Stealth and Risk

For combat stuff that doesn't fit into any of the other forums.
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Stealth and Risk

Post by Acarin »

The recent update causes enemies to stay paranoid for longer and increases risk by 12 per search. For most, the increased risk decreases stealth rerolls. I'll get to this in a bit. More importantly, there is one warrior specialization that is more impacted by changes in risk than any others (nightblades) since nightblade tactic bonuses (or penalties) are entirely based on risk level with everyone 20 accumulated risk reducing rerolls for offense and defense by 1 (down to -2). This means that if I induce paranoia in an opponent, it takes 7-8 searches for the enemy before I am somehow weaker in combat than someone in a non-combat guild. This tactic is already more restricted and specific than other warrior tactics (in encumbrance, weapons that can be used, etc) and is the only one with potential to make you weaker than any other guild.

I find that those change makes nightblades almost unplayable. It is fine if you kill an enemy after one shot and are never noticed but being noticed and rehiding or attacking in ranged essentially means that 1) I will quickly be spotted and 2) drawn out of hiding with negative defensive and offensive rerolls.

As an example, with 400+ stealth, I was not able to stay in hiding through more than 3-5 searches against a vagrant (about as easy as it gets) and was drawn out nearing negative rerolls. Does that sound right?

Here's my suggestion... drop risk accumulation to 3-4 per search. Align nightblade tactics reroll bonuses with what duelists end up getting from combat analysis (after any adjustments) but in reverse (minimum of 0 rerolls). Give a chance for any ranged attack to be traced back to the attacker (stealth rerolls increase with risk). INCREASE RISK ACCUMULATION FROM RANGED ATTACKS (not melee which exposes your location) to 15-20. Remove the extra defense pushdown reroll that ranged gives over melee.

What this does: Allows ranged users to still attack and remain hidden but means that if they try to rapidly attack (rather than choosing specific well placed shots at a reasonable rate), they will be seen quickly.

Again, the way things are set up now penalizes nightblades way more than others and I do not even feel I can participate in solo combat right now using stealth. Keep in mind that risk takes time to go down so this also means extra huge downtime during combat in addition to the possibility of potentially waiting long periods for mark to activate. Given that 3 stealth checks are required to hide, stalk, stalk and some enemies autostart, this change feels like it destroys the nightblade class.

Regarding risk reducing stealth rerolls, we already have to pass multiple checks to do anything in combat. I dont have an idea of what stealth rerolls actually look like but I do believe that reducing rerolls below 0 at max risk would be relatively unfair given that this is a skill investment and is weighed against perception. If anyone can just decide to search a room 6 times to max risk and be near guaranteed to find whoever is in it, what is the point of training the skill for those that invest?
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Re: Stealth and Risk

Post by Acarin »

Just to add to this, I would suggest that multi-throw count as a single attack for risk accumulation purposes (no matter how many knives are actually thrown).
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
User avatar
Rias
DEV
Posts: 2024
Joined: Sun Sep 03, 2017 4:06 pm
Location: Wandering Temicotli

Re: Stealth and Risk

Post by Rias »

