UPDATE: I tested the effect of weapon level on evade rate. It's at the bottom.

Here are the results of my research to see what effect (if any) weapon spec and damage stats have on a target's evade rate. I have organized my research in to 2 major tests. Test 1 measures the effect of weapon spec only. Test 2 measures the effect of increased damage stat combined with different weapon specs.

Each test is further split in to 2 targets: A and B.

Target A is a level 50 Ranger with 181 Dexterity and 113 Quickness. Her unmodified chance to evade an attack is 19.5%.

Target B is a level 50 Nightshade with 105 Dexterity and 83 Quickness. His unmodified chance to evade an attack is 37.2%.

The attacker is a level 50 Ranger, Blades spec, equipped with a Practice Bronze Sword (level 1 weapon).

You can download the combat logs here: http://www.mediafire.com/?hzygtj2dmdo


Here are the results:


---------------------------------------------------------------


TEST 1A


Test 1A1 - Blades 1, Strength 65, Weapon Skill 388
-Attacks: 1468
-Evades: 400
27.3% evade rate

Test 1A2 - Blades 25, Strength 65, Weapon Skill 481
-Attacks: 1630
-Evades: 447
27.5% evade rate

Test 1A3 - Blades 38, Strength 65, Weapon Skill 531
-Attacks: 1538
-Evades: 388
25.3% evade chance

Test 1A4 - Blades 50, Strength 65, Weapon Skill 577
-Attacks: 1521
-Evades: 312
20.5% evade rate

Test 1A5 - Blades 50+20, Strength 65, Weapon Skill 688
-Attacks: 1586
-Evades: 373
23.6% evade rate


TEST 1B


Test 1B1 - Blades 1, Strength 65, Weapon Skill 388
-Attacks: 1709
-Evades: 867
50.8% evade rate

Test 1B2 - Blades 25, Strength 65, Weapon Skill 481
-Attacks: 1694
-Evades: 826
48.8% evade rate

Test 1B3 - Blades 38, Strength 65, Weapon Skill 531
-Attacks: 1757
-Evades: 880
50.1% evade rate

Test 1B4 - Blades 50, Strength 65, Weapon Skill 577
-Attacks: 1522
-Evades: 713
46.9% evade rate

Test 1B5 - Blades 50+20, Strength 65, Weapon Skill 688
-Attacks: 1635
-Evades: 778
47.6% evade rate


---------------------------------------------------------------


Test 2A


Test 2A1 - Blades 1, Strength 122, Weapon Skill 493
-Attacks: 1543
-Evades: 375
24.3% evade rate

Test 2A2 - Blades 1, Strength 267, Weapon Skill 755
-Attacks: 1498
-Evades: 260
17.5% evade rate

Test 2A3 - Blades 50, Strength 122, Weapon Skill 733
-Attacks: 2100 (I fell asleep lulz)
-Evades: 432
20.6% evade rate

Test 2A4 - Blades 50, Strength 267, Weapon Skill 1122
-Attacks: 1544
-Evades: 206
13.4% evade rate


Test 2B


Test 2B1 - Blades 1, Strength 122, Weapon Skill 493
-Attacks: 1572
-Evades: 744
47.3% evade rate

Test 2B2 - Blades 1, Strength 267, Weapon Skill 755
-Attacks: 1601
-Evades: 571
35.7% evade rates

Test 2B3 - Blades 50, Strength 122, Weapon Skill 733
-Attacks: 1552
-Evades: 592
38.1% evade rate

Test 2B4 - Blades 50, Strength 267, Weapon Skill 1122
-Attacks: 1526
-Evades: 438
28.7% evade rate


---------------------------------------------------------------


The Weapon Level Test


Target B - Blades 50, Strength 122, Weapon Skill 733, level 50 Diamond Seal sword
-Attacks: 1482
-Evades: 537
36.3% evade rate

You can download the combat log here: http://www.mediafire.com/?kawlmm5qfmh


---------------------------------------------------------------


Conclusion


We still need to figure out the correct formulas, but here are my initial thoughts:

--Weapon spec up to 50 composite does effect target evade rates. This is clearly demonstrated in Test 2.

--Weapon spec over 50 composite has no additional benefit for defense penetration, or the benefit is tiny.

--Having weapon spec below 50 or strength below 122 actually increases the target's evade rate. The Nightshade with a 37.2% base chance to evade consistently reached the evade cap (50%) against a low weapon spec, low strength attacker!

--Damage stat obviously has a huge impact on target evade rates.

--Weapon level has no effect on evade rate.


Comments?

