work. Whether they work well is a matter of opinion, but one can run combats with them by the book. Yes, one can contrive scenarios where they produce odd results, but so what? One can contrive scenarios for any set of rules that make them go badly.
To touch briefly on the scenario that started this thread, I'd probably drop into combat a couple of rounds before the party makes combat with the ambusher.
Jumping into combat a couple of rounds before making contact is not a great idea in my view. That's the GM practically insisting that the players actively metagame and roll perception checks until they succeed. If the GM constructed the situation with an ambusher designed to be undetected, I don't know why they would deliberately undermine it like that.
The thing is, they don't know why we're in combat time. Are there traps? Hazards? Enemies ready to rush out of the fortress gate? Snipers above? Etc.
(I also might introduce some traps or hazards, just to keep them on their toes.)
By dropping in early, IMO you reduce the players' ability to correctly predict what and when will happen.
But really, the only metagamey thing I'd expect players to do in the supposed problem situation here is prepare actions like attacking or cantrips that, narratively, are perfectly plausible things that one would do when an enemy pops out of nowhere by surprise. Nobody's gonna drop a fireball just because.
(As a side note, the great menace of metagaming is massively overrated. The problem, such as it is, is bad play.)
Would you as a GM have players roll initiative when traveling on an open road with no enemies around, just to fake players out?
I cannot disagree more. The answer cannot depend on what spell is being cast. No time machines. The GM should be able to decide what happens without being told what spell it is. (That is, yeah, he should be told, but his decision shouldn't change based on what he's told. The world needs to be consistent and causality linear barring prophecy, divination, or other explicit means of accessing the future).
Indeed, my preference as a DM would be for players to come up with what the components of each of their spells are, and narrate the components when they cast instead of declaring what spell they're casting.
OK, it's becoming quite clear that this is a you problem. You have a particularly idiosyncratic idea of how combat in D&D should behave, and it's incompatible with the 5e rules.
The 5e rules work. Whether they work well is a matter of opinion, but one can run combats with them by the book. Yes, one can contrive scenarios where they produce odd results, but so what? One can contrive scenarios for any set of rules that make them go badly.
To touch briefly on the scenario that started this thread, I'd probably drop into combat a couple of rounds before the party makes combat with the ambusher.
No time machines and linear causality is particularly idiosyncratic? Really?
Dude. You are talking about how the DM shouldn't know what spells the players are casting. Elsewhere you asserted that combat doesn't begin until everybody knows they're in a fight.
You appear to have this thing where you want the mechanics to be completely subservient to the fiction. Your "time machine" thing is because you're unwilling to accept that invoking the combat rules prior to an attack being made is perfectly fine and normal. (It's also how just about all non-surprise combats happen.)
I have yet to see the 2024 surprise rules played as you interpret them in any real game.
So? How many real games do you happen to observe?
It's also the kind of thing where a lot of tables have their own house practices. Plenty of people are still running surprise rounds in 24 games, likely because they are unaware there's a change. I might run it by the book, but my current game has virtually no fights where surprise is a factor, so it hasn't come up. (Combination of setting details, very negotiaty players, and not a lot of fights.)
OK, it's becoming quite clear that this is a you problem. You have a particularly idiosyncratic idea of how combat in D&D should behave, and it's incompatible with the 5e rules.
The 5e rules work. Whether they work well is a matter of opinion, but one can run combats with them by the book. Yes, one can contrive scenarios where they produce odd results, but so what? One can contrive scenarios for any set of rules that make them go badly.
To touch briefly on the scenario that started this thread, I'd probably drop into combat a couple of rounds before the party makes combat with the ambusher.
Jumping into combat a couple of rounds before making contact is not a great idea in my view. That's the GM practically insisting that the players actively metagame and roll perception checks until they succeed. If the GM constructed the situation with an ambusher designed to be undetected, I don't know why they would deliberately undermine it like that.
The thing is, they don't know why we're in combat time. Are there traps? Hazards? Enemies ready to rush out of the fortress gate? Snipers above? Etc.
(I also might introduce some traps or hazards, just to keep them on their toes.)
By dropping in early, IMO you reduce the players' ability to correctly predict what and when will happen.