it takes 7-8 searches for the enemy before I am somehow weaker in combat than someone in a non-combat guild.
I'm looking at lowering the rate of risk accumulation, but I'll also suggest stealth users not stay stealthed by an enemy when they are searching and increasing the stealther's risk. If a stealth user has somehow tipped off an enemy and they start searching, I recommend either moving away and approaching later, or starting the attack immediately before the enemy actually finds or further increases the risk of the stealther. Part of the stealth combat experience, and that of Nightblades in particular, is deciding when to attack and when to withdraw and come at it again later. Sometimes they may have their hand forced early and need to attack without being able to finish that perfect setup.
being noticed and rehiding or attacking in ranged essentially means that 1) I will quickly be spotted and 2) drawn out of hiding with negative defensive and offensive rerolls.
See above about making strategic withdrawals or starting an attack earlier, rather than trying to remain in hiding while the enemy is actively searching. The idea is that once the enemy is aware of a hider's presence and searching, the hider is forced to act in some way other than remaining in hiding in front of the person searching for them and just waiting it out.
As an example, with 400+ stealth, I was not able to stay in hiding through more than 3-5 searches against a vagrant (about as easy as it gets) and was drawn out nearing negative rerolls. Does that sound right?
I don't feel like anyone should be able to stay hidden indefinitely in front of any kind of creature that's determinedly searching for them. Again, people nearby have been made paranoid and started searching, stealth users should be forced to act rather than waiting right in front of the enemy until the enemy gets bored and gives up. Cue Skyrim meme of the bandit guard with an arrow in his head giving up his search and saying "must have been the wind."
drop risk accumulation to 3-4 per search.
At most I'd drop it back to the original 8, which Nightblades and others seemed to be managing quite well with before these changes. I will be reducing it, though.
Align nightblade tactics reroll bonuses with what duelists end up getting from combat analysis (after any adjustments) but in reverse (minimum of 0 rerolls).
1) Duelists' Analysis bonuses are going to be toned down soon - the Duelists themselves recommended as much, stating it felt too powerful. 2) The idea is that Nightblades get to start off strong with positive rerolls but eventually lose them to the point of a disadvantage over time, while Duelists need to spend time making actions (unhidden and being attacked the whole time, mind) to build up to their advantages over time. So they're not exactly equal in value and able to be mirrored but in opposite directions. I may decrease the Nightblade negative reroll cap, but I still want it to end up at a negative after a while so the Nightblade is encouraged to finish their job quick, or withdraw and make a new approach if things go south. Start strong with an advantage over other Warriors - with no effort required for those starting positive rerolls - but get the job done quick.
Give a chance for any ranged attack to be traced back to the attacker (stealth rerolls increase with risk).
Possibly. I feel like with the increasing risk per search system they're going to be detected soon anyway (see the experience with the laborer), while still getting a chance for a couple shots shot before their hiding spot is found. We don't have a lot of ranged snipers running around yet so we still need to get a feel for how much this needs tweaking. I'm keeping an eye on it.
INCREASE RISK ACCUMULATION FROM RANGED ATTACKS
This may well happen.
Remove the extra defense pushdown reroll that ranged gives over melee.
I don't feel this is necessary.
the way things are set up now penalizes nightblades way more than others and I do not even feel I can participate in solo combat right now using stealth.
The intent was not to hurt Nightblades specifically more than other stealth users, so I do want to mitigate that feeling.
Given that 3 stealth checks are required to hide, stalk, stalk and some enemies autostart, this change feels like it destroys the nightblade class. [...] Regarding risk reducing stealth rerolls, we already have to pass multiple checks to do anything in combat.
The update to risk-related reroll penalties only applies to staying unnoticed when something is actively searching for the hider. Everything else is the same. So the update only affects stealth users once their target has been made paranoid. And when that happens, stealth users should take some kind of action, be that starting their attack, sneaking away, or otherwise.
If anyone can just decide to search a room 6 times to max risk and be near guaranteed to find whoever is in it, what is the point of training the skill for those that invest?
To cue up my broken record response: If a stealth user has tipped someone off to their presence, the stealth user should have to take some kind of action, not just remain passively in hiding assuming the other person will never be able to find them. What might help is to have some increasing (up to a cap) roundtime per additional search within a short time, so people put themselves increasingly at risk by excessive just-in-case searching. I've talked a lot about how stealth users should take action if their presence is suspected, but the same goes for people who suspect they're in danger of someone being hidden nearby. Rather than searching over and over to depend on the Risk buildup to out the hider, it would probably be a better idea to pack up and get away if one suspects an ambush is about to take place and one can't quickly find the location of the ambusher within a search or two.

Regarding a comment I saw somewhere about wechuge: The wechuge do randomly search, but they don't do it rapidfire/constantly. I've just hid in front of one and watched it idle around for 20 seconds before it walked away, no search. Another randomly searched once after I entered the room successfully in stealth, then also idled for a good 30 seconds before wandering off.
<Rias> PUT ON PANTS
<Fellborn> NO
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Re: Stealth and Risk

Post by Acarin »

To offset this for nightblades if you feel this is already balanced for other stealth classes (although not sure it is), a risk mitigation planning ability could be added as well:

Nighblades always keep a backup plan in mind. Reduces risk accumulation from enemy actions NOT one's own actions to 1/3 or 1/4 of normal value and mitigates the penalties created by risk on tactics: nightblade - rerolls now range from 3/3 to 0/0 instead of 2 to -2. Could be a stealth requirement of 200-400 on this one as well.
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Re: Stealth and Risk

Post by Acarin »

