test content
Logo
What is the Arc Client?
Install Arc
Damage Resistance and Runaway Returns: A Study of a Mod 6 Anomaly (Long, Math)
a097832846
Member Posts: 13 Arc User
TL;DR version: HP is too good as a defensive stat, but reducing the ratio of HP to other stats on gear makes defense too good as a defensive stat. Deflect is always inferior as a defensive stat.
Premise 1:
Stats in Neverwinter can (and generally do) have two kinds of diminishing returns: absolute and relative. Absolute diminishing returns apply when each point of a statistic provides a smaller contribution to the associated rating percentage than the previous point.
Example (On live):
2900 -> 3000 defense increases your damage resistance from 16.4% to 16.69% (.29%)
3000 -> 3100 defense increases your damage resistance from 16.69% to 16.96% (.27%)
Relative diminishing returns are when each increase in a rating is a smaller portion, relative to the whole, than the previous increase.
Example (On live):
2900 -> 3000 power increases your damage output from 117.4% to 118% (.511% improvement)
3000 -> 3100 power increases your damage output from 118% to 118.6% (.508% improvement)
Premise 2:
Defense, deflect, and hp as defensive stats can be usefully compared by converting into "effective hp" using the following formula:
H / (D * (1-S) * (1-M) + (1-D) * (1-M))
where: H = hit point
D = deflect chance
S = deflect severity
M = damage resistance
So, if one has 25000 hit points, 15% deflect chance, 50% deflect severity, and 20% damage resistance, the effective hp can be calculated as:
25000 / (.15 * (1-.5) * (1-.2) + (1-.15) * (1-.2))
giving an effective hp value of ~ 33784. This means a character with these stats could survive as much damage on average as a character with no deflect or damage resistance and 33784 hit points. Note that the assumption here is that deflect will always provide its average expected defensive value (obviously false in practice, but can be assumed to average out this way over an infinite number of hits).
Premise 3:
A couple of unusual effects arise from this comparison. The first is that the more HP you have to work with, the better the returns from increasing defense and deflect. This makes sense when you consider that these stats are working as a multiplier on the base value of your hit points. The second is that unlike most statistics, each percentage point of increase in deflect chance and damage resistance gives a better relative return than the last. This seems counterintuitive, but consider what damage resistance actually means: less damage taken. If you think of 20% damage resistance as 80% damage taken, you begin to see the source of the improved relative returns. Going from 20 to 21% resistance is really going from 80 to 79% damage taken, a 1/80th improvement. Going from 21 to 22% resistance is really going from 79 to 78% damage taken, a slightly larger 1/79th improvement.
This is mitigated on live by the absolute diminishing returns of defense and deflect - eventually these stats contribute so little to the overall damage resistance equation that the relative returns begin to diminish for a constant increment of statistic.
Premise 4:
With the new linear ratings formulae for statistics on test, the absolute diminishing returns are removed, and the increasing relative returns are not compensated for, meaning each point invested in defense or deflect stats is better than the previous point. Consider the situation for this Guardian Fighter character (most recent proposed stat formulae, 400 stat = 1% rating):
Deflect chance: 9.10% (2440 deflect, 3% from Dex), 50% severity
Damage resistance: 35.18% (7441 defense, 16.5% from AC)
Hit points: 51854 (effective HP 83807)
adding 1 point of deflect increases effective HP by ~1.1 and each point is better than the previous point
adding 1 point of defense increases effective HP by ~3.24 and each point is better than the previous point
adding 1 hit point increases effective HP by ~1.62 and each point increases the returns on future points of defense and deflect.
This character has already reached the point where the returns on defense have begun to "run away," increasing faster than the other stats can catch it. Since there is a hard cap on damage resistance (80% on live, and 100% at the worst since mitigating more than 100% of an attack is meaningless), defense will eventually stop giving returns. Deflect will likewise increase in value until 100% deflect chance is reached. HP, at a 1:1 (or even 2:1) ratio will never have a chance to catch up until the other two stats hit absurd hard caps.
We know currently HP is typically provided on gear at a 4:1 ratio with other stats. This makes it the only defensive stat worth having on live, and in this example and at this ratio it continues to provide superior returns until around 103250 HP, at which point defense returns overtake even these huge HP returns.
Conclusions:
A) HP, at the current ratios, continue to be the dominant defensive stat. I and others have made the suggestion to reduce the ratio on gear to better balance with other defensive stats. I have established that this may not work, however, because:
Defense will quickly overtake HP at a lower ratio, and once it does it will effectively never be worth investing in the other two stats again.
C) Deflect is always inferior to both of the other stats by it's very nature. Increasing deflect severity helps a little, but not enough to make it worthwhile, and even boosting it to 100% severity you'd just be creating another problem stat like defense, albeit a rather spiky one.