But really, the only metagamey thing I'd expect players to do in the supposed problem situation here is prepare actions like attacking or cantrips that, narratively, are perfectly plausible things that one would do when an enemy pops out of nowhere by surprise. Nobody's gonna drop a fireball just because.
(As a side note, the great menace of metagaming is massively overrated. The problem, such as it is, is bad play.)
In my opinion, it's up to the PC's to decide whether to make Perception checks when doing things such as approaching an entrance to an enemy fortress. Is it smart for them to do so, absolutely. If they don't choose to make any, then it's up to their passive perception. If they are relying on passive perception, or they do make an active perception check but fail, them's the breaks. I'd typically only allow one or two rolls for Perception - either two players separately or one rolling with advantage.
Dropping them into initiative with no obvious threat means that they will almost certainly roll Perception checks, whether or not they already did that and failed. If they're dropped into initiative 2 or more rounds before making contact with the enemy, as you suggested, then they will get 2+ rounds to each make Perception checks, instead of just one each (and maybe they even made one before initiative started). The DM is the one who set up the situation with the Orc Garotter hiding near the entrance to begin with, and giving the players that many chances undermines the very scenario they set up.
By dropping in early, IMO you reduce the players' ability to correctly predict what and when will happen.
IMO, that's completely wrong. You yourself just said that you wouldn't have PC's roll initiative on an open road where nothing is going to happen, because that's a waste of everyone's time. That's correct. Which is why rolling initiative absolutely tells the players something IS going to happen, and there's an element of danger or risk to it. That's the exact opposite of reducing the players ability to correctly predict what and when something will happen.
But really, the only metagamey thing I'd expect players to do in the supposed problem situation here is prepare actions like attacking or cantrips that, narratively, are perfectly plausible things that one would do when an enemy pops out of nowhere by surprise. Nobody's gonna drop a fireball just because.
You'd think that, but earlier in the thread, dropping a fireball in response to an attacker which was silent, invisible, undetected and hadn't attacked yet (because they rolled lower on initiative) was one of the possibilities suggested:
The top of the order (before ambushers) is able to react to whatever is about to happen because they are just that much faster with their reflexes. If that means they take the dodge action, or ready a sword swing, or blow up the room, then so be it. Again, the ambusher is—supposedly—attacking in those same 6 seconds, so play into that. As the attack suddenly becomes visible, the PC who chose dodge was able to duck out of the way (or not, if the ambusher targeted another character). As the attack suddenly becomes visible, he gets stuck in the gut by the godly reflexes of the rogue (or not, because the attacker wasn't within range). As the attack becomes suddenly visible, the room explodes as the double 20 rolled by the wizard for initiative resulted in a fireball (or they ended up killing the tied up price they didn't spot and the attacker was actually 80 ft away with a xbow and shoots the wizard anyway).
Some of the quasi 5.24e rules proposed in this thread are even more awkward because it incentivizes the ambusher to roll low (i.e. they don't want advantage) so the party wastes their turn "doing nothing while waiting for combat to 'actually' start", effectively giving the "surprise round" that was supposed to be done away with.
***
I don't see any conflict with the rules as written by by letting the Garotter who the PC's failed to perceive get off the first attack and then rolling initiative with surprise afterwards (disadvantage for the surprised people, advantage for any attackers who are unseen - but no advantage on the one who just attacked). This is not the same thing as the 2014 surprise round where all the ambushers get a turn and all the ones ambushed are flat-footed and do nothing, it's one single attack which the attacker gets off first because they were undetected, and then initiative gets rolled.
For what it's worth, the probability an undetected invisible assailant (advantage on initiative) wins initiative against a single surprised opponent (disadvantage on initiative) with the same bonus to initiative is only 81.58%. So almost 1 in 5 times, a single ambushed enemy goes first.
This gets worse the more surprised enemies there are. P(ambusher with advantage goes first | all combatants have the same initiative modifier):
2 surprised opponents: 66.56%
3 surprised opponents: 54.30%
4 surprised opponents: 44.30%
5 surprised opponents: 36.14%
6 surprised opponents: 29.5%
So even with just 3 opponents, you've only got about a 1 in 2 chance of going first as the ambusher. Against typical party sizes, you expect at least one of them to act before you.
If any of the surprised opponents has a higher initiative bonus than the ambusher, these odds drop even further.