Good Job with your testing. I suspect this also applys to other defenses like parry and blocking also. One issue i see though is when comparing only weapon spec, your WS is higher or lower and is hard to make a comparison. One way to see the effect of only weapon spec is to lower dmg stat when you raise your weapon spec to keep the same WS. I suspect also, that its the WS that is lowering the defenses, and if that is the case, just need to know the formula to turn WS value into reduction of defenses. Which your spec and dmg stat has an effect, since more weapon spec = less dmg stat for more WS.

Eclipsed, I thought the same thing at first, but compare these two tests:

Test 1B5 - Blades 50+20, Strength 65, Weapon Skill 688
-Attacks: 1635
-Evades: 778
47.6% evade rate

Test 2B1 - Blades 1, Strength 122, Weapon Skill 493
-Attacks: 1572
-Evades: 744
47.3% evade rate

The strength test wins, even with almost 200 points less in Weapon Skill. WS is such a useless stat. I don't know why Mythic even bothered with it.

ya i see that issue. But another issue in that comparison is that the 50+20, is giving it WS for the RR and +11 skill , which you found that it may not give much bonus past composite 50, so kinda throws off the WS comparison for that. Which possibly the displayed WS value has no effect on the actual bonus, but some formula between spec + skill + dmg stat and what ever formulas , is what makes up the bonus. Which does reflect the WS a little, since all those things give WS, its just the defense penetration bonus is figured diffrently or has its caps on skill and so on.

So what you need to fiqure out is, what bonus does each lvl of spec give, how much dmg stat gives what bonus, and how do they relate combined. Like in the situation of WS, the more spec you have, the more WS you gain per stat. Maybe defense penetration has some simularity, like how much penetration do you get per stat with low blade spec, and how much per stat with higher blade spec.

What ever the case, spec and dmg stat does have an effect vs your enemys defense, wether its a flat -%, or a -% of there current defense %. Like lets say they have a 50% chance to evade and your stats and spec is giving you a 20% bonus , it would ether be 50 - 20 = 30% evade chance left, or 50 - 20% = 40% evade chance left. I would think it would be the second way, this way someones bonuses could never totaly null there defenses, unless somehow they get enough to reduce it by 100% lol.

Eclipsed wrote:ya i see that issue. But another issue in that comparison is that the 50+20, is giving it WS for the RR and +11 skill , which you found that it may not give much bonus past composite 50, so kinda throws off the WS comparison for that. Which possibly the displayed WS value has no effect on the actual bonus

That's a good point, so let's compare two more tests:

Test 1A3 - Blades 38, Strength 65, Weapon Skill 531
-Attacks: 1538
-Evades: 388
25.3% evade chance

Test 2A1 - Blades 1, Strength 122, Weapon Skill 493
-Attacks: 1543
-Evades: 375
24.3% evade rate

Strength wins again, by an even larger margin than before, with 38 points less in WS.

WS may give the player a general idea of how good her character is at defense penetration, but I don't believe it can be used in calculations.

So what you need to fiqure out is, what bonus does each lvl of spec give, how much dmg stat gives what bonus, and how do they relate combined. Like in the situation of WS, the more spec you have, the more WS you gain per stat. Maybe defense penetration has some simularity, like how much penetration do you get per stat with low blade spec, and how much per stat with higher blade spec.

Yeah, I was hoping the community could help out with this. My brain is fried.

What ever the case, spec and dmg stat does have an effect vs your enemys defense, wether its a flat -%, or a -% of there current defense %.

When we compare the Ranger evade rates to those of the Nightshade, it is clear that defense penetration is based on a -% of the current total defense %. It is not a flat -% increase/decrease.

Like you pointed out, if it were a flat -% then it would be possible to reduce a target's defense rate to 0%. Which we know is not the case.

Yes i have a few comments:

- Evade is a defensive skill so acts same way like the block test i ve done. If u r using a low lvl weapon u havent tested much as the evade increases alot in my opinion and you will see v high numbers that dont help so much, better equip the target with artefacts armours and use on attacker artefact weapon ,both chars lvl 50 rr5+ .
- Weapon spec over 50 and 50 its almost the same dmg stats like str or dext for attacker will lower evade rate alot more then + to skill in weapon even if it adds same amount to WS.
- You used ranger as attacker, it is a class from the bottom of the dmg table, i m prety sure a tank like BM or Hero would have given lower evade % , you can test this also .
- In my opinion better to use a friar as defender with no weapon equiped because you can heal from time to time and he ll evade only that way.
- testting an increasing spec is good to see a pattern either its skill in weapon or a dmg stat yet keep in mind that you got a marginal error of +- 2% if not more and its gona be dificult to make conclusions if diference in % is low.

