Dexterity impact on flurry and procs.

[Cydrenlin]

At start I wanted to figure out by myself how important is dexterity related to all kind of "procs" (procs being seen in a large way: critics, flurry, "buff procs" -aten gloves-, and "classic" procs).

Everything I have ever read/heard was that "safehouse/monkly business/whoever says dex isn't very important related to proc rates".
My expectations were to see a very slight improvment on proc rates with a higher dex but I was clueless about flurry, "buff" procs (aten gloves) and critics.

I have maxed offensive AAs (ferocity 3, ambidextry, combat fury 3, flurry 3, fury of the ages 3, raging flurry 3).

So I ran the 3 following tests on a poor SH halfling guard:

1°) Bloodfrienzy+BoC (unbuffed, dex 185, "classic" lazy xp group)

2°) Bloodfrienzy+primal spear (avatar stayed up for all the parse, 275 dex)

3°) Bloodfrienzy+priceless spear (unbuffed, dex 175)

I parsed:
The total number of swings (don't care hit or miss and damage)
The total number of Enraging blow procs.
The total number of call of fire procs.
The total number of flurry.
The total number of critics/cripplings.

And I calculated proc rates (number of proc / number of swing*100).

Tests were done unbuffed. Point was just to see what was my best xp group setup if i was too lazy or to busy to swap weapons for avatar.
I don't calculate the proc ratio of EB in 1rst test as well as flurry ratio since I can't know how many attacks are related to main hands and offhand. The most interesting tests to compare are 2 and 3 imo.


TEST 1
(185 dex, bf/boc):

5395 swings (all slash)
108 EB procs (na)
61 call of fire procs (61/5395*100=1.13%)*see note
64 flurry (na)
164 critics/cripplings (164/5395*100=3.03%)

TEST 2
(275 dex, avatar up all the time, bf+primal spear):

5634 swings (3107 slash, 2527 pierce)
123 EB procs ( 123/3107*100 = 3.95% )
30 avatar procs ( 30/2527*100 = 1.187% )
62 call of fire ( 62/5634*100 = 1.1%) *see note
52 flurry ( 52/(3107-2*52)*100 = 1.73% ) **see note
277 critics/cripplings (277/5634*100 = 4.916%)

TEST 3
(175 dex, bf+priceless spear):

5258 swings (2934 slash, 2334 pierce)
75 EB procs (75/2934*100=2.55%)
51 call of fire procs (51/5258*100=0.969%) *see note
51 flurry ( 52/(2934-2*51)*100= 1.83%) **see note
165 critics/cripplings (165/5259*100=3.138%)


Notes:
*: I assumed call of fire can get a proc out on any swing, main hand or offhand. So I calculated procrate using total number of swings.

**: I assumed flurry was only triggering on main hand attacks. In fact, we don't know really.. maybe there is a roll at the start of the round (you can't "double flurry" in the same round). Anyway, since I can hardly take a melee round as a parsing basis, i decided that flurry could be counted as if it was a mainhand only proc. I also removed flurry attacks from the count of the main hand attacks in order to calculate this ratio (since you can't flurry on a flurry).


So...
- clearly the best xp group setup for aggro is to keep primal offhand (unless you got a nice shaman in group or unless your dex is nearly maxed unbuffed anyway).

- Flurry and call of fire (proc that is not on a weapon) does not seem to trigger more with a higher dex as I secretly hoped... so this is a rather bad news for me.

- Impact of 100 dex on the crit rate is really impressive..

- Proc rate of bf is greatly improved by +100 dex, enough to keep primal over boc if I don't want to switch all the time.


So.. I guess nothing new on this parsing for most of you, but I never seen anyone parsing call of fire or flurry and trying check if dex was meaning something for it (seeing how dex have a huge impact on critics, one could have assumed it does also impact flurry).

Maybe the parse are a bit small so I guess i'll do them again with around 20,000 attacks when I will have the time.

edit: spelling, grammar etc.... and sorry if there is still a lot of mistakes, english isn't my mother tongue :rolleyes:

[Sylviania]

interesting parse ....
now one small thing to add : shamen when they did test proc rates & dex, discovered proc rate depended on the weapon delay : in fact you had and average of 1 proc every 45 seconds, whetever weapon u used ( well , that was before the cleric's hammers.. ).
so to test dex efficiency , you have to check nbr_of_proc / duration_of_fight... could you give us those values please ?

