The Combat mechanics revisited - Page 8

• frmorrison · 09/14/06, 6:40 PM

BWL drakes require 315 skinning. I am not sure about Teremus, but he should require 315 skinning, since he is a level ?? Boss (which is 63 everywhere else, why would it be different for him?)

No counting the Uldaman bosses, they are strange.

Celandro · 09/14/06, 7:02 PM

Originally Posted by Kenco

result   count    percent
-------------------------
dodge      27     8.2%
parry      41    12.5%
miss       29     8.8%
block       9     2.7%
crit        7     2.1%
hit        74    22.5%
glance    142    43.0%
-------------------------
TOTAL     329

The dodge, miss and expecially parry numbers seem higher than I would expect, while the crit and block numbers are lower than expected. The net hit + crit cap of 24.6 is substancially lower than the 37.6 I would have expected for a 1h weapon. Dual wield with its 18% higher miss rate would have a hit+crit cap of only 5.6% which is substancially below the 18.6% of the time Id expect.

Something funny is going on here.

Emeraude · 09/14/06, 9:21 PM

http://ctprofiles.net/3702467

Breaking the 85% Avoidance barrier O_O

This is about 92%

sp00n · 09/15/06, 3:58 AM

Originally Posted by Celandro

Originally Posted by Kenco

result   count    percent
-------------------------
dodge      27     8.2%
parry      41    12.5%
miss       29     8.8%
block       9     2.7%
crit        7     2.1%
hit        74    22.5%
glance    142    43.0%
-------------------------
TOTAL     329

The dodge, miss and expecially parry numbers seem higher than I would expect, while the crit and block numbers are lower than expected. The net hit + crit cap of 24.6 is substancially lower than the 37.6 I would have expected for a 1h weapon. Dual wield with its 18% higher miss rate would have a hit+crit cap of only 5.6% which is substancially below the 18.6% of the time Id expect.

Something funny is going on here.

Ain't 329 swings a bit too less for definitve conclusions?

Celandro · 09/15/06, 3:33 PM

No, 329 is plenty to get fairly accurate numbers. I wouldnt say dodge, miss, block, glance or crit are outside the error ranges but combined, they put the hit number far out of the expected error ranges

Using the confidence interval calculator here: http://www.geocities.com/profemery/samplesize.htm put in 329 for sample size and 24.6 for the percentage. You get 4.65 for 95% and 6.13 for 99%. This means you can be 95% certain that the crit cap against this particular mob is between 19.95 and 29.25, and 99% certain that the crit cap against this particular mob is between 18.47 and 30.73. I dont remember enough statistics to tell you how likely the 37.6% crit cap is, but its far under 1% probably closer to 0.1% or less.

Anyhow my point is that Kenco's parse conflicts greatly with how I've always believed the dodge/parry/miss/block/crit/hit/glance table to work.
I would have expected 5.6% dodge, parry, miss, block, 4.4% crit, 40% glance and 33.2% hit.

Emeraude · 09/16/06, 2:02 AM

I just noticed this was the same Kenco I ran into on the Warrior boards. Heeeeeey =D

Kenco · 09/18/06, 2:59 AM

I just made the fattest data set ever! I have a level 25 warrior vs a level 28 elite mob. Warrior has 125/125 sword skill, using sword + shield. His stated crit rate is 6.44% (he has 2 points in cruelty and 40 agility). All attacks are made in front of the mob. The mob is Death's Head Acolyte in Razorfen Kraul (he casts renew at 50% so we can get plenty of attacks). He is classed as a Paladin.
We have 1496 attacks in total. The Min and Max columns refer to the 95% confidence interval from the page Celandro linked.

|-------------------------------------|
|Effect   Count Fraction   Min     Max|
|-------------------------------------|
|dodge  |   87    5.8%    5.2%    6.4%|
|parry  |  192   12.8%   12.0%   13.7%|
|miss   |  114    7.6%    6.9%    8.3%|
|block  |   58    3.9%    3.4%    4.4%|
|crit   |   25    1.7%    1.3%    2.0%|
|hit    |  410   27.4%   26.2%   28.5%|
|glance |  610   40.8%   39.5%   42.0%|
|-------------------------------------|
|total  | 1496 100.0%                 |
|-------------------------------------|

A few points of interest. Firstly the parry rate is nowhere near 5%, and this is replicated by my other data above. I can only suggest that Warrior or Paladin type mobs get an additional 5% parry or something similar.
Secondly, the crit rate is way down. The warrior is quoted at 6.4%, but the actual rate is probably below 2.0%. There seems to be at least -1% per level difference happening.
The block value could be 0% +1% per level + defense bonus.
Lot's of weird stuff though.

Kalman · 09/18/06, 6:13 AM

As odd as it is, some mobs use shield block. It's a thought, at any rate.

Apate · 09/18/06, 7:55 AM

Originally Posted by Kalman

As odd as it is, some mobs use shield block. It's a thought, at any rate.

I've even noticed it on mobs that don't visibly have a shield (can't recall specifically which). As a side note, I've also been surprised by the number of mobs with 1x or 2x thrash attacks, as noted by the most recent version of spellalert that I had (before it started spitting out lots of errors).

Emeraude · 09/18/06, 8:48 AM

Originally Posted by Kenco

I just made the fattest data set ever! I have a level 25 warrior vs a level 28 elite mob. Warrior has 125/125 sword skill, using sword + shield. His stated crit rate is 6.44% (he has 2 points in cruelty and 40 agility). All attacks are made in front of the mob. The mob is Death's Head Acolyte in Razorfen Kraul (he casts renew at 50% so we can get plenty of attacks). He is classed as a Paladin.
We have 1496 attacks in total. The Min and Max columns refer to the 95% confidence interval from the page Celandro linked.

|-------------------------------------|
|Effect   Count Fraction   Min     Max|
|-------------------------------------|
|dodge  |   87    5.8%    5.2%    6.4%|
|parry  |  192   12.8%   12.0%   13.7%|
|miss   |  114    7.6%    6.9%    8.3%|
|block  |   58    3.9%    3.4%    4.4%|
|crit   |   25    1.7%    1.3%    2.0%|
|hit    |  410   27.4%   26.2%   28.5%|
|glance |  610   40.8%   39.5%   42.0%|
|-------------------------------------|
|total  | 1496 100.0%                 |
|-------------------------------------|

A few points of interest. Firstly the parry rate is nowhere near 5%, and this is replicated by my other data above. I can only suggest that Warrior or Paladin type mobs get an additional 5% parry or something similar.
Secondly, the crit rate is way down. The warrior is quoted at 6.4%, but the actual rate is probably below 2.0%. There seems to be at least -1% per level difference happening.
The block value could be 0% +1% per level + defense bonus.
Lot's of weird stuff though.

I found something similar if you take a lever 25 toon and take it up against a level 35 character, the rates at which the mob parries/dodges can become quite insane.

Andrise · 09/18/06, 10:13 AM

Originally Posted by Kenco

A few points of interest. Firstly the parry rate is nowhere near 5%, and this is replicated by my other data above. I can only suggest that Warrior or Paladin type mobs get an additional 5% parry or something similar.

This definitely seems to be true. The Four Horsemen, for example, seem to parry at ~10% +/- 1. They happen to be the first mobs I specifically noticed it against, but it's certainly the case. (5/5 deflection? lol)