Given a party of adventurers is generally 4-6 players, it isn't a rare occurrence for an ambushing enemy to be beaten on initiative. It's actually the expected result.
(Method: generated the probability of any given result for both 2d20 drop highest and 2d20 drop lowest. Computed the probability of each die pairing by multiplying the vectors and forming a matrix. multiplied all elements of this matrix by the corresponding elements of a matrix of the same size that was 1 wherever 2d20 drop the lowest > 2d20 drop the highest (which is a triangular matrix with a diagonal of 0s, and 1s above it), then summed all squares of that final matrix).
There's no time machine. When action initiate combat its not resolved before Initiative but after in turn order, as explained in the DMG which can still be anticipated. Wether you're Invisible or surprised by combat starting doesn't make any mention changing that fact, all it does it say is affecting Initiative with Advvantage & Disadvantage respectively.
So when the narratives transition to violence everything pause to follow combat steps by steps how combat unfold; 1. Establish Positions 2. Roll Initiative 3. Take Turns
At this point you know you're surprised by combat starting with Disadvantaged to Initiative, while the Invisible creature have Advantage to Initiative.
You still can't see anyone Invisible who's action initiate combat which you anticipated somehow. How much is revealed at this point is up to DM.
There's many ways this can be described the rules don't elaborate on this. Did you see, heard, smelled or otherwise sensed something moments before the attack? Or suspect something due to lack thereof from unusual silence. Perhaps birds took off right before the attack following a noise. Ambusher jump out of the woods about to attack etc The more you fill the blank, the more narrative support to rules mechanic there will be.
That is essentially what the Devs were saying in the Dragontalk: Sage Advice when suggesting to come up with narrative ways to describe it when the D20 throw you a curve ball in unexpected ways.
Yeah, except the dragontalk assumes 2014 surprise rules, where the surprised characters don't get to act in round 1. Totally different, because at least if they're not acting, they don't have a time machine and it doesn't disturb causality, and not at all applicable to the 2024 rules. (Nor does it help that there's no narrative way to handle the unobserved invisible attacker or the subtle spell sorceror - if they're delayed for some reason, combat simply doesn't happen until they do manage to act). Even in the context of the 2014 rules, their discussion about narrative ways to explain weirdness only makes sense in the context of 'i draw my dagger in the middle of the masquerade ball' and similar situations, where people can react to things they can see (the dagger visibly being drawn) before the attack is actually made. If the invisible assassin gets distracted by seeing his aunt sally at the ball or a massive pile of almond pastries, no one does anything in the meantime, because nothing is happening as far as everyone but the assassin is concerned.
Here's the problem: Some characters are surprised but act before anything happens that would have surprised them - that's a time machine in action. The cause is in the future.
And the initiating character does not even expect to act first most of the time. Nor are they committed to do the action that would have initiated combat. So combat just starts for no reason, and could potentially not involve any combat at all (the attacker chooses to withdraw unobserved because the enemies act first). That's non-linear causality, quite possibly resulting in no causality at all (the initiating character gets killed before they get a chance to do the thing that was theoretically initiating combat, and is thus killed for no reason at all).
Both of these are huge problems, and no one could reasonably think things should work like that.
And they also lead to narrative problems. The evil chancellor orders his guards to kill the PCs, and the guards catch the PCs by surprise somewhere in the castle. The PCs win initiative and kill the guards before they act. Now instead of acting in self defense, the PCs are murderers under your interpretation - the guards never actually did anything! The PCs just suddenly slaughtered them, and that's how any eyewitness would describe the events that actually happened. The chancellor comes off as the good guy when he advises the king to imprison the PCs. It's an absolute disaster from a narrative perspective, because the PCs only attacked the guards because they were told they were being attacked!
Initiative is an mechanic used to make narration of actions happening in quick succession easier to manage. Seeing, hearing or otherwise sensing something is covered by Perception, not Initiative. If someone did not have a high enough active or passive Perception to notice an ambusher, having a high Initiative does not mechanically provide a benefit to Perception. A gut feeling due to an unusual silence, or animal reactions might be Insight. The same issue arises. There's no mechanical benefit to Insight from a high Initiative.
An attacker jumping out of the woods and being spotted doesn't apply to the OP's scenario, which has an ambusher already in hiding using an ability which makes them magically transparent and their footsteps silent.