Things not taken into account for this analysis:
Damage resistance is reduced by the foe's resistance ignored rating. Do mobs even have this rating?
Deflections may be unaffected by resistance ignored and therefore have utility in situations where damage resistance is negated. Need to confirm.
Deflect has utility as a DPS source for some classes/builds that is outside the scope of this analysis.
Damage is generally not dealt in an unending stream of tiny hits. Stacking deflect or HP could provide a chance of surviving a huge attack that defense alone would not.
Premise 1:
Stats in Neverwinter can (and generally do) have two kinds of diminishing returns: absolute and relative. Absolute diminishing returns apply when each point of a statistic provides a smaller contribution to the associated rating percentage than the previous point.
Example (On live):
2900 -> 3000 defense increases your damage resistance from 16.4% to 16.69% (.29%)
3000 -> 3100 defense increases your damage resistance from 16.69% to 16.96% (.27%)
Relative diminishing returns are when each increase in a rating is a smaller portion, relative to the whole, than the previous increase.
Example (On live):
2900 -> 3000 power increases your damage output from 117.4% to 118% (.511% improvement)
3000 -> 3100 power increases your damage output from 118% to 118.6% (.508% improvement)
Premise 2:
Defense, deflect, and hp as defensive stats can be usefully compared by converting into "effective hp" using the following formula:
H / (D * (1-S) * (1-M) + (1-D) * (1-M))
where: H = hit point
D = deflect chance
S = deflect severity
M = damage resistance
So, if one has 25000 hit points, 15% deflect chance, 50% deflect severity, and 20% damage resistance, the effective hp can be calculated as:
25000 / (.15 * (1-.5) * (1-.2) + (1-.15) * (1-.2))
giving an effective hp value of ~ 33784. This means a character with these stats could survive as much damage on average as a character with no deflect or damage resistance and 33784 hit points. Note that the assumption here is that deflect will always provide its average expected defensive value (obviously false in practice, but can be assumed to average out this way over an infinite number of hits).
Premise 3:
A couple of unusual effects arise from this comparison. The first is that the more HP you have to work with, the better the returns from increasing defense and deflect. This makes sense when you consider that these stats are working as a multiplier on the base value of your hit points. The second is that unlike most statistics, each percentage point of increase in deflect chance and damage resistance gives a better relative return than the last. This seems counterintuitive, but consider what damage resistance actually means: less damage taken. If you think of 20% damage resistance as 80% damage taken, you begin to see the source of the improved relative returns. Going from 20 to 21% resistance is really going from 80 to 79% damage taken, a 1/80th improvement. Going from 21 to 22% resistance is really going from 79 to 78% damage taken, a slightly larger 1/79th improvement.
This is mitigated on live by the absolute diminishing returns of defense and deflect - eventually these stats contribute so little to the overall damage resistance equation that the relative returns begin to diminish for a constant increment of statistic.
Premise 4:
With the new linear ratings formulae for statistics on test, the absolute diminishing returns are removed, and the increasing relative returns are not compensated for, meaning each point invested in defense or deflect stats is better than the previous point. Consider the situation for this Guardian Fighter character (most recent proposed stat formulae, 400 stat = 1% rating):
Deflect chance: 9.10% (2440 deflect, 3% from Dex), 50% severity
Damage resistance: 35.18% (7441 defense, 16.5% from AC)
Hit points: 51854 (effective HP 83807)
adding 1 point of deflect increases effective HP by ~1.1 and each point is better than the previous point
adding 1 point of defense increases effective HP by ~3.24 and each point is better than the previous point
adding 1 hit point increases effective HP by ~1.62 and each point increases the returns on future points of defense and deflect.
This character has already reached the point where the returns on defense have begun to "run away," increasing faster than the other stats can catch it. Since there is a hard cap on damage resistance (80% on live, and 100% at the worst since mitigating more than 100% of an attack is meaningless), defense will eventually stop giving returns. Deflect will likewise increase in value until 100% deflect chance is reached. HP, at a 1:1 (or even 2:1) ratio will never have a chance to catch up until the other two stats hit absurd hard caps.
We know currently HP is typically provided on gear at a 4:1 ratio with other stats. This makes it the only defensive stat worth having on live, and in this example and at this ratio it continues to provide superior returns until around 103250 HP, at which point defense returns overtake even these huge HP returns.
Conclusions:
A) HP, at the current ratios, continue to be the dominant defensive stat. I and others have made the suggestion to reduce the ratio on gear to better balance with other defensive stats. I have established that this may not work, however, because:
Defense will quickly overtake HP at a lower ratio, and once it does it will effectively never be worth investing in the other two stats again.C) Deflect is always inferior to both of the other stats by it's very nature. Increasing deflect severity helps a little, but not enough to make it worthwhile, and even boosting it to 100% severity you'd just be creating another problem stat like defense, albeit a rather spiky one.