[Cydrenlin]

Yes,
test 1
[Mon May 05 02:42:13 2003] Auto attack is on.
[Mon May 05 03:20:24 2003] Auto attack is off.
=> 38min 11sec

test 2
[Mon May 05 16:43:05 2003] Auto attack is on.
[Mon May 05 17:22:40 2003] Auto attack is off.
=> 39min 35 sec

test 3
[Mon May 05 18:00:19 2003] Auto attack is on.
[Mon May 05 18:36:57 2003] Auto attack is off.
=>36min 38sec

I don't think you should compare duration of fight but number of attacks. Number of swing doesn't take into account the haste (wich is good since it's not the purpose of this test).

edit: since you mentionned weapon delay, bf, boc, primal and priceless have same delay (20). Bf is known for having a higher proc rate than boc but the purpose of this parse wasn't to prove it :)

[Sylviania]

well , what showed shamen the proc rate a constant based on time and not swings is the fact having haste did not make proc more often than without haste ( good here again , as haste is not purpose of this parse :P ). proc average being considered (for a given weapon) constant over time and not swings, that's why I asked for duration of your parses

btw, CoFire is indeed working for both hands ( not even need a weapon ,fun to stun a monk with punch :D ).
And as a ranger , for me the result shown here for dex effect on CoF proc rate is just ... bah ... *throws away her pocket shaman*

[Varatho]

You know, someone should round up all the stat test threads and sticky them or something

[Cydrenlin]

"is the fact having haste did not make proc more often than without haste"

I don't believe that I proc at the same rate without 60% haste or with it.. you can count on me for parsing that tomorrow after I get some sleep...

That would puzzle me quite a bit... I always considered proc was checked at each attack (hence the ability to proc on a miss for exemple) and not over time...

[Isk]

Actually yea it's basically over time. If you're slowed you have a higher chance per swing to proc.

Varaho (or anyone else) if you'd like to collate the stat tests I'm sure one of us mods can hook ya up with a sticky.

[Sylviania]

that sure is puzzling :)
but as Iskandar explained, in fact at each swing a check is made, and the chances of the check are dependant on your current delay (more change /swing when slowed, less per swing when hasted). basicaly check is made versus some_weapon_constant*delay

this allow most weapons to have same chances to proc in a given amount of time, without having to bother about delay ( fast or slow ), and allow 2hand weapon to have good chances to proc anyway ....

just though of that ... the weapon with most damage ( or aggro, for ya warriors ) based on proc can be muuuuch more efficient versus mob that are slowing .... as slow won't effect procs rate , and so u will still do near as many damage and/or aggro as if u weren't slowed....
biggest problem : opposite is true too ... weapons based on proc will lose efficiency ( well , compared to non proccing weapons )when you are hasted...

[Cozlo]

this allow most weapons to have same chances to proc in a given amount of time
I was under the impression it had already been concluded that different weapons have different innate proc rates?

I can tell you for a fact, in any set of circumstances, my FB procs more than my WC, and the FG proc rate is just sick.

[Isk]

The key word is most. A few weapons have different proc rates. Also parse your proc rates, you might be surprised.

[Tinuvil]

...and how about different procrate in different Zones?

Could be easily done when you have a different random number generator for zones.
I dont think thats a myth.

Anyone ever thought about parsing that?

[Korak Sarathai]

Could be easily done when you have a different random number generator for zones.

I have been more than a little curious about that. I tend to disbelieve it, simply because it is unnecessary complexity. However, I swear I can't proc any of my weapons to save my life in most Ykesha zones. I helped my roomate get his charm, then I got the hell out.

Once again, I don't want to know the answer badly enough to parse it myself. Perhaps someone else does.

[Ladas Sindarin]

Call of Fire, like the other "imbue proc" buffs, only works on your mainhand attack, iirc. Put a Stien in your primary and you won't get any procs.

[Mourningloree]

Another interesting thing to see would be what's the effect of the % of hits that were flurries if you changed to a 2h weapon. Are 1.7% of swings w/a 2h flurry as well or does it go up as the delay goes up maintaining a constant flurry over time percent.

