Miss cap for DK or diminishing returns on def rating

[Warku]

Hi,

I recently found out that the author of RatingBuster could not find out the miss cap for other classes than warriors:
"-- I've done extensive tests that show the miss cap is 16% for warriors.
-- Because the only tank I have with 150 pieces of epic gear required for the tests is a warrior,
-- Until someone that has the
will and gear to preform the tests for other classes, I'm going to assume the cap is the same(which most likely isn't)" - https://github.com/raethkcj/RatingBu...rath_Logic.lua

So I thought that I could do that for death knights, but what I found in classic and TBC raids (ICC data coming soon) was that the miss chance from defense rating does not seem to be diminished.

Can anybody verify?

[Artashir]

Hi,

I recently found out that the author of RatingBuster could not find out the miss cap for other classes than warriors:
"-- I've done extensive tests that show the miss cap is 16% for warriors.
-- Because the only tank I have with 150 pieces of epic gear required for the tests is a warrior,
-- Until someone that has the
will and gear to preform the tests for other classes, I'm going to assume the cap is the same(which most likely isn't)" - https://github.com/raethkcj/RatingBu...rath_Logic.lua

So I thought that I could do that for death knights, but what I found in classic and TBC raids (ICC data coming soon) was that the miss chance from defense rating does not seem to be diminished.

Can anybody verify?

Hello, I dont think i understand you well, but if im mistaken "excuse moi" , so isnt this think same like Runfogeing enchant Stonskin Gargoyle think>? The core is its give you pure defense, NOT rating defence , so this mean like i say you can on theory evade hits infinity coz its not under Diminishing Returns like rating was. For example you get 3% melee avoidance and 1% spell avoidance form that enchant and if you are godlike on every roll on dice you can miss forever coz you always have that % Up separate form dinimishing return. One more point this think scale with levels coz when you level up your scale on defense is rise up to 400 on level 80 so every extra youve got you do it coz mobs and bosses are high level than you so you actually play against level 83 . Sorry if im wrong i dont tank , its just my try to help from bottom of my heart .

[WePlayWithMath]

Yes, the runes provide additional defense, but my test was not stretched over different levels. I did measure at lvl 80 with 0-121 defense from defense rating against mobs at lvl 60-63, 70-72, and 83/Boss lvl. If Whitetooths formula is correct, then we would only need to determine one more unknown: miss cap. A test series at 121 additional def against lvl 60-83 mobs would show a linear fall and if there is diminishing return on additional defense, then the data would be below 0.04*(400+26+121 - target's lvl*5), because there should be some f(x) with f(121)<121 instead of 121, but it is not. It is not Whitetooth's methodology, but should be valid anyway.

I cannot upload my plot, so I have to post my data for 121 defense from rating:
Target Weaponskill Miss Miss error
300 0.202205882352941 0.024353302069167
305 0.182473655019412 0.009096055481096
310 0.159420289855072 0.017991239215146
315 0.180064308681672 0.02178830871938
350 0.166145708102909 0.005771586216055
355 0.168227121715494 0.006500859716739
360 0.153434433541481 0.010764377276426
415 0.14218009478673 0.010752048870469 (Bosses from AQ20 and Kara)

Miss are the observed number of misses divided by total number of attacks and Miss error is the statistical error on that. I have filtered out dual wielders, range attackers and attacks not against me from the data by hand.

[WePlayWithMath]

P.S.: I forgot to mention constants in the miss formula: 5% base miss, 3% from talents.

[WePlayWithMath]

Including the ICC bosses I get now a miss chance of (14.6 +/- 0.9)% against level 83 bosses and mobs (AQ20, Kara, ICC), while 13.3% are expected before(!) diminishing returns for 121 def from rating (neither insect swarm nor scorpid sting were applied). That means that what ever the miss cap is, it is well outside the margin of error.
Usually one data point is not sufficient to draw conclusions, but before continuing the measuring, I would like to know, maybe even from a developer, if DR on def rating is actually implemented.

[WePlayWithMath]

For those interested there is a post about the DR on def rating for Wrath Classic: https://github.com/magey/wotlk-warrior/issues/17
Don't know to what degree this is applicable to Warmane, but it seems that even Blizzard did not implement it right with classic.
Same discussion is also on the Discord channel Blight Club

[WePlayWithMath]

I found my mistake: I thought the Nightelven racial gives 2% dodge, not miss chance. Then I also arrive at the 16% miss cap within the margin of error.