54 réponses à ce sujet

[Veovim]

Sorry, everyone was only looking at half the mechanic for that damage boost. The system that checks whether enemies are near you does stack but the math coming back does not. Thus having two effects with this buff will not increase your damage output but will increase the reliability of that damage boost (though it'll be negligible in most real cases). IF, hypothetically, a new item had this effect but at a different value - say 10% per enemy instead of 5% - that would stack.

 

As for other effects stacking or not this is mostly a case by case basis.

What do you mean by the reliability of the damage boost?  Does the distance that gets scanned for enemies stack with multiple items, so two items with +damage for each enemy within 8m would be equivalent to one item with +damage for each enemy within 16m; or is the code that scans for enemies unreliable, so wearing two items with that effect gives enemies a second chance to be detected if they're missed on the first scan?

[Drasca]

Currently there is a 'soft' cap for the benefits due to the hard cap in place for defense which is 80%. As for the other benefit of each stat those currently increase indefinitely. So depending on your class/passives/gear you'll start getting less out of them somewhere in the 100-150 range.

 

To expand, there's two classes of benefits, basic survivability, and offense. Survival broken down to health which has no cap and defense, which caps at 80%. Offense into attack power (willpower) and critical chance (cunning) which caps at 100%

Note promotes are not the only way to increase critical chance, or to reduce damage.

 

Health

Got enough health, you're alive, don't have enough, and you're dead. However, with enough heal on kill, offense and defense combined, any additional health is less useful because you're already surviving. Or if you have barrier all the time, health is not useful. That said, more health allows you to be more lazy with your playstyle and there is no limit except time.

 

Defense

melee, ranged, magic. These are capped at 80%. Note magic resistence is seperate from magic defense.

 

There is no gain in defense benefit beyond certain plateaus, and less relative benefit even with absolute gain in stats indefinitely, depending on class/gear/passives/conditions.

 

Defense plateaus being different for classes example, a 1H warrior can gain Turn the Bolt passive, which raises their ranged defense by 50 and is additive with front defense of shields. From the front condition and 1H Warrior class, there is no additional defense benefit from cunning.

 

Some classes get cunning or consitution or willpower passives, and they each add to the promotes, and their respective defenses, lowering the overall requirement to maximize benefit.

 

Offense

Offense is broken down to added critical chance, which benefits critical damage bonus and passives, and attack power.

 

Attack power affects every attack ability differently on the damage formula from trivial benefit (1% to 10% gain in final calculated at most for a 100-200 willpower gain) in abilities with lots of 'damage bonus' on the ability itself (Power Chord, Longshot) to 1:1 gains in abiltiies that are pure weapon damage (Bolt)

 

Critical Chance: There is mathematical proof on geometrically diminishing relative benefit to critical chance. The damage component of critical chance however can be boosted with Critical Damage Bonus gear, maxing Critical Chance more valuable as CDB gear and dexterity passives (boosting critical damage) gets equipped.

 

You'll need more cunning to gain the same amount of relative benefit gained previously for any set amount. The benefit from going from 5% to 10% critical chance is a factor x2, and requires 10 cunning. To gain the same factor, you need to go from 10% to 20%, requiring 20 cunning. So on and so forth, the relative marginal benefit reduces.

 

Classes that have automatic critical chance passive benefit less from higher cunning as critical chance becomes guaranteed. Sneak attack and Opportunist benefit the most at about the 50-100 range. Mercy killing / Death Blow does not benefit at all.

 

Critical chance dependent passives sometimes have internal cool downs, such as Flashpoint's 10s cooldown. Once critical occurs every 10 seconds, additional critical chance is less useful. Also less useful once criticals occur frequently enough to overlap.

[Rolenka]

Currently there is a 'soft' cap for the benefits due to the hard cap in place for defense which is 80%. As for the other benefit of each stat those currently increase indefinitely. So depending on your class/passives/gear you'll start getting less out of them somewhere in the 100-150 range.

 

Thanks!