I doubt that weapon level effects evade rate. As you can see, a full buff Ranger with 50 weapon spec already reduces the target's evade rate by a huge amount (about 1/3rd). A Warrior with the same stats would reduce the evade rate by about 1/2. Adding any more penalties to defense rates would be extreme.

Glacius, I don't have access to good equipment. Can you equip a level 1 weapon on your test characters and see if that increases defense rates?

Just do a thousand swings with your test characters and compare the results to your blockrate test. If the blockrate is higher against a level 1 weapon then we will have our answer.

It's easier to test down, than it is to test up.

Seyha wrote:Eclipsed, I thought the same thing at first, but compare these two tests:

Test 1B5 - Blades 50+20, Strength 65, Weapon Skill 688
-Attacks: 1635
-Evades: 778
47.6% evade rate

Test 2B1 - Blades 1, Strength 122, Weapon Skill 493
-Attacks: 1572
-Evades: 744
47.3% evade rate

The strength test wins, even with almost 200 points less in Weapon Skill. WS is such a useless stat. I don't know why Mythic even bothered with it.

Weapon skill isn't a stat, which is where most of the confusion about it comes into play. It's a calculation and it's (somewhat) useful if you know how it could mislead you.

Wyrd proved a long time ago that your test spec (Blades 50+20) is unnecessary as far as weapon skill is concerned with regard to evade rate. What you should aim for is (at least) 51 composite spec in a base weapon line and max out any style line (if they happen to be the same, max them out as best you can) for damage purposes. For defense penetration it should be at 50. In other words, speccing that extra 1 point (51 vs. 50) gets you a little more damage, a little less damage variance, but no additional defense penetration.

Composite spec = Spec pts. + pts. on armor + bonus pts. due to Realm Rank.

For example: At RR6 with capped spec for a weapon = 35 spec + 11 (capped armor) + 5 (RR bonus) = 51 composite spec.

Here's where the weapon skill can mislead you. After 50 composite spec you will not penetrate defenses any better though your WS will continue to rise. Your strength will give you a bonus to penetrating defenses, however (or 1/2 str + 1/2 dex if you are using piercing weapons). The weapon skill calculation might be higher because of all the spec points in a line but you'll only penetrate defenses better with higher weapon statistics (even if the overall weapon skill calculation is lower).

You've basically confirmed Wyrd's tests. His data are still online for the tests, even from way back in 2004 it seems.

Rat, I admit I'm a bit confused by Wyrd's testing. I don't think Wyrd's tests were nearly as comprehensive as mine. If you look at the tests he posted in that thread, the only time he tests a weapon spec below 50, he ends up at the evade cap (50%). This means that the test was not accurate because the real evade rate was hidden by the cap.

I was able to rectify this in Test 2:

Test 2B1 - Blades 1, Strength 122, Weapon Skill 493
-Attacks: 1572
-Evades: 744
47.3% evade rate

Test 2B3 - Blades 50, Strength 122, Weapon Skill 733
-Attacks: 1552
-Evades: 592
38.1% evade rate

That is a big difference between 1 weapon spec and 50 weapon spec. Maybe we can use Weapon Skill to calculate evade rates if we know to discard WS gained from over-speccing a weapon, but that seems very complicated. Different classes get different amounts of WS from weapon spec based on their damage table.

It just seems very inconsistent.

Seyha wrote:Different classes get different amounts of WS from weapon spec based on their damage table.

It just seems very inconsistent.

Yep, though most (when they're speccing) can jot down their WS at 50 composite spec and realize that their defense penetration is not going up beyond that. They can do it as they level up and finally get 50 composite in their weapon spec and they can find that number out at any point where they have to respec (since they'll be trying to hit 51 or higher).

Any gains in WS after that point should be ignored, and gains due to str (slash/crush) or str/dex (pierce) are still valid. It's a bit tricky. For most classes who have to worry about it (usually stealthers and soloers) there's a good chance that they're only going to composite spec to 51 in a weapon line anyway, so ALL gains in WS for them will be valid. They'll put the rest of their points into a different line entirely for maxing out style damage in that line (critical strikes, left axe, etc.). Infiltrators will still have to think about it since they're probably going to be putting all their points into one line in order to get dragonfang.