Really bad at this quoting thing so apologies if this doesn't break up right
Rias wrote: Sun Feb 14, 2021 10:43 am
it takes 7-8 searches for the enemy before I am somehow weaker in combat than someone in a non-combat guild.
I'm looking at lowering the rate of risk accumulation, but I'll also suggest stealth users not stay stealthed by an enemy when they are searching and increasing the stealther's risk. If a stealth user has somehow tipped off an enemy and they start searching, I recommend either moving away and approaching later, or starting the attack immediately before the enemy actually finds or further increases the risk of the stealther. Part of the stealth combat experience, and that of Nightblades in particular, is deciding when to attack and when to withdraw and come at it again later. Sometimes they may have their hand forced early and need to attack without being able to finish that perfect setup.
being noticed and rehiding or attacking in ranged essentially means that 1) I will quickly be spotted and 2) drawn out of hiding with negative defensive and offensive rerolls.
See above about making strategic withdrawals or starting an attack earlier, rather than trying to remain in hiding while the enemy is actively searching. The idea is that once the enemy is aware of a hider's presence and searching, the hider is forced to act in some way other than remaining in hiding in front of the person searching for them and just waiting it out.
The description of nightblade tactics is that we should not be staying engaged with a target too long, not staying in the room disengaged and hiding. Note that a single mark can take minutes to activate and you have coded a clear delay there. That means that if an enemy decided to start searching, regardless of whether we have tipped them off or it's auto, this causes problems. I do not have control over when this ability activates and should not be penalized this severely for remaining in a room if something decides to search while waiting for a minute for Mark to activate. Especially since they may not even be looking for me. After I attack, they also maintain there paranoia as well and you have just extended the period that this lasts so waiting for this to end is a huge annoyance.
As an example, with 400+ stealth, I was not able to stay in hiding through more than 3-5 searches against a vagrant (about as easy as it gets) and was drawn out nearing negative rerolls. Does that sound right?
I don't feel like anyone should be able to stay hidden indefinitely in front of any kind of creature that's determinedly searching for them. Again, people nearby have been made paranoid and started searching, stealth users should be forced to act rather than waiting right in front of the enemy until the enemy gets bored and gives up. Cue Skyrim meme of the bandit guard with an arrow in his head giving up his search and saying "must have been the wind."
So basically you're saying that stealth should just be useful for an initial check and then it doesn't matter how much stealth you have... you should be found if someone searches? This sounds absurd to me and again... it requires 3 passed checks to avoid tipping someone off... this is way more than just driving a sword into someone's head. This is against a vagrant... the second weakest mob in the game... and this sounds reasonable to you?
drop risk accumulation to 3-4 per search.
At most I'd drop it back to the original 8, which Nightblades and others seemed to be managing quite well with before these changes. I will be reducing it, though.
if it was set to 8 before, maybe this was a bug or I was just clearly not paying attention because it seemed like we were only getting 3 or 4 per search.
Align nightblade tactics reroll bonuses with what duelists end up getting from combat analysis (after any adjustments) but in reverse (minimum of 0 rerolls).
1) Duelists' Analysis bonuses are going to be toned down soon - the Duelists themselves recommended as much, stating it felt too powerful. 2) The idea is that Nightblades get to start off strong with positive rerolls but eventually lose them to the point of a disadvantage over time, while Duelists need to spend time making actions (unhidden and being attacked the whole time, mind) to build up to their advantages over time. So they're not exactly equal in value and able to be mirrored but in opposite directions. I may decrease the Nightblade negative reroll cap, but I still want it to end up at a negative after a while so the Nightblade is encouraged to finish their job quick, or withdraw and make a new approach if things go south. Start strong with an advantage over other Warriors - with no effort required for those starting positive rerolls - but get the job done quick.
I meant align to what they end up with. I'm sorry but learning someone's combat style to some degree should not have significantly higher reroll caps than attacking someone who is completely unaware they are about to be attacked (after studying them for a while... just as the duelist is doing). Maybe this means an adjustment to the actual ambush rerolls instead of overall though.

Not understanding how a warrior that is trained in combat is ever going to be worse than a farmer though and therefore penalized. Are we supposed to remain in open combat or not after ambushing? I asked this the other day but the intent is becoming increasingly unclear. There is significant motivation to have rerolls and I do think a penalty is not necessary. I usually try to disengage after risk gets over 20-40 anyway...
Give a chance for any ranged attack to be traced back to the attacker (stealth rerolls increase with risk).
Possibly. I feel like with the increasing risk per search system they're going to be detected soon anyway (see the experience with the laborer), while still getting a chance for a couple shots shot before their hiding spot is found. We don't have a lot of ranged snipers running around yet so we still need to get a feel for how much this needs tweaking. I'm keeping an eye on it.
Cool... seems like we're applying ranged specific issues (who can stay in hiding indefinitely) to stealth overall. If someone is throwing things at me or firing, it's likely that I'm going to see the trajectory (or know what direction it came from if it hits) and have a chance to locate the combatant... keep in mind that even if spotted, it still takes time to close on them and push them out of hiding.
Remove the extra defense pushdown reroll that ranged gives over melee.
I don't feel this is necessary.
With aimed shot, that's two extra rerolls... but cool. I don't understand the reasoning at all but its your game.
the way things are set up now penalizes nightblades way more than others and I do not even feel I can participate in solo combat right now using stealth.
The intent was not to hurt Nightblades specifically more than other stealth users, so I do want to mitigate that feeling.
Thanks. I look forward to seeing what you end up with bit that is clearly the end result.
Given that 3 stealth checks are required to hide, stalk, stalk and some enemies autostart, this change feels like it destroys the nightblade class. [...] Regarding risk reducing stealth rerolls, we already have to pass multiple checks to do anything in combat.
The update to risk-related reroll penalties only applies to staying unnoticed when something is actively searching for the hider. Everything else is the same. So the update only affects stealth users once their target has been made paranoid. And when that happens, stealth users should take some kind of action, be that starting their attack, sneaking away, or otherwise.
When hunting like skill mobs, checks will be failed regularly. This means that mark is now impossible to set up on like-skill enemies even with rehiding. This may be fine in theory but does not work out in practice... again, hitting nightblades harder than others.
If anyone can just decide to search a room 6 times to max risk and be near guaranteed to find whoever is in it, what is the point of training the skill for those that invest?
To cue up my broken record response: If a stealth user has tipped someone off to their presence, the stealth user should have to take some kind of action, not just remain passively in hiding assuming the other person will never be able to find them. What might help is to have some increasing (up to a cap) roundtime per additional search within a short time, so people put themselves increasingly at risk by excessive just-in-case searching. I've talked a lot about how stealth users should take action if their presence is suspected, but the same goes for people who suspect they're in danger of someone being hidden nearby. Rather than searching over and over to depend on the Risk buildup to out the hider, it would probably be a better idea to pack up and get away if one suspects an ambush is about to take place and one can't quickly find the location of the ambusher within a search or two.
Adding increasing rts for searching may mitigate this a bit. I ask again though... if I invested 600 in stealth, do you think that it's reasonable that someone with 0 perception is basically guaranteed to find me after x searches? I like talking to people from hiding too with no intent of killing them entirely outside of combat. Guess I won't be doing that... in a CvC situation as well, now it's not worth the risk of giving someone a warning to give them a chance so there are other consequences of this.
Regarding a comment I saw somewhere about wechuge: The wechuge do randomly search, but they don't do it rapidfire/constantly. I've just hid in front of one and watched it idle around for 20 seconds before it walked away, no search. Another randomly searched once after I entered the room successfully in stealth, then also idled for a good 30 seconds before wandering off.
And there are others that are searching right when you walk in the room and multiple times. They have high perception and good chance of seeing you even at 400+ stealth. Gives them a free chance at paranoia if they notice you and additional rerolls pushdowns for nightblades. It's fine. I just won't hunt them knowing that my risk is going to skyrocket randomly.
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Re: Stealth and Risk

