Crit specific gear will need serious bonuses from talents and other sources to match pure damage bonuses.
It's really confusing if you ask me. High end items gives only +90 Precision (usually) wich is only +5% critical rate, fairly weak compered to +90% damage from +90 Attack or +90% Class Damage. The whole point of stats (Attack, Precision, Vitality, Defence) is that one point in any of those should have the same "power level". So, if you find a +20 attack ring it should give the same damage bonus as a +20 precision ring (and both items would have the same level requeriment).
I wish people would stop equating 1 Attack to +1% total damage output.
Let's say you start with 600 Attack. Your 100 base damage spell will deal 700 damage. Now you gain 1 Attack. Your spell deals 701 damage. 1/700 = +0.14% damage increase, not 1%.
Precision still increases total damage less than Attack does, but it is not by a ridiculously large ("15:1") margin.
I'm aware of your numbers. Anyway, despite not knowing the formula (you guessed: it's a server side calculation) we can however see a trend. This is for the Wizard.
If we assume that you get 4 HP per Vit (based on the tooltip text):
Lv 1: 76 HP = 9*4 + 40
Lv 2: 84 HP = 10*4 + 44
Lv 3: 96 HP = 12*4 + 48
Lv 4: 104 HP = 13*4 + 52
Lv 5: 116 HP = 15*4 + 56
Lv 6: 124 HP = 16*4 + 60
Lv 7: 136 HP = 18*4 + 64
etc.
The newer posts in the Incgamers thread (using an emulator to datamine lv 14-60 numbers) seem to indicate that the formula for 1-30 is different from the formula for 31-60. This shows the problem with extrapolating any 1-13 numbers to the level 60 game.
To be balanced, %Crit stacking CANNOT give as much damage as +%Damage stacking. Stacking +%Damage just gives you damage, nothing else. Crits tend to generate extra Fury/Arcane Power and can sometimes proc entirely different effects (aoe explosions etc). So of course the raw damage increase from Precision will be lower than Attack.
The thing about Crit in D3 is that it's not just +50% damage, once Runes come into play your Crits can also proc lots of special effects.
While 1000 precision may be significantly less upfront damage than 1000 attack, it may still be worthwhile if you build a character with a ton of proc-on-crit effects.
You can set L = 1 to 60. It sets the players level. Note that the result is for enemies of the same level only.
Those formulae appear to fit all known data points. Excellent!
I am surprised by how simple the level-scaling is (pure linear), and also by the rather high minimum threshold for Defense. I presume that most classes will be close to the threshold just due to levelup points, so gear will basically add to that.
The latter formalism (X instead of Y) is slightly easier to work with because it directly includes effective HP (= 100% + Armor*X), although you could always use (100% + Armor/Y).
Another thing we are unsure of is if there are any hard caps or diminishing returns thresholds. I suspect that kind of info will only become available once people are at level cap though >_>.
Good point. Although, if there are good uniques equippable at a low level (like a diablo 3 version of sigon's set) it may be possible to test out Armor cap with lowbies. (since it takes less Armor to get the same % damage reduction at low levels)
Figuring out Defense, Armor and Resistance scaling should be very easy if the mechanics are similar to WoW.
The formula from WoW is:
%DR = 100% - (100% / (x*Armor + 100%)), where x is a different number for each character level.
On the other hand, it appears that Precision is directly proportional to Crit, as in:
%Crit = y*Precision, where y is a different number for each character level
Since both of these formulae are directly proportional (ie, linear no threshold), you technically only need one datapoint to figure out the scaling value - simply divide one value by the other. However, to account for rounding error you will want to average out at least 3 datapoints.
If %DR is the displayed damage reduction (from Defense, Armor, or Resistance), then let %ieHP equal incremental effective HP (ie, the percent gain in effective HP over zero damage reduction):
%ieHP = (100% * %DR) / (100% - %DR)
This can be shown to be directly proportional to Armor as below:
However, if there is a threshold for Defense (ie, below a certain amount of Defense you get zero damage reduction) the formula will no longer be direct proportional. This appears likely as I've seen several screenshots of players with positive amounts of Defense but 0% damage reduction.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
I wish people would stop equating 1 Attack to +1% total damage output.
Let's say you start with 600 Attack. Your 100 base damage spell will deal 700 damage. Now you gain 1 Attack. Your spell deals 701 damage. 1/700 = +0.14% damage increase, not 1%.
Precision still increases total damage less than Attack does, but it is not by a ridiculously large ("15:1") margin.
If we assume that you get 4 HP per Vit (based on the tooltip text):
Lv 1: 76 HP = 9*4 + 40
Lv 2: 84 HP = 10*4 + 44
Lv 3: 96 HP = 12*4 + 48
Lv 4: 104 HP = 13*4 + 52
Lv 5: 116 HP = 15*4 + 56
Lv 6: 124 HP = 16*4 + 60
Lv 7: 136 HP = 18*4 + 64
etc.
To be balanced, %Crit stacking CANNOT give as much damage as +%Damage stacking. Stacking +%Damage just gives you damage, nothing else. Crits tend to generate extra Fury/Arcane Power and can sometimes proc entirely different effects (aoe explosions etc). So of course the raw damage increase from Precision will be lower than Attack.
While 1000 precision may be significantly less upfront damage than 1000 attack, it may still be worthwhile if you build a character with a ton of proc-on-crit effects.
Those formulae appear to fit all known data points. Excellent!
I am surprised by how simple the level-scaling is (pure linear), and also by the rather high minimum threshold for Defense. I presume that most classes will be close to the threshold just due to levelup points, so gear will basically add to that.
The two formulae are mathematically equivalent:
QED.
The latter formalism (X instead of Y) is slightly easier to work with because it directly includes effective HP (= 100% + Armor*X), although you could always use (100% + Armor/Y).
Good point. Although, if there are good uniques equippable at a low level (like a diablo 3 version of sigon's set) it may be possible to test out Armor cap with lowbies. (since it takes less Armor to get the same % damage reduction at low levels)
The formula from WoW is:
On the other hand, it appears that Precision is directly proportional to Crit, as in:
Since both of these formulae are directly proportional (ie, linear no threshold), you technically only need one datapoint to figure out the scaling value - simply divide one value by the other. However, to account for rounding error you will want to average out at least 3 datapoints.
If %DR is the displayed damage reduction (from Defense, Armor, or Resistance), then let %ieHP equal incremental effective HP (ie, the percent gain in effective HP over zero damage reduction):
This can be shown to be directly proportional to Armor as below:
However, if there is a threshold for Defense (ie, below a certain amount of Defense you get zero damage reduction) the formula will no longer be direct proportional. This appears likely as I've seen several screenshots of players with positive amounts of Defense but 0% damage reduction.