[Angelus_de_Mortiel]

Critical Chance: There is mathematical proof on geometrically diminishing relative benefit to critical chance. The damage component of critical chance however can be boosted with Critical Damage Bonus gear, maxing Critical Chance more valuable as CDB gear and dexterity passives (boosting critical damage) gets equipped.

 

You'll need more cunning to gain the same amount of relative benefit gained previously for any set amount. The benefit from going from 5% to 10% critical chance is a factor x2, and requires 10 cunning. To gain the same factor, you need to go from 10% to 20%, requiring 20 cunning. So on and so forth, the relative marginal benefit reduces.

 

You entire post was very informative and well written for the most part.

Although, trying to read this from the perspective of someone that is a newer player and/or may not "speak" math (different than being able to "do" math), these two paragraphs were a big vague and or confusing. Now, bear in mind what I am about to say is not attempting to discredit the value in what you say, just helping it along. I do not want this to become a big debate. I am just relaying how I am reading things.

With Critical Chance specifically, I honestly am having difficulty processing "relative" results, since it is a linear (not curved) scale.

 

Overall, this second paragraph about Critical Chance seems like a completely unavailing and contrived statement that attempts to marginalize a linear result through relative comparison. Not insulting. I'm just laying out how it appeared.

That's not really how it works, as it is like saying that when running a mile, the distance from the 0.25 mile mark to the 0.5 mile mark is the same distance as the succeeding 1/2 mile in total because the rate of distance from 0.25 to 0.5 is 2x, as is the rate from 0.5 to 1. That's just not how that works with Critical Chance, as each 1% of Critical Chance gained is a 1% increase in the chance for a Critical Hit.

 

With Critical Chance, would it not be better to use data comparative to your very accurate point about the Legionnaire? Critical Chance gained from promotions is relative to the character/gear: Characters/gear have ways to boost Critical Chance either outright or through bonuses to Cunning?

[Drasca]

That's not really how it works, as it is like saying that when running a mile

 

If you cannot handle the math, you're not going to understand directly.

 

If we're down to analogies, it is closer to comparing time required to be faster rather over a set distance. It takes more than twice the effort to get twice the speed. You get diminishing returns for your effort.

 

Let's say it takes you 3 hrs a week of exercise to maintain a light jog over the course of a partial marathon. To get faster, you need to put in a lot more effort, and to be in the top running, you need to spend most of your time everyday, consumed in nutrition, balanced meals, exercise plans, etc.

 

You need more effort to gain ever smaller results, as you improve. The biggest improvements are at the beginning with the least relative effort.

 

If you insist on using a "when running a mile," I actually know quite a bit about people that run a mile, a mile and a half, etc and marathoners. It takes training to do it, because people run out of their initial reserves of short-term energy quickly. To be able to even run a mile at a minimum pace, you need to train, and it takes more effort to improve distance just as it does speed when you start from people that don't have any fitness at all. Extremely fit people are already committed, and there's a limit to how much time they can spend.

 

The absolute goals, the absolute distance or speed differences, may be the same, but the effort required to get there are not. In fact, you'll never get over 100% critical chance, as I'm fairly certain you'll never run 50 mph*, or faster than the fastest human alive , or roughly 28 mph on level surfaces, so the effort required is impossible.

[Texasmotiv]

He's not arguing that 10/5 and 20/10 don't equal 2. Merely questioning the relevance in the context of crit rate. Just because you double your crit going from 10 cunning to 20 cunning doesn't mean getting another 10 cunning (up to 30) doesn't yield the same increase in real damage. It's still an extra 5% crit and you still get .5% for each point up to the cap.

Like Mortiel said. It's a linear gain to the cap. The fact that it isn't an exponential gain is largely irrelevant outside of assuming a player doesn't realize the difference until they hit these arbitrary "double up" thresholds.

[Denrok1]

Currently there is a 'soft' cap for the benefits due to the hard cap in place for defense which is 80%. As for the other benefit of each stat those currently increase indefinitely. So depending on your class/passives/gear you'll start getting less out of them somewhere in the 100-150 range.

 

What about cunning Luke? I assume 100% crit is the max thus 210 is the hard cap correct?