Post by Acarin »

I don't want to argue with your vision at all but the overall issue that I see is that the rate of accumulation of risk based penalties feels unreasonably high and disproportionately impacts nightblades over all others by quickly compromising overall combat capabilities.

I realize that you have the best intentions in mind for the game and the community. I've given my opinion here (and mean no disrespect, of course) and will respectfully withdraw. I do hope that additional changes are made to balance this out more for all stealth characters. I am going to take a break from playing until I am less frustrated though.
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
User avatar
Rias
DEV
Posts: 2024
Joined: Sun Sep 03, 2017 4:06 pm
Location: Wandering Temicotli

Re: Stealth and Risk

Post by Rias »

I appreciate your willingness to share your feedback, as I'm sure all players of Nightblades do. Hearing from players is the best way for me to reach the right balance with all this. Sorry to hear about the frustration, but I understand taking a break. I very much appreciate you taking the time to share your thoughts and experience before doing so. Hopefully we're in a better place real soon and you can feel comfortable jumping back into the saddle.

I do want to clarify that the paranoia duration was only increased (from 30 to 60) on actual ambush attacks. Paranoia duration from any other action (like getting a feeling of being stalked or marked) has not changed.
The description of nightblade tactics is that we should not be staying engaged with a target too long, not staying in the room disengaged and hiding. Note that a single mark can take minutes to activate and you have coded a clear delay there.
Right, the time spent sneaking up on the target and setting up the ambush is considered pre-engagement preparation. Once you've engaged in the actual fight by making an attack (or being attacked), that's when the intended Nightblade "style" is to get the encounter to end as quickly as possible - whether that's by taking the target down, or disengaging if the fight drags out too long to make a second approach later.
That means that if an enemy decided to start searching, regardless of whether we have tipped them off or it's auto, this causes problems. I do not have control over when this ability activates and should not be penalized this severely for remaining in a room if something decides to search while waiting for a minute for Mark to activate. Especially since they may not even be looking for me.
If an enemy starts immediately rapid-searching after being stalked or marked, they made their Perception check and I would suggest not using Mark, but instead attacking them or leaving to engage them again later. Mark is a powerful ability but it's not meant to be the sole way an adaptable Nightblade engages in combat.

If a mob just randomly starts rapid-searching with nothing to start that off, that sounds like a mob AI error. Things like wechuge that have a chance to randomly search may occasionally do it multiple times in a row if the RNG rolls that way, but they've got other actions they are programmed to consider, including "do nothing", which should make this rare. If it does happen, it may be worth sneaking away to avoid getting caught, and sneaking back in after you've taken a few moments to let your Risk tick back down. As I've been playing with wechuge today I have yet to see any random rapid multi-searches happen (though one did search me quickly and nearly take my leg off when I attacked ...), but there may be some particular circumstance that's incorrectly causing them to just search over and over. If anyone sees behavior like this, please use the BUG command in the room it happened in, and include what type of mob it was.
After I attack, they also maintain there paranoia as well and you have just extended the period that this lasts so waiting for this to end is a huge annoyance.
One possible solution might be making paranoia go down at an accelerated rate if the attacker has left the room. Hmmm.
So basically you're saying that stealth should just be useful for an initial check and then it doesn't matter how much stealth you have... you should be found if someone searches?
If they search long enough, and the hider does nothing to respond to the situation, then yes: eventually. It would have happened even before this change due to the open nature of the rolls in this game. The change just made the chances more weighted.
it requires 3 passed checks to avoid tipping someone off...
This is the battle that stealth fighters engage in for the benefit of being safe and in control as they set up their ambush, being able to get out of the situation if they detect things have gone bad. Contrast to non-stealth fighters who are making constant combat checks for offense and defense, and being attacked all the while, as they try to build up their own advantages (Balance, Analysis, Fury, etc). If an ambusher gets a bad roll and makes a target paranoid and they prefer not to go through with the attack on a paranoid target, they can always move off to another mob in the area and come back to the first one later when the paranoia duration has ended.
if it was set to 8 before, maybe this was a bug or I was just clearly not paying attention because it seemed like we were only getting 3 or 4 per search.
It was definitely 8, as shown here in my version control commit. Risk does tick down every second so it may have gone down a bit before you noted it, maybe?
I'm sorry but learning someone's combat style to some degree should not have significantly higher reroll caps than attacking someone who is completely unaware they are about to be attacked (after studying them for a while... just as the duelist is doing). Maybe this means an adjustment to the actual ambush rerolls instead of overall though.
The difference is that the duelist is engaging them in actual combat and analysing their combat movements and capabilities. The Nightblade setting up a Mark is not watching the creature's actual combat actions (unless it's attacking someone else I suppose, but that's not a consideration for the mechanics of the ability), but waiting for the perfect moment when the target is not at guard and has an exposed weak point. And don't forget that Mark is not only getting Nightblades an offense reroll, but pushing down the target's defense rerolls on top of the usual stealth pushdown. Nightblades are often getting the double benefit of not only taking the best of their own offense rolls, but forcing their target to take the worst of defense rolls as well.
Not understanding how a warrior that is trained in combat is ever going to be worse than a farmer though and therefore penalized.
The concept is supposed to come across as the Nightblade utilizing a risky combat style which gives them an advantage over everyone else at the start, but the risks eventually catch up with them and they start to suffer for it, simulated by the negative they end up in after staying engaged in combat for too long. It's not a 100% accurate-to-life move-by-move execution for sure, but is meant to be a more abstract simulation of the concept. And, of course, it serves as a balance to Nightbaldes starting out with a very powerful +2 rerolls to all attacks that require no time or active buildup to achieve.
Are we supposed to remain in open combat or not after ambushing?
Yes, just not for a prolonged amount of time. Strike hard and fast, do your best to finish the job, but if things draw out too long that you're now suffering negative rerolls, fall back and set up a fresh attack. There's a lot of room between "attack once, then retreat" and "never stop attacking once in the open".