Things not taken into account for this analysis:
Damage resistance is reduced by the foe's resistance ignored rating. Do mobs even have this rating?
Deflections may be unaffected by resistance ignored and therefore have utility in situations where damage resistance is negated. Need to confirm.
Deflect has utility as a DPS source for some classes/builds that is outside the scope of this analysis.
Damage is generally not dealt in an unending stream of tiny hits. Stacking deflect or HP could provide a chance of surviving a huge attack that defense alone would not.
Post edited by a097832846 on
0
Comments
Well done.
Three types of damage resistance (although HP is technically not one) only make sense if there are attacks that can ignore each one.
We have attacks that negate deflect (prones) and limited piercing and DR reducing attacks. What's missing is something that hits for like 30% of your HP or whatever.
So given these facts, HP is even stronger than you suggest.
So if
"adding 1 point of deflect increases effective HP by ~1.1 and each point is better than the previous point
adding 1 point of defense increases effective HP by ~3.24 and each point is better than the previous point
adding 1 hit point increases effective HP by ~1.62 and each point increases the returns on future points of defense and deflect."
Your not adding 1 HP your adding 4 HP meaning its 6.48 EHP which is TWICE as good as defense?
Most calculations I have run, defense is actually a very poor stat to stack. Also whats funny is you have nearly 8k Defense but ONLY 35.18% DR!!!!!!!!
Talk about crappy DR. Youll nearly ALWAYS be mitigated down to zero via ARP in PVP. I hope they adjust AC to provide MUCH more DR % to compensate tanky classes.
Not sure about everything else, but my Pally has a chest piece thats like lvl 30 with 11 AC on it, ive never played GF so i dont know if thats just the way it is with PLate Gear etc. but just thought id mention that
somethign needs to be confirmed, I was told that Armor Pen would continue to be useful even past teh targets DR%
So if I have 50% Armor Pen, and you have 0 DR, I am still gonna do another 50% of damage to you, where as in Live, all that extra armor pen is useless if the target has no DR to deal with.
I dont know if this is correct however, as i mentioned, it was what the Devs seemed to be hinting towards, Armor Pen Infinitely being useful no matter how much u stack and no matter who u attack.
if you need something worth testing, you shoudl test how much more damage Armor Pen can do vs targets with varying levels of DR to see if what the devs mentioned is in or not
Well the only side comment is that defense gives an "increase" to healing on EHP
Meaning if you are healed for 10k with 50% Dr that effective heal is really more like 20k Heal, where as if you have only 25% DR your effective heal is for like 13k
So "self healing" as well as actual healing are much more effective with DR. So if you do look over a longer term fight I wonder how much healing/duration it would take them for it to even out.
In PVP though with classes stacking ARP like mad, DR will be worthless and most people will just stack HP.
I dont know about 200:1 but maybe putting defense back at 300:1 might make sense there. Just because of the healing effect on EHP.
What this means is HP will still be better for RAW EHP, but when healing is involved you will probably be better with DR.
What would come next is Deflect since 400=1% is very poor considering Deflect is on "par" for .5% DR.
So if Defense is 300:1 Deflect needs to be about 150:1. However with TR having such high severity, maybe it needs to be catered to TRs and put it at 225:1 on Deflect.
Currently, it is 50% severity for TRs on preview, with absolutely no response to queries as to whether this is a bug or an intentional change.
Neverwinter Census 2017
All posts pending disapproval by Cecilia
Im not surprised there is no response, either itll get fixed in an new patch or itll be release before they announce a reason why.
If this is true, then deflect @ 150:1% seems fair with 300:1% Defense:DR
That way all stats seems "inline" with EHP (HP/Deflect/Defense)
It's not only ArPen though.
As I stated, Deflect and Defense can be offset by prones (PVP/PVE) and piercing damage (PVP). That's why they are even less desirable. As you've said, simply tweaking the ratio won't help, we need an attack that directly aims at HP. If mobs or players had attack that would eat a percentage of HP mitigated by DR and Deflect, all those shiny HP won't help, you'd need to stack the other stats anyway.
If HP is the dominant Defensive stat, is that really a problem? Because you only get SOME choices on stats, while others might be pretty much set.
If a piece drops with X Defense Y HP and Z Deflect, it's a more defensive piece that something that drops with X Defense Y Power and Z Critical hit.
Is it really bad to have a system where (as far as defensive stats go) HP>Defense>Deflect, with the understanding that really tanky classes are going to have a healthy mix of all three?
Nonoononononononononono Piercing damage is the WORST thing to ever hit this game - arguable atleast on par with RP for arti-gear.
We dont need more attacks that bypass DR causing coding and balance problems.