[Texasmotiv]

There is a base 5% crit I believe so you are shooting for 200 cunning.

[Texasmotiv]

...For lvl 1 with no crit on your gear that is.

[Denrok1]

There is a base 5% crit I believe so you are shooting for 200 cunning.

 

Luke wrote there is a soft cap for all stats which led me to believe cunning has another component other than crit chance and ranged defense. That would be news to me though.

[Texasmotiv]

By soft cap he is referring to when the stat no longer provides defensive benefit. Cunning is hard capped at 200 because at this point it gives NO benefits all.

[Drasca]

Like Mortiel said. It's a linear gain to the cap. The fact that it isn't an exponential gain is largely irrelevant outside of assuming a player doesn't realize the difference until they hit these arbitrary "double up" thresholds.

 

You won't 'feel' / notice the smaller gains without comparing these larger relative gains, and effort required to obtain these relative gains is measureable.

[Denrok1]

By soft cap he is referring to when the stat no longer provides defensive benefit. Cunning is hard capped at 200 because at this point it gives NO benefits all.

 

BOOOOOO

[kmeeg]

Sounds like everyone is talking past each other.

 

A 200 damage sword, starting at 5% critical chance and 1.4 damage multiplier

Avg Damage: with 5% crit chance and 1.4 damage multiplier (gain 4)

(200*(1.4-1)*0.05)+200 = 204

 

Avg Damage: with 10% crit chance and 1.4 damage multiplier (gain 8)

(200*(1.4-1)*0.10)+200 = 208

 

Avg Damage: with 15% crit chance and 1.4 damage multiplier (gain 12)

(200*(1.4-1)*0.15)+200 = 212

 

Avg Damage: with 20% crit chance and 1.4 damage multiplier (gain 16)

(200*(1.4-1)*0.20)+200 = 216

 

The gain is x2 from 5% to 10% and x2 from 10% to 20%. So takes more and more effort to get x2 the gain. 

 

• Drasca says gain is x2, which is correct 
• Mortiel speaks of linear scale, which also is correct

 

-> I guess it is all a point of view

[Texasmotiv]

For the record. I understand what Drasca means and will yield to his assertion about which point players will notice these changes. My assertion is that it's too gradual to notice that your crit has increased unless you are periodically polling you performance.

Great information, and well presented Mr. Drasca.

[Drasca]

For the record. I understand what Drasca means and will yield to his assertion about which point players will notice these changes. My assertion is that it's too gradual to notice that your crit has increased unless you are periodically polling you performance.

Great information, and well presented Mr. Drasca.

 

Thanks Texas. You do 'notice' it once you're at around 50-100 on cunning though for assassin and archer, and finally katari. These all start doing criticals automatically with the passives aforementioned.  Approaching the 50% point natively, and 100% on critical chance passive, is a big deal. It gets downhill from there though.
 

Still, I would argue every crit felt is part of 'polling performance'. Lots of things require tools for precise measurement, but you do 'feel' it more at the aforementioned x2 factors. It isn't about average damage but the big numbers are clearly visible.

 

There's also whole idea of the minimum resolution "the smallest interval measurable by a scientific (especially optical) instrument; the resolving power" where you need a minimum size to see. I would argue you would notice certain bigger jumps. . . and in order to achieve those noticeable gains them you do need even more effort to produce them.

 

You might not notice a sprinkle of salt on your meal, but dump the bottle and you'll notice it. Somewhere in between is your actual ability to detect differences in effects.

[Angelus_de_Mortiel]

If you cannot handle the math, you're not going to understand directly.

 

I understand the math just fine, and attempting your usual diffusion of me "not understanding math" will not work. 

 

All I was saying is that Critical Chance is not a geometrically progressive scale, it's linear. Geometric is exponential (bⁿ =  b¹  q ^n - 1). The scale does not depreciation nor appreciation on an exponential scale, and is therefore not geometric. Simple as that. 