Playing my Nightblade I'm able to:
- Make an initial ambush with +2 rerolls to offense and defense, plus whatever pushdown the ambush does to enemy rerolls. +3 rerolls and even more pushdown if I use Mark. Then it's open combat where I can:
- make 3 attacks while enjoying +2 rerolls to offense and defense,
- then make 3 attacks while enjoying +1 reroll to offense and defense,
- then make 3 attacks at no bonus to offense and defense.

So once I get to my 11th attack, that's when I start feeling the negatives. If I as a Nightblade am unable to take an opponent down after an ambush and 9 additional attacks, I'm going to slip away and make a fresh approach. Preferably an ambush. 10 attacks at 5 RT each is 50 seconds, which means in 10 more seconds the paranoia from the initial ambush is going to be gone. By the time I've re-hid and waited for my Risk to go back down, that paranoia effect has long been over.
keep in mind that even if spotted, it still takes time to close on them and push them out of hiding.
That is a good point.
When hunting like skill mobs, checks will be failed regularly. This means that mark is now impossible to set up on like-skill enemies even with rehiding.
This statement made me particularly worried, so I wanted to get a more accurate picture of what your specific experience was like. I took the liberty of cloning your Nightblade character on my local copy of the game server and took him out for a wechuge hunt. Out of 20 wechuge, 3 detected me: 1 when sneaking into the room, 1 on a random search, and 1 after stalking it. The rest were the same: Sneak into the room, stalk twice, mark, attack. 14 of the 17 I set up undetected died in a single blow prompted by Mark. 2 of those that detected me, I took down after a few aimed attacks to the chest. The remaining one I ran away from after losing my positive rerolls to Risk, and came back to stalk, mark, and one-shot later. In the end I had a slight wound to my arm as evidence of my trek into dangerous wechuge territory.

After that experience I'm feeling that combat is very manageable as a Nightblade after these updates, and I'm wondering if possibly we have staggeringly polar opposite RNG luck. Are you having a different experience?
I ask again though... if I invested 600 in stealth, do you think that it's reasonable that someone with 0 perception is basically guaranteed to find me after x searches?
Yes. It's also basically guaranteed that someone with 0 Melee skill is going to hit someone with 700 Dodge skill after x attacks. With the RNG everything is going to happen eventually. It's the nature of how the roll system is set up in this game. I get the concern, but just like a skilled warrior shouldn't assume he's untouchable just standing passively in front of someone who's attacking him, a skilled stealth ... ior ... shouldn't assume he's unfindable just sticking around while someone is continuously searching for him.
<Rias> PUT ON PANTS
<Fellborn> NO
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Re: Stealth and Risk

Post by Acarin »