I think people are also underestimating ARP. ARP is going to be SO much easier to stack, even now you can get to 30% ARP very easily. For example a GWF with 25 CON has 15% built in.
Even if a defensive character stacks 12k Defense, thats only 30% Defense. It will only take you about 3k ARP to completely negate that.
I also DONT like HP being dropped in effectiveness due to PVP concerns (people will 1 shot everyone). I think what HAS to happen is Defense needs to be brough back down to about 300:1%.
Even if Defense is the "go to" stat, thats fine since it can be mitigated. Most of my calculations assume for PVP they have 30% ARP.
I mean even at 300:1% that means youll need 10k Defense to get to 33% DR. Thats pretty high..... Unless they are going to start throwing defense at us like crazy and give armor MORE value to defense. that may be possible but it just means defensive slots will be ALL Radiants.
I dont like the idea that EVERYTHING is 400:1 In fact I really hate it. I like the idea of flat bonuses for the most part.
Defensive Slots:
HP>Defense>Deflect. We have already stated HP is about 2x as effective as Defense which at 400:1 is also 2x better than deflect (50% severity) So what makes SENSE to me:
Defense moved to 300:1%
Deflect moved to 150:1%
Offensive Stats:
Power/ARP/Crit/Recovery
Ill leave recovery alone for now but....
If ARP is 100:1% This means Armor Pen is around 700% MORE effective than Power. Here is why that is.
If PVE mobs have 80% DR. Lets say you have 40% Armor Pen. You attack with 10k base and they net out another 40% DR that was not taken out by your ARP leaving 6k Damage.
Well... You can get 100 points in ARP and this gives you 39% ARP, meaning your NET damage is now 6,100. You just buffed damage by 100 points which was a 1.667% damage boost
Well how much POWER do you have to add to give 1.667% dmg boost NET?
Its not 400 power because thats only a 60 damage boost (1% dmg boost) so if you need a true 100 dmg boost NET youll need about 668 Power to EQUAL 100 ARP
Well what about Crit?! Crit is even worse since MOST people (not including vorpal) have less than 100% crit severity. So 400 Crit ends up being about a .75-.85% DPS boost.
So using .75% DPS boost. You will now need about 890 CRIT to equal 100 ARP
ALL this means is that NOONE will be using Radiants or Azure in offensive slots. Everyone will be trying to pick up crit via feats/features and maybe even Stats like "Dexterity".
I am very surprised they are taking this approach TBH, basically for PVE is Dark Enchants or go home, you will need to stack upwards of 6k+ ARP to be any good in PVE.
The bad to this is that in PVP it will basically "wash" all classes to being equal.
A GF wont have any more DR than a CW. In truth the CW will be much tankier because of "shield"
Since in PVE players will need to stack as much ARP as possible. In PVP players will stack 40%+ARP and negate nearly every classes DR.
So attacking a TR or a GF will be the same end net result. Thats the bad.
By making Defense stronger, people now have to make a choice of stacking it more which it MAY get negated with ARP or not.
If a GF can get back up to 50%+ DR in PVP (remember Tenacity is separate and cannot be mitigated by ARP) Then he stands a real chance to have MORE DR than his opponent will ARP.
I do not remember if it was the " Abaddon " or " points" or " iron" that said: "What will happen is that or invest in X or Y". Good in fact this will cause chars have more " specialized " and with that maybe a tradeoff between the stats.
In the background is not and has never been reasonable to allow in a char had plenty of everything in terms stats. Is not reasonable as some argue here that a tank is immortal and still have 10-12k pow and still lf , deflect and absurd DR neither one nor the other tank class!
I also agree that we need to make some "adjustments" to all stats have a role within the entire system.
...my two cents and bad congratulations contribution.
ps .: these super technical posts and with a lot of math : Serious must be some disguised dev ! rofl
Currently I'm running next to 60% total DR(if you ignore the deflectrate) with defense + feats + AC. In PvP it's mostly worth 30% or less. For over 5k stats. Now with the new calculation for 70, it will be next to 45% or so for my 5k defense I have. So I need to get over 50% more defense to be at the same level as one with 4k offstats???
Anyways, this means that it is 4 times easier to stack enough armor pen get through whatever amount of defense you can come up with in mod 6. This 4 to 1 ratio might be ok with pve due to the fact that top tier dungeons will have up to 90% DR potentially (as mentioned on the live stream introducing the paladin) but in pvp it will make things very difficult. So stack as much HP as you can for pvp because defense will not matter at all.
In fact, assuming that Armor pen can still not bring you into negative DR territory I would consider stacking deflect over defense if I had the choice in mod 6 because armor pen can't negate the EHP bonus that deflect gives you.
We are searching for slave labor, will pay with food from our farm!
Then something like ARP in PVP is 400:1 just like other stats, but then in PVE it could be the 100:1