 

The reason there is no appreciation nor depreciation since 1 Cunning always returns 0.5% Critical Chance (up to 200 Cunning, of course), you are attempting to marginalize the results as if they did appreciate or depreciate. There will never be an instance where 1 Cunning grants 0.25% Critical Chance, so the results never diminish. This is not a negotiation. It is a fixed variable.

 

And saying it "feels" different is qualitative. A scale can not have geometric progression based on qualitative data of a single data sample (your own), and furthermore shows entirely different results that Cunning/Critical Chance progression. How you "feel" critical hits are landing based upon your promotions does not correlate to the actual linear progression of Cunning/Critical Chance. It is it's own data by itself.

All I was doing is trying to help clarify topics to make it easy to be explained. I am not going to argue any more on this. You made a good post, and had a lot of good data in there. You simply are over-complicating that one facet.

[yarpenthemad21]

Critical Chance: There is mathematical proof on geometrically diminishing relative benefit to critical chance. The damage component of critical chance however can be boosted with Critical Damage Bonus gear, maxing Critical Chance more valuable as CDB gear and dexterity passives (boosting critical damage) gets equipped.
 
You'll need more cunning to gain the same amount of relative benefit gained previously for any set amount. The benefit from going from 5% to 10% critical chance is a factor x2, and requires 10 cunning. To gain the same factor, you need to go from 10% to 20%, requiring 20 cunning. So on and so forth, the relative marginal benefit reduces.
 
Classes that have automatic critical chance passive benefit less from higher cunning as critical chance becomes guaranteed. Sneak attack and Opportunist benefit the most at about the 50-100 range. Mercy killing / Death Blow does not benefit at all.
 
Critical chance dependent passives sometimes have internal cool downs, such as Flashpoint's 10s cooldown. Once critical occurs every 10 seconds, additional critical chance is less useful. Also less useful once criticals occur frequently enough to overlap.

You can think about that in this sense, but crit chance is also a stat which I would call linear in nature.

You can introduce a "multiplier" which I call average crit damage.
It's (crit damage - 1) * (crit chance) + 1
so for basic crit damage, and 50% crit chance it's
(1.4 - 1) * (0.5) + 1 = 1.2
So it feels like it should. half of my hits are regular, half are crit with 40% more damage, so on average I deal 20% more damage.

You simplify it to:
(crit dam bonus) * (crit chance + added crit chance) , we add 1 always So i will skip it.
let simplify it
crit dam bonus * crit chance + added crit chance * crit dam bonus
It's easy to see that additional crit chance bonus depends only of crit chance. So you can for example take 1% as additional and have basic formula
0.01 * crit dam bonus, so for example 0.01 * 0.4 = 0.004
So in that sense we can think about crit chance as linear, where it only depends on crit damage bonus and each 1% add the same additive bonus to calculation.
If we now take this behavior into damage calculation we have something like this:
base_damage * average crit damage bonus.
Because of base math I can now calculate how much damage each 1% gives, for our example it's 0.004 * base damage.
So when base damage = 300 I get 1.2 damage for every 1% crit chance with basic crit damage bonus.
So damage gain is linear in nature.

Where base damage is calculated based on formula using other multipliers aside from crit (so attack, abilities, bonus damage )

[Silversmurf]

My head hurts

[Thadrial]

My head hurts

Lighty rub the sides of your head in a clockwise and counter-clockwise motion with the pointer and middle finger. Then sacrifice a Druffalo and 3 Golden Nugs to gain 10% crit chance. It's simple really.

[Drasca]

saying it "feels" different is qualitative.

 

You've missed the concept of limit of resolution completely. Texas hasn't, he just didn't paint the whole picture. Instruments can only detect changes within a certain size, just as microscopes are limited by their scanning resolution (everything smaller is fuzzy and unuse-able), the primary instrument here is the brain and will only notice bigger changes. To get those bigger changes, you need bigger factors of relative change, which also requires even more effort as mathematically proven.

 

As Texas pointed out, you won't notice the smaller granular changes on a linear scale, without polling your performance with tools outside the human brain to assist you -- barring some savant ability to memorize everything beyond 8 bits without depositing the data elsewhere.

 

You can think about that in this sense, but crit chance is also a stat which I would call linear in nature.