Rias wrote: Sun Feb 14, 2021 2:30 pm I appreciate your willingness to share your feedback, as I'm sure all players of Nightblades do. Hearing from players is the best way for me to reach the right balance with all this. Sorry to hear about the frustration, but I understand taking a break. I very much appreciate you taking the time to share your thoughts and experience before doing so. Hopefully we're in a better place real soon and you can feel comfortable jumping back into the saddle.
Thanks. Understood. And happy to continue to provide feedback around this. Nightblades are very cool and you did an amazing job coding them and making them fun. That being said, I probably won't be on for a while.
do want to clarify that the paranoia duration was only increased (from 30 to 60) on actual ambush attacks. Paranoia duration from any other action (like getting a feeling of being stalked or marked) has not changed.
I wasn't aware that Mark had a check associated with it... cool... so 4 checks. The problem comes from any failed check resulting in a rapid increase in risk. I am less concerned with the duration. Although If I am caught and then attack and engage for a bit... the risk accumulation continues after rehiding. This seems reasonable though.
Right, the time spent sneaking up on the target and setting up the ambush is considered pre-engagement preparation. Once you've engaged in the actual fight by making an attack (or being attacked), that's when the intended Nightblade "style" is to get the encounter to end as quickly as possible - whether that's by taking the target down, or disengaging if the fight drags out too long to make a second approach later.
The change is penalizing in the pre-engagement part and not following engagement (i.e. after successful ambush). In other words, it's a pain before I make an attack but does nothing if I succeed in making an ambush in the first place because that does not trigger paranoia for melee attacks. Getting caught sneaking triggers the paranoia. This actually motivates characters to stay engaged instead so they don't start taking risk penalties if they try to rehide.
If a mob just randomly starts rapid-searching with nothing to start that off, that sounds like a mob AI error. Things like wechuge that have a chance to randomly search may occasionally do it multiple times in a row if the RNG rolls that way, but they've got other actions they are programmed to consider, including "do nothing", which should make this rare. If it does happen, it may be worth sneaking away to avoid getting caught, and sneaking back in after you've taken a few moments to let your Risk tick back down. As I've been playing with wechuge today I have yet to see any random rapid multi-searches happen (though one did search me quickly and nearly take my leg off when I attacked ...), but there may be some particular circumstance that's incorrectly causing them to just search over and over. If anyone sees behavior like this, please use the BUG command in the room it happened in, and include what type of mob it was.
Other characters in a hunting area frequently trigger paranoia. This can result in immediate risk increases if you walk in on someone. I have not seen this (besides wechuge) independent of this situation but have seen multi-searches with wechuge.
One possible solution might be making paranoia go down at an accelerated rate if the attacker has left the room. Hmmm.
This is an interesting idea. Alternatively, risk could be specific to a particular mob rather than a global effect on a character. If I steal from one person, I dont see why I would have greater risk when I move a room over and steal from the next guy.
If they search long enough, and the hider does nothing to respond to the situation, then yes: eventually. It would have happened even before this change due to the open nature of the rolls in this game. The change just made the chances more weighted.
Agree. The key here is how quickly this occurs.
This is the battle that stealth fighters engage in for the benefit of being safe and in control as they set up their ambush, being able to get out of the situation if they detect things have gone bad. Contrast to non-stealth fighters who are making constant combat checks for offense and defense, and being attacked all the while, as they try to build up their own advantages (Balance, Analysis, Fury, etc). If an ambusher gets a bad roll and makes a target paranoid and they prefer not to go through with the attack on a paranoid target, they can always move off to another mob in the area and come back to the first one later when the paranoia duration has ended.
And these players pay for this extra defense by a commitment in skill points above the normal to have this added benefit. In addition, melee fighters are subjected to defense rolls after an attack that are ranged are not (meaning 1 attack occurs before rehiding is possible). Stealth is a huge benefit but the point is that trying to remain in hiding and rapidly accumulating risk resulting in negative rerolls for a nighblade is not a decision. You just run and wait because by the time you can attack again, you have to wait down 100 risk, -2 rerolls, and 30-60 seconds of paranoia. The risk also carries to the next mob. It doesn't drop when you leave the room so don't get this. Basically, you're saying that we should never attempt to attack if an enemy has paranoia. I'm going to be blunt and say that is not fun at all. If we're forced to run, wait, and come back... then that's going to get boring fast.
It was definitely 8, as shown here in my version control commit. Risk does tick down every second so it may have gone down a bit before you noted it, maybe?
Maybe
The difference is that the duelist is engaging them in actual combat and analysing their combat movements and capabilities. The Nightblade setting up a Mark is not watching the creature's actual combat actions (unless it's attacking someone else I suppose, but that's not a consideration for the mechanics of the ability), but waiting for the perfect moment when the target is not at guard and has an exposed weak point. And don't forget that Mark is not only getting Nightblades an offense reroll, but pushing down the target's defense rerolls on top of the usual stealth pushdown. Nightblades are often getting the double benefit of not only taking the best of their own offense rolls, but forcing their target to take the worst of defense rolls as well.
Importantly, nightblades are also not able to drop opponent rerolls by 5 to 8 due to sweep/tackle/stagger/other during this time so the reroll pushdown on an ambush is far less than one would typically see before the kill shot as a duelist. Right now, as a duelist using armor chink with a morale bonus and sweep, I can be looking at 5 rerolls versus -8 rerolls after a sweep. I think we can all agree that this is absurd. It is currently impossible for a nightblade to setup an opponent prior to an ambush. I therefore do not think your statement accurately captures the entirely of the situation.
The concept is supposed to come across as the Nightblade utilizing a risky combat style which gives them an advantage over everyone else at the start, but the risks eventually catch up with them and they start to suffer for it, simulated by the negative they end up in after staying engaged in combat for too long. It's not a 100% accurate-to-life move-by-move execution for sure, but is meant to be a more abstract simulation of the concept. And, of course, it serves as a balance to Nightbaldes starting out with a very powerful +2 rerolls to all attacks that require no time or active buildup to achieve.
I do understand the intent here. If there is no logic behind this and it is just to make nightblades take a hit (most of the time off of non-combat actions), then how can I debate it?
Yes, just not for a prolonged amount of time. Strike hard and fast, do your best to finish the job, but if things draw out too long that you're now suffering negative rerolls, fall back and set up a fresh attack. There's a lot of room between "attack once, then retreat" and "never stop attacking once in the open".
The issue is not remaining in the open right now. It's that rehiding now can give penalties much faster than the +10 risk from attacking in the open. So basically... Im rewarded for staying in the open for 30-60 seconds rather than withdrawing. If I withdraw go back in hiding (on a successful check), I'll be accumulating way more risk than staying in the open and waiting an extra couple seconds between attacks.
Playing my Nightblade I'm able to:
- Make an initial ambush with +2 rerolls to offense and defense, plus whatever pushdown the ambush does to enemy rerolls. +3 rerolls and even more pushdown if I use Mark. Then it's open combat where I can:
- make 3 attacks while enjoying +2 rerolls to offense and defense,
- then make 3 attacks while enjoying +1 reroll to offense and defense,
- then make 3 attacks at no bonus to offense and defense.
Again, this is in the open and not relevant to the impact of the changes (rapid risk accumulation due to paranoia). I would suspect that nightblades would disengage long before this and the point is that we should be able to do so without running to the next room assuming we pass the stealth check for rehiding.
So once I get to my 11th attack, that's when I start feeling the negatives. If I as a Nightblade am unable to take an opponent down after an ambush and 9 additional attacks, I'm going to slip away and make a fresh approach. Preferably an ambush. 10 attacks at 5 RT each is 50 seconds, which means in 10 more seconds the paranoia from the initial ambush is going to be gone.
I would hope it's gone by then. I guess we're supposed to stay in the open for 10 attacks to wait this down then? And then wait down the 60+ accumulate risk after rehiding, right? Again, I'd rather only go a couple rounds in the open before disengaging. Is the intent for us to stay in the open for this long?
That is a good point.
Thanks. Just trying to say that this is an addition advantage of ranged so even when a non-nightblade ranged stealth user hits high risk, its way less of an issue.
This statement made me particularly worried, so I wanted to get a more accurate picture of what your specific experience was like. I took the liberty of cloning your Nightblade character on my local copy of the game server and took him out for a wechuge hunt. Out of 20 wechuge, 3 detected me: 1 when sneaking into the room, 1 on a random search, and 1 after stalking it. The rest were the same: Sneak into the room, stalk twice, mark, attack. 14 of the 17 I set up undetected died in a single blow prompted by Mark. 2 of those that detected me, I took down after a few aimed attacks to the chest. The remaining one I ran away from after losing my positive rerolls to Risk, and came back to stalk, mark, and one-shot later. In the end I had a slight wound to my arm as evidence of my trek into dangerous wechuge territory.