Guards attacking PC's in the castle is not a good example to support your case, because there could be any number of indications, expressions, body stance, pace of movement, and obviously the sound of swords being drawn.
The surprise system only really has problems for an unseen attacker who kicks combat off, but their initiative puts unaware characters ahead of them and lets them react to actions which haven't happened yet.
While action initiating combat aren't resolved before Initiative, something is iniatiated to make others surprised by combat starting. Something is told to DM to start combat.
DM preferring to limit Initiative randomness can always use Initiatice Score;
Ambusher 15 + DEX mod
Surprised 05 + DEX mod
So what I actually do is a system that amounts to "if you successfully surprise the enemy, you win initiative". To cut down on die rolls I just combine the check for whether you successfully surprise the enemy (stealth vs perception most of the time, sometimes deception vs insight) with the initiative roll. This does mean the person who kicked off the combat might not go first, but at that point there's a narrative explanation: you botched your attempt to surprise them and they saw you coming.
The thing is, they don't know why we're in combat time. Are there traps? Hazards? Enemies ready to rush out of the fortress gate? Snipers above? Etc.
(I also might introduce some traps or hazards, just to keep them on their toes.)
By dropping in early, IMO you reduce the players' ability to correctly predict what and when will happen.
But really, the only metagamey thing I'd expect players to do in the supposed problem situation here is prepare actions like attacking or cantrips that, narratively, are perfectly plausible things that one would do when an enemy pops out of nowhere by surprise. Nobody's gonna drop a fireball just because.
(As a side note, the great menace of metagaming is massively overrated. The problem, such as it is, is bad play.)
No; that's wasting everybody's time.
Dude. You are talking about how the DM shouldn't know what spells the players are casting. Elsewhere you asserted that combat doesn't begin until everybody knows they're in a fight.
You appear to have this thing where you want the mechanics to be completely subservient to the fiction. Your "time machine" thing is because you're unwilling to accept that invoking the combat rules prior to an attack being made is perfectly fine and normal. (It's also how just about all non-surprise combats happen.)
So? How many real games do you happen to observe?
It's also the kind of thing where a lot of tables have their own house practices. Plenty of people are still running surprise rounds in 24 games, likely because they are unaware there's a change. I might run it by the book, but my current game has virtually no fights where surprise is a factor, so it hasn't come up. (Combination of setting details, very negotiaty players, and not a lot of fights.)
In my opinion, it's up to the PC's to decide whether to make Perception checks when doing things such as approaching an entrance to an enemy fortress. Is it smart for them to do so, absolutely. If they don't choose to make any, then it's up to their passive perception. If they are relying on passive perception, or they do make an active perception check but fail, them's the breaks. I'd typically only allow one or two rolls for Perception - either two players separately or one rolling with advantage.
Dropping them into initiative with no obvious threat means that they will almost certainly roll Perception checks, whether or not they already did that and failed. If they're dropped into initiative 2 or more rounds before making contact with the enemy, as you suggested, then they will get 2+ rounds to each make Perception checks, instead of just one each (and maybe they even made one before initiative started). The DM is the one who set up the situation with the Orc Garotter hiding near the entrance to begin with, and giving the players that many chances undermines the very scenario they set up.
IMO, that's completely wrong. You yourself just said that you wouldn't have PC's roll initiative on an open road where nothing is going to happen, because that's a waste of everyone's time. That's correct. Which is why rolling initiative absolutely tells the players something IS going to happen, and there's an element of danger or risk to it. That's the exact opposite of reducing the players ability to correctly predict what and when something will happen.
You'd think that, but earlier in the thread, dropping a fireball in response to an attacker which was silent, invisible, undetected and hadn't attacked yet (because they rolled lower on initiative) was one of the possibilities suggested:
***
I don't see any conflict with the rules as written by by letting the Garotter who the PC's failed to perceive get off the first attack and then rolling initiative with surprise afterwards (disadvantage for the surprised people, advantage for any attackers who are unseen - but no advantage on the one who just attacked). This is not the same thing as the 2014 surprise round where all the ambushers get a turn and all the ones ambushed are flat-footed and do nothing, it's one single attack which the attacker gets off first because they were undetected, and then initiative gets rolled.