You can introduce a "multiplier" which I call average crit damage.

You simplify it to:

 

This analysis has two major flaw limiting it: We can change the critical bonus damage factor easily, and noticing (or useful) damage isn't about 'average' damage but lethal damage given we can quickly kill most enemies, and how many hits until they're lethal.

 

Texas is quite right about not noticing averages and granular changes too small for the brain to process. However I argue you'll tend to notice big numbers, number of hits (or time required) to kill and dead guys.

 

Big numbers from Crits is self explanatory.

 

Number of hits: If I require 6 hits of 200 damage each to kill an enemy, I'll notice the time required as he's doing lethal damage to me in return i.e. I'm dead before then if I don't dodge / barrier / cover / block or otherwise defend.

 

Big hits and dead guys with number of hits: I notice when people enemies are dead, and when enemy life bars suddenly drop. If I normally require 6 hits of 200 damage each, and I crit twice at 100% CBD, that'll reduce the hits to 4. Crit four times, it'll reduce to 2 (which is a big deal). If I do enough damage to crit with to OHKO dead, I'll notice lethal damage vs normal hits that don't kill them.

 

Its the number of hits and time to kill that matters. You'll notice see this behaviour with rogues, assassins in particular, who are punished (with damage or death) when their initial strike fails to kill.

 

Average damage is not very useful unless we have more traditional MMO style boss enemies where we're just hacking away at a tree of seemingly unlimited health that doesn't move. That doesn't apply in DAMP where enemy health is relatively low, and time to kill is done is recorded in relatively few seconds vs entire minutes (or hours) of attacking. Even our High Dragon runs away after a certain damage threshold.

 

Critical Bonus Damage is not set in stone at base, but variable:

 

Rings, Weapon upgrades, Passives (both dexterity and duelist crow's nest lookout), rendering the relative bonus of CBD and average damage adjusted for crits quite different as we gain progressive upgrades in equipment and passives.

 

As CBD gets to and exceeds 100%, you'll gain a greater than 1:1 benefit on the linear scale for critical chance increases. Assuming the same base weapon damage to reach this CBD%, at 100% Critical bonus damage total, you'll gain 1% average damage for every 1% increase in critical chance. If you have higher than 100% CDB, you'll gain more for your critical chance.

 

The relative scale for benefit still quickly diminishes, just the base benefit starts much higher, with more CBD.

[yarpenthemad21]

This analysis has two major flaw limiting it: We can change the critical bonus damage factor easily, and noticing (or useful) damage isn't about 'average' damage but lethal damage given we can quickly kill most enemies, and how many hits until they're lethal.
 
Texas is quite right about not noticing averages and granular changes too small for the brain to process. However I argue you'll tend to notice big numbers, number of hits (or time required) to kill and dead guys.
 
Big numbers from Crits is self explanatory.
 
Number of hits: If I require 6 hits of 200 damage each to kill an enemy, I'll notice the time required as he's doing lethal damage to me in return i.e. I'm dead before then if I don't dodge / barrier / cover / block or otherwise defend.
 
Big hits and dead guys with number of hits: I notice when people enemies are dead, and when enemy life bars suddenly drop. If I normally require 6 hits of 200 damage each, and I crit twice at 100% CBD, that'll reduce the hits to 4. Crit four times, it'll reduce to 2 (which is a big deal). If I do enough damage to crit with to OHKO dead, I'll notice lethal damage vs normal hits that don't kill them.
 
Its the number of hits and time to kill that matters. You'll notice see this behaviour with rogues, assassins in particular, who are punished (with damage or death) when their initial strike fails to kill.
 
Average damage is not very useful unless we have more traditional MMO style boss enemies where we're just hacking away at a tree of seemingly unlimited health that doesn't move. That doesn't apply in DAMP where enemy health is relatively low. Even our High Dragon runs away after a certain damage threshold.
 
Critical Bonus Damage is not set in stone at base, but variable:
 
Rings, Weapon upgrades, Passives (both dexterity and duelist crow's nest lookout), rendering the relative bonus of CBD and average damage adjusted for crits quite different as we gain progressive upgrades in equipment and passives.
 