After that experience I'm feeling that combat is very manageable as a Nightblade after these updates, and I'm wondering if possibly we have staggeringly polar opposite RNG luck. Are you having a different experience?
I had a very different experience with risk accumulation on easy mobs but haven't yet tried wechuge since the changes. Good to know I can still kill them efficiently. In the past though, I have had a vastly difference experience with multi-searches.
Yes. It's also basically guaranteed that someone with 0 Melee skill is going to hit someone with 700 Dodge skill after x attacks. With the RNG everything is going to happen eventually. It's the nature of how the roll system is set up in this game. I get the concern, but just like a skilled warrior shouldn't assume he's untouchable just standing passively in front of someone who's attacking him, a skilled stealth ... ior ... shouldn't assume he's unfindable just sticking around while someone is continuously searching for him.
This is more of a question of rate of being found than being eventually found. Yes, I expect to eventually be found, but no... I do not expect to be found after 3 or 4 searches... although it may happen with RNG. That being said, it shouldn't be the norm that high stealth users are detected quickly by low perception searchers based on rapid risk accumulation. I expect to be seen. That also doesn't mean that I should no longer be able to hide after being seen because my risk went way up due to repeated searching.

Rias, I appreciate you testing this. Again though, this is a massively disproportionate nerf to nightblades and I have to assume that this was not intended so pointing it out. I do feel that nightblades were targeted with this change when it was initially intended to balance ranged stealth (maybe not?). Also, I hope you enjoyed playing Crimson and think he's in a good place. He has certainly been fun.
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
User avatar
Rias
DEV
Posts: 2024
Joined: Sun Sep 03, 2017 4:06 pm
Location: Wandering Temicotli

Re: Stealth and Risk

Post by Rias »

I wasn't aware that Mark had a check associated with it... cool... so 4 checks. The problem comes from any failed check resulting in a rapid increase in risk. I am less concerned with the duration.
Is this a theoretical concern? I rarely get noticed going up against similar-skill targets. See my wechuge experiment, and wechuge are considered some of the more perceptive ones.
This actually motivates characters to stay engaged instead so they don't start taking risk penalties if they try to rehide.
If an ambusher has been detected, they should indeed either engage and stay engaged, or withdraw and come back after the paranoia has ended. Going back into hiding and trying to use stealth against a paranoid target is not recommended. That's the point of the paranoia effect.