My understanding is that these tests do not hold for 2-handed wielders, however, and that defense penetration is supposed to be higher for 2-handers in general. I'm not sure of the source of that information since I played several Wyrd specced classes (so I remembered the source) and don't recall having ever played a 2-handed wielder. I wish I could point everyone directly to a source on that anyway, sorry. :-/

I will do a direct weapon skill comparison (50 Weapon Spec and 60 strength vs. 1 Weapon Spec and 120 Strength). These two setups should have about the same listed WS.

weaponskill is a client side # only

that has been stated numerous times in many grab bags.

the client itself performs that calculation based on the modified skill of the weapon equipped, and the "damage stats" which is str, str/dex, or dex depending on the weapon.

the weaponskill # you see on your paperdoll is calculated by the client. not the server.

it even goes up by the same amounts when you hit hard and soft caps that the server imposes.

the soft caps are 350 and hard caps are 400, other than qui which hard caps at 190 points of benifit.

so on your paper doll, soft cap is 410, and hard caps is 460, and qui hard caps at 250.

even when you go above those #s, wskill will still go up in a linear fashion even though it has no effect or diminished effect.


while i will admit that 2k attacks is a pretty good sample, the test results do not bear out the consistency.

oh, and if you are using a low level weapon such as a level 1 weapon, then your target CAN use level 50/51 armor, but the weapon and armor must share the same bonus % or else that will throw off your #s.

if you were using a level 1 weapon and level 50/51 armor with a full 35% bonus, then the tests are less than useable as they have flaws.

if you are going to use these sample sizes, you need to spread them out. mythics RNG seed is estimated to be gotten every 10-15 minutes, and some have claimed every 30 mins. what this means is that you need to break up your samples into no more than 200 "tests" per RNG seed. if you just did a continous test, that may be the cause of the inconsistencies we see in your results.

Seyha wrote:I doubt that weapon level effects evade rate. As you can see, a full buff Ranger with 50 weapon spec already reduces the target's evade rate by a huge amount (about 1/3rd). A Warrior with the same stats would reduce the evade rate by about 1/2. Adding any more penalties to defense rates would be extreme.

Glacius, I don't have access to good equipment. Can you equip a level 1 weapon on your test characters and see if that increases defense rates?

Just do a thousand swings with your test characters and compare the results to your blockrate test. If the blockrate is higher against a level 1 weapon then we will have our answer.

It's easier to test down, than it is to test up.

If you would have look at my block results you would have seen that attacker s weapon beeing lower lvl is blocked more , and with more then 10% actually , also for block having a lower lvl shield as defender will block less. ( It was tested before you opened your thread)

All of your results i think that are higher with a certain % because you have used a low lvl weapon VS a lvl 50 player , test under same stats with a lvl 51 weapon.
I have dressed so far more then 20 chars on pendragon full artefacts , and u come and tell me that you dont have acces to good gear..

Seyha wrote:I will do a direct weapon skill comparison (50 Weapon Spec and 60 strength vs. 1 Weapon Spec and 120 Strength). These two setups should have about the same listed WS.

You cant be so sure they will give you same WS, i advice you to choose them in such a manner that they will .
I have done this test for block and conclusion was that dmg stat like str lowered block more then spec in weapon at same WS.

@ Therat I havent tested for block to see blockrate for 50 weapon spec or with 50+23 weap spec on attacker ,but i will ..will be a WS increase and even so from 1- 50 weapon spec the blockrate wasnt less with much.

Murzin wrote:while i will admit that 2k attacks is a pretty good sample, the test results do not bear out the consistency.

oh, and if you are using a low level weapon such as a level 1 weapon, then your target CAN use level 50/51 armor, but the weapon and armor must share the same bonus % or else that will throw off your #s.

if you were using a level 1 weapon and level 50/51 armor with a full 35% bonus, then the tests are less than useable as they have flaws.

The Nightshade was completely naked. The Ranger had some crap Merchant gear on with 0% bonus. It made no difference. The evade rate % seems to increase and decrease at the same rate for both targets when you account for the evade cap and margin of error. The results between the tests are actually quite consistent with only one or two "anomalies," again due primarily to the evade cap hiding the real numbers in a few of the tests.

if you are going to use these sample sizes, you need to spread them out. mythics RNG seed is estimated to be gotten every 10-15 minutes, and some have claimed every 30 mins. what this means is that you need to break up your samples into no more than 200 "tests" per RNG seed. if you just did a continous test, that may be the cause of the inconsistencies we see in your results.

If I can't see what the RNG seed is or when it begins or how long it lasts then what is the point of spreading the tests out? I would have to do thousands of attacks spread out over several hours to make any difference and how do you know the RNG even effects evade rates? I've never seen anything to suggest that it does.

No offense but I think you're over-complicating this.

Glacius wrote:
If you would have look at my block results you would have seen that attacker s weapon beeing lower lvl is blocked more , and with more then 10% actually , also for block having a lower lvl shield as defender will block less. ( It was tested before you opened your thread)

All of your results i think that are higher with a certain % because you have used a low lvl weapon VS a lvl 50 player , test under same stats with a lvl 51 weapon.

I looked over your test. I did not see the part where you tested with a lower level weapon. Please copy/paste that part of the test.