As CBD gets to and exceeds 100%, you'll gain a greater than 1:1 benefit on the linear scale for critical chance increases. Assuming the same base weapon damage to reach this CBD%, at 100% Critical bonus damage total, you'll gain 1% average damage for every 1% increase in critical chance. If you have higher than 100% CDB, you'll gain more for your critical chance.
 
The relative scale for benefit still quickly diminishes, just the base benefit starts much higher, with more CBD.

I know that I can change crit damage bonus, it's obvious. What is also obvious is that's the only damage related bonus based purely around gear and maybe few passives (so those few + dex bonuses, not higher than 15% and one passive)

It's also obvious that crit chance on it's own don't give you anything. You need crit damage for it it's also obvious.
So it's obvious that how much crit chance gives relative damage increase depends purely (and what is important only) on crit damage.


Your idea of time to kill and "big" numbers isn't the way to go. Because I could easily prove that even 50% damage increase does not change "time to kill" at all. 100 hp enemy. You hit for 60 damage. You need 2 hits, hits to kill = 2
now 50% damage bonus. Your damage is 90 now. Still 2 hits to kill.
Using this type of idea for damage calculation lead to total mess. You need to take into account enemy health, take crit chance more like probability of crit streaks etc. In this approach even order of crits matter, because most of the time you would prefer crit damage at start (so it woun't be an overkill).
This idea is maybe more realistic but I would say it can't calculated in easy manner at all.

As CBD gets to and exceeds 100%, you'll gain a greater than 1:1 benefit on the linear scale for critical chance increases. Assuming the same base weapon damage to reach this CBD%, at 100% Critical bonus damage total, you'll gain 1% average damage for every 1% increase in critical chance. If you have higher than 1%

Yes if you think about "average" damage as base damage, not relative ones.
1% crit chance for CBD 100% (so 1) is 1%. So for every 1% you get 1% base damage more. In relative sense it's always less.
In the same way works +% attack.
You can look at +1% attack as every 1% of attack gives me 1% more of base damage. Difference between those two and reason why for ppl with high crit chance CBD gives "more" is in relative damage increase, to other words for them
(1 + attack_bonus + damage_multiplier + type_bonus) is just higher than (1 + critical_damage_bonus + flanking_bonus)

[Angelus_de_Mortiel]

The relative scale for benefit still quickly diminishes, just the base benefit starts much higher, with more CBD.

I think you forgot the part where you claimed there is a "geometrically diminishing relative benefit" for Critical Chance .

Critical Damage Bonus and whether or not the brain can notice granular change is irrelevant to your statement. The fact is that Critical Chance itself is a linear scale and does not diminish. Period.

However, you are correct of the inability for the brain to "notice" granular change. As I already said, however, that is qualitative data and shows a completely different result than relative increase in Critical Chance as Cunning increases. You are, however, presenting this data as quantitative, which adds more confusion.

I would also note that adding Critical Damage Bonus in as a factor dilutes and confuses the topic. When establishing a scale on whether or not Cunning has diminishing returns, Critical Damage Bonus should always be set to it's base multiplier to avoid skewing results, as Cunning has no direct effect on Critical Damage Bonus. Again, you are over-complicating the subject.

[Drasca]

This idea is maybe more realistic but I would say it can't calculated in easy manner at all.

 

Long time for us, not so for machines (in absolute terms, I have no proof of machines perceiving time in a relative state). This is what calculating machines (computers) and programming is for. I never declared you had to hand calculate everything, or by yourself.

 

 

In relative sense it's always less.

 

This is always true, but actually calculating how much less is more important.

 

 

Your idea of time to kill and "big" numbers isn't the way to go. Because I could easily prove that even 50% damage increase does not change "time to kill" at all. 100 hp enemy. You hit for 60 damage. You need 2 hits, hits to kill = 2

 

I used real game examples, as there are enemies with actually 1.2k health, and players that do 200 damage. There are none with 100 hp. Your proof only valid for your one hypothetical case that does not exist. Both time to kill and hits to kill recognize the idea of overkill whereas average dps models do not.