I just got flurry 1 last night and as suspected it sucks. Horribly. In 40% of an AA in an xp group it has only gone off once, but it did do 1474 damage that one round it went off. If even after PoP Flurry 3 it's only on 1.7% swings it would be a bit dissapointing...

[Rodney]

Over the past two years I have tested proc rate on every weapon I have had from WC,FB types to epics to GSoT and they all have the same rate of 1 proc every 30-35 sec with a 275 dex and it seems to be linear with dex. This data is the first I have seen to show a proc rate different then 1 proc every 30 sec. The way random number gererators work you can get very very different results of short periods (30 -60 mins but over time (> 2 hrs) it comes into 1 per 30 sec. I have never found a difference based on zones.

test #2 1 proc every 19 sec
test #3 1 proc every 30 sec

275/175 = 1.571
30/19 = 1.578
proc rate in this range looks to be very linear based on dex


Rodney
Overlord CT

[SlimyDaTroll]

So.. I guess nothing new on this parsing for most of you, but I never seen anyone parsing call of fire or flurry and trying check if dex was meaning something for it (seeing how dex have a huge impact on critics, one could have assumed it does also impact flurry).

I would submit that the large increase of crits is due to the +100ATK from Avatar rather than the extra Dexterity.

Slimy

[Dinian]

You really think the atk rating of a player increases your proc rate?

[SlimyDaTroll]

I actually said crit rate, not proc rate.

[Haass]

A crit is, in essence, nothing more than a proc that multiplies your damage.

[SlimyDaTroll]

Regardless, all I'm saying is that with Dex maxed at 305 unbuffed, I see a lot more crits with Ferine on me than without.

Obviously not empirical evidence, but it certainly seems to be the case.

Slimy

[Varatho]

Have you seen the same increase of crits with different +atk buffs?

[Cydrenlin]

Ok.. so TODO list:

- run a test with a 2hander and parse flurry rate
- get +100ish attack from buffs and run test 2 again (to see if dex or attack is modifying crit rate).

I'll try to do that tomorrow, 2am here atm :)

edit: and use a larger period of time (~2hours).

[issu]

one question here
If i didn't remember wrong , CoF proc seems to be treask as innate proc which only proc when you hit something.
Cloud you re-run your parse and treat CoF as proc only when hit ?
I am not sure that CoF will proc only on mainhand or not either.
Maybe give out the exactly hit number of all tests would be a good idea to do some extra result.

[Graal the Dorf]

A very exhaustive study was done on the effect of dex on weapon proc rates by several people on the Zeb server. Tests were done with 10k swing samples of about 15 different delay weapons, plus samples with some of the those same weapons hasted. The samples were also done at different levels of DEX. It was an amazingly comprehensive study. (Notice I didn't use the term "parse", it went far, far beyond that)

There were a few exceptions, but generally all weapons with a proc had the same proc rate equalised to delay for a given level of DEX.

If I remember correctly:
Proc_rate/delay=~0.002 at 255 DEX for most weapons

I think normal churn has eaten the post on the shaman's crucible, but if you are interested in the raw data, post there asking if Ruca still has it.


BTW, proc buffs are completely unaffected by DEX.

[Sylviania]

here may be an answer to the increased criticals when +atk :
+atk mean a greater hit/miss ratio, so more hits, so more chances to critical. ( as i guess u can't crit a miss :p )
and yes u should give numbers of actual hits too in the parse, to check.

[Isk]

Is that a theoretical? I was under the impression that increased raw ATK (not through weapon skill and offense skill) didn't affect melee accuracy.

[Sylviania]

well , i admit i dunno ... but i though it would. problem is i dunno where to find a formula explaining how is calculated atk ( it is say atk is made of 2 part , like AC is). and yes that wazs a theory, i am always too lazy to do parses myself :p

[Ladas Sindarin]

Again, Call of Fire only fires from the mainhand, and it will proc on a "miss". Its based on the swing, not a hit.

Increasing your attack from 1200 to 1800 has no effect on your accuracy that I have ever seen in all my parses, but it does shift your damage to a higher portion of the weapon range.