When playing my Nightblade, once something starts searching for me I attack or I leave. I don't wait for it to increase my risk for several rounds by letting it search for me. Tip: Attack them right after they search and are in roundtime with lowered defense. An ambush without Mark is still quite effective.
Other characters in a hunting area frequently trigger paranoia.
The joys of sharing combat zones. Ask them to be more careful, ask them if they want to team up, engage in open combat against paranoid targets, go hit another zone for a while.
If I steal from one person, I dont see why I would have greater risk when I move a room over and steal from the next guy.
Risk is an abstract stat. Think of if as pushing your luck a bit more each time you perform or are affected by a "risky" action within a short timeframe. You're not pushing your luck specifically in relation to another person. I understand if you don't like the concept, but it's implemented and working as intended.
trying to remain in hiding and rapidly accumulating risk resulting in negative rerolls for a nighblade is not a decision.
It is. Nightblades should take action or leave before their Risk accumulates further. Simply staying in hiding while someone searches for them is a decision.
You just run and wait because by the time you can attack again, you have to wait down 100 risk, -2 rerolls, and 30-60 seconds of paranoia.
Why wait until 100 Risk? I have literally never had my Nightblade's Risk get that high other than when deliberately allowing it to happen for testing purposes.

Why refrain from attacking due to paranoia? The paranoia effect means I attack earlier and don't get the full spectrum of bonuses available to me, but I don't expect to have a perfectly-stacked deck in every single situation. If I haven't managed to stalk to Melee range, I just leave and find another mob and come back later.
Basically, you're saying that we should never attempt to attack if an enemy has paranoia.
No, I am not. I have repeatedly said that Nightblades should choose to either attack, or leave and come back later in this situation. I've had great success attacking paranoid targets and just relying an the "regular" +2 rerolls and occasional ambush pushdown on a good roll despite paranoia. For players that refuse to attack paranoid targets because they're not at peak performance and therefore have to spend time waiting the paranoia out, that's their decision and that's on them.
(Things about Duelists)
I would recommend you roll up a Duelist character and try them out before making further comparisons between them and Nightblades. My Nightblade has had a far easier time landing hits, blasting through enemies, and leveling up than my Duelist. Duelists are absolutely very cool and get some reroll/"de-reroll" management, but they are more tedious to play and far more prone to injury and death in my experience.
Im rewarded for staying in the open for 30-60 seconds rather than withdrawing. If I withdraw go back in hiding (on a successful check), I'll be accumulating way more risk than staying in the open and waiting an extra couple seconds between attacks.
Yes. Once a Nightblade has been detected, immediately re-hiding and attempting stealth against the same target is intentionally discouraged. They are very capable at open combat.
I would suspect that nightblades would disengage long before this and the point is that we should be able to do so without running to the next room assuming we pass the stealth check for rehiding.
No. Nightblades are discouraged from immediately re-hiding and attempting stealth against a target that is paranoid. This is when a Nightblade should engage in open combat. They are very capable at open combat.
Is the intent for us to stay in the open for this long?
Yes, or if that is not one's preference, leave the room and the target completely and come back at them later. Not immediately. Hop to another mob or wait it out in another room, that's up to the player. Personally I have had great success in simply engaging in open combat when discovered or dealing with paranoid targets.
I had a very different experience with risk accumulation on easy mobs
Out of curiosity, how are you getting such high Risk accumulation when going against things that are below your skill level? My Risk level stayed low enough that I never felt the need to retreat from an encounter in all but 1 of my 20 wechuge encounters in my most recent test with your character. I have had a similar experience with my own Nightblade going against like-skill targets. I can't imagine it would somehow be worse against enemies of less skill than the Nightblade, but maybe something is buggy.
Again though, this is a massively disproportionate nerf to nightblades
I'm really not feeling that. My tests with both my and your Nightblade characters show they still easily blow through enemies with ease compared to my Duelist, Marauder, Dreadnought, and Guardian, whose encounters feel longer (this might be due to not one-shotting most of the time even if they aren't actually longer - waiting for Mark to kick in sometimes takes a bit) and all of whom take more frequent hits and have to go get patched up at the infirmary. Try going out and hitting those wechuge when you're ready to play again, and see if it feels like a massive nerf in experience to match your expectations.
<Rias> PUT ON PANTS
<Fellborn> NO
Acarin
Posts: 202
Joined: Fri Sep 04, 2020 10:49 pm

Re: Stealth and Risk

Post by Acarin »

Thanks for taking the time to clarify and test. I'll just assume that I have had extraordinarily bad luck then with being seen and inducing paranoia... and that I am probably not playing my nightblade as intended. Since there are clearly no issues here and everything is functioning as intended, I will move on.
You reach toward ((DEV Rias)) ... Pull(d225([1]x)):214 vs Mark(d1100):714 = -500 (-222%)
You notice ((DEV Rias)) glance your way, causing you to quickly withdraw your hand from his wool
drawstring pouch (open).
Roundtime: 5 seconds.
Post Reply