 

To be more realistic, we can scale these hypothetical numbers up by a factor of 10, changing to 600 damage and 1000 hp enemy. The 50% damage alone is obviously not enough, but there's more than one damage bonus available in the game, and an additional 10% damage multiplier would be, note flanking alone is 25%, though added to critical bonus damage.

 

Perilous enemy values are pretty limited.  They're staggered at around 1k health, 2.3k, 3.4k, 3.8k, 4.5k, 7k, etc. I'm sure someone can provide a full list if they so chose to investigate. It is also not hard to calculate actual damage, how many hits are actually useful, and how much additional damage is actually useful.

 

The time to kill model is certainly valid when you use actual numbers. Definitely more valid than the average dps model when enemy health is low.

 

 

You can look at +1% attack as every 1% of attack gives me 1% more of base damage.

 

Except it isn't, when you use the context of ability damage with bonus damages that don't multiply with attack power.

 

 

(1 + attack_bonus + damage_multiplier + type_bonus) is just higher than (1 + critical_damage_bonus + flanking_bonus)

 

Just the opposite, the attack bonus often gives less benefit than critical damage bonus if CDB >= 100%, since the full formula is

 

final_damage = (base_damage * rand(0.95 to 1.05) - armor * (1 - armor_penetration))
* (ability_multiplier)
* (1 + critical_damage_bonus + flanking_bonus)
* (1 + attack_bonus + damage_multiplier + type_bonus)
* (1 - magic_resistance)

 

The critical damage bonus multiplies against everything, whereas the attack_bonus {attack power} merely adds to damage_multiplier {damage bonus on abilities}

 

This formula is imperfect too, and was built around single player, but we have to work with it until a better model is made.

[yarpenthemad21]

Drasca
It's not about calculating data by machines, it's about building a model.

For time to kill model I need:
1) Exact values of enemies
2) exacts values of damage done, abilities, attack order, rotation, everything pretty much connected with attacking.

So this model works for specific enemy, for specific build and specific gear. I've calculated this type of models in many games, mostly for burst type pvp builds, but in general I can easily say what is generic result of this.
It makes crit chance slightly more "worth" it. Just it. In DAIMP, calculating this type of model is just waste of time.
Every class out there has different way to deal with "time to kill" scenario and it's all about skills. Avvar can use LW as damage source to kill almost dead targets (so it makes his time to kill more stable), on the other hand archer with long shot rely fully on crits to kill something in one shot (so lack of crit in that instance hurts his time too kill a lot more).
If you want to analyze every type of that combination of skills, classes and enemies you can. For me it's not worth it, average damage works good enough. This time to kill model also gives result as probabilistic, you can't ignore it, plus what is obvious results can vary depending of enemy etc.
And average is just average. When someone ask basic question "should I take crit ring or crit damage ring" it will give fast answer for it, not set of probabilistic answers for each skill/enemy like time to kill model.

final_damage = (base_damage * rand(0.95 to 1.05) - armor * (1 - armor_penetration))
* (ability_multiplier)
* (1 + critical_damage_bonus + flanking_bonus)
* (1 + attack_bonus + damage_multiplier + type_bonus)
* (1 - magic_resistance)

As I said many times, we have 2 basic multipliers in damage formula.

attack one - (1 + attack_bonus + damage_multiplier + type_bonus)
crit one - (1 + critical_damage_bonus + flanking_bonus)

In general attack one is higher, we have more attack bonus in general than crit damage bonus + we have those damage multiplier bonuses from several skills.
Still how much crit damage is worth depends ONLY on crit chance as I showed in previous posts. When you take into into high crit chance scenario "crit multiplier" instead of being from time to time, starts to be always. This makes 1% in crit damage worth more than 1% of attack bonus in most cases.
But in everything else, crit damage works only on crits, and attack works on everything. If you introduce time to kill model,
crit damage can give you nothing in "bad luck" kill streak and attack will give you always something. This depends on enemy health can lead to either "no" change or change.