[Sylviania]

thx for clarification Ladas

[Tinguran]

bah we need sony to open source this sheep, we spend way to much time testing data, and reading others tests :P


ahh well we can all wish on something :)

[Haass]

+atk mean a greater hit/miss ratio, so more hits, so more chances to critical. ( as i guess u can't crit a miss

This is false. Raw ATK does not improve your chance to hit, it increases your average hit.

[Cydrenlin]

Ok after the posts here I ran another test with a 2hander to check if flurry was based on the number of attack rounds or, as for normal procs, based on time.
I was also curious about call of fire proc rate on a 2h.

Weapon used was Norge'tal (53/45), and i'm not including Norge'tal procs in the result since it's not the purpose of the test.

Test was still performed in SH, on the same mob, still unbuffed. Dex was 155.

Test 4:
Duration 1 hour 8 seconds:
2266 attacks (1301 hits, 965 miss)
67 call of fire procs
35 flurry
75 crit/crip


To compare to the previous datas:


Test 2 (275 dex, bf+primal)
Duration 39 min 35 sec
5634 swings (3107 slash, 2527 pierce)
62 call of fire
52 flurry
277 critics/cripplings


Test 3 (175 dex, bf+priceless)
Duration 36 min 38 sec
5258 swings (2934 slash, 2334 pierce)
51 call of fire procs
51 flurry
165 critics/cripplings



Flurry ratio doesn't seem to be based on time at all but on number of attacks from the main hand:

test 4:
35/(2266-35*2)*100= 1.59%

test 2:
52/(3107-2*52)*100 = 1.73%

test 3:
51/(2934-2*51)*100= 1.80% (made a mistake in 1rst post)

The more main hand attacks you get (and probably the more attacks "rounds" you get, meaning the lower delay your weapon is) the more flurry AA triggers.. that seems to be it.

It's not based on time and doesn't seems affected by dex..
I guess I would need to test during sereval hours to get a more exact ratio for flurry but around 1.7% seems rather correct.

Now.. about call of fire: It seems safe to say that it's based on time and not on number of attacks (still have to test that main hand / offhand stuff).
However, as for flurry, the exact ratio proc /sec will be hard to get exactly without a very very long parse.

I'm going to check now if crit rate is based on attack or on dex.. i will come back to post a result when I get one.

[Cydrenlin]

Originally posted by SlimyDaTroll
I would submit that the large increase of crits is due to the +100ATK from Avatar rather than the extra Dexterity.

Slimy

I decided to test that.. so I ran a 5th test
I used bf / priceless but got my dex up to 273 with focus / diaku from a shaman.
So my stats are nearly exactly the same that in bf+primal test but without the +100 attack.

My dex was 273 (instead of 275 in the other test).

And since I wanted to get sure, I ran it for nearly 1hour 30 min (1 hour 26 min 12 sec).

I got 528 critics/crippling for 6679 hits. So the ratio "number of critics" / "number of hits" is for the 3 tests:

bf+priceless (175 dex) 5.7%
bf+primal(275 dex) 8.7%
bf+priceless+dexbuffs (273 dex) 7.9%

I guess the 0.8% difference between primal/dexbuffs is due to the fact that primal test is only 36 min long and maybe I got slightly lucky on proc.
But without changing attack I still get a +2.2 on crit rate so imo it's only based on dex. Attack just makes you hit harder, so maybe you "notice" your crits more with avatar.

[Meety Da'Shield]

Nice Parsing but,.... deep inside i'll keep believeing asking for a dex buff will make me run a proc-a-thon.

[Cydrenlin]

I believed my previous post showed that dex increased proc rate /shrug.
It doesn't increase call of fire proc rate or flurry rate but I don't think it's an issue at 54 :)

[Sylviania]

well Cyndrenlin .... great parsing , lot of infos here ....
now i'll have to have a few SK and rangers to test out the others proc buff to check if they work same way than Call of Fire... ( they prolly do anyway :/

[Boub]

Could you make a test without CoF imbue, wonder if the number of EB procs would go up, had a discussion with someone who meants CoF is useless since it reduces the chance for EB proc.

[Gnomos]

ACtually, it's not attack rating itself that affects proc rates. It's attack rating vs mob defense that affects proc rates. You proc more vs mobs that are easier to hit. I proc like a mad man vs spiders in velks for example, vs giants in BoT I often shake my weapons at them in the hope they will do something. The same seems to be true for crits.