Half the man I used to be ... (maybe a quarter of the elf actually)

The point is, percent means "parts per hundred" which means it has already been adjusted for the sample size. But since you don't understand, I will explain:

Based on that parse data:

2H procs FA 63 times off 710 autoswings and 30 times off 436 CAs.
In other words, 8.87 times out of every 100 autoswings, and 6.88 times out of every 100 CAs.

2H procs GS 29 times off 710 autoswings and 17 times off 436 CAs.
In other words, 4.08 times out of every 100 autoswings, and 3.89 times out of every 100 CAs.

DW procs FA 14 times off 832 autoswings and 10 times off 195 CAs.
In other words, 1.68 times out of every 100 autoswings, and 5.13 times out of every 100 CAs.

DW procs GS 13 times off 832 autoswings and 4 times off 195 CAs.
In other words, 1.56 times out of every 100 autoswings, and 2.05 times out of every 100 CAs.

What this means is that for DW to proc FA 63 times, DW has to hit 3744 times (63/3744*100=1.68 hits per 100). Which is 5.27 (3744/710) times more swings than 2H. Clearly, there is no way DW weapons hit that many times more than 2H in any conceivable situation.

For DW to proc GS 29 times, DW has to hit 1856 times (29/1856*100= 1.56 hits per hundred). This is 2.6 times more swings than 2H.

Similiar results can be had for the other proc numbers.

The size of each test sample is irrelevant, except when determining margin of error. As Mordock posted, we are only interested in measurements versus time, such as proc % per second, or damage per second. It does not matter if one test covered 25 minutes and the other 10, as long as we make comparisons with respect to some unit of time.

That said, there were actually more combat art hits reported per unit of time while using DW than with 2H, so there was no bias against DW in this regard.
2H: 436 combat art hits in 1633* seconds (0.267 CA per second)
DW: 195 combat art hits in 661.5* seconds (0.295 CA per second)
* The time calculations are for comparitive use only, because we don't have access to the number of misses, or the haste of the tester. I reservedly assume that these two factors (hit rate, haste) were similar for both tests since it isn't explicitly shown, and would cancel itself out.

Arbitrarily doubling pieces of data like you suggest leads to improper conclusions like the ones Mordock reached.
najena.konk

With both of us using different words to say the same thing, I hope it gets across! In any case, we have proc %s, which can be applied to any time period anyway.

konofo, when you were linking all those pics from beta and stuff, did you have to upload those to another site then insert the image here using the URL or is there a way to just directly upload a image here?

I would still like you to attempt to justify the above claim that increasing the denominator on the right side of the equation [Proc % = Procs / Swings] increases the Proc %.

Try it, please.

SoE has no webspace for us as far as I know.

Actually, Solokov, there was no "unfair advantage" in figuring it that way...if in 1633 seconds he only cast a CA once every 3.7+ seconds, and in 622 seconds, he cast them every 3.1+, that's HIS doing, not mine...I simply multiplied by the difference in time spent in combat...if he had actually done a fair comparison of damage/time, we'd have better figures to work with, but he didn't, and seems to be actively refusing to face the fact that he's "stacking".

It doesn't matter if you multiply the 15 procs in 622 seconds by the figure that matches the ratio of compared times, then the 2 CA generated ones, and add them, or add them, THEN multiply by the figure...you get the same answer. Any "advantage" given the DW there was HIS doing, for firing them more often in the comparitavely short period thathe ran the DWs, and I can't help that.

Did you happen to notice that his total proc worked out much differently than he represented it, as well? THAT is what I was insinuating. "total auto-swing hits" is counting each weapon that impacts...which happens twice per swing...so coming up with a 2% figure is a SINGLE WEAPON figure...divide the melee hits by 2 to get the "per swing" chance.
Message Edited by GurgTheBashur on 09-26-2005 04:17 PM

You can't figure DPS without there being a "seconds" part of the equasion, man!

And it's NOT a fair comparison to count prcs you got in X time with one set to procs you got in Y time with another setup, NOR is it accurate if you alter the CA ratio to auto-swing...you end up comparing apples and oranges...and if you don't see that....well, all I can say is that I've got a GREAT Corvette to sell you, nevermind the "AMC" emblem on the side...
The assertation you've been denying this whole time is "with equivalent imbued weapons, the DW has an advantage in AS and proc damage done over the same period of time"...you haven't tested them over the same period of time, you haven't tested the as and proc damage alone over that time, and you've done every damned thing you could to skew it in favor of your pet theory by mismatching CA damage and procs in the mix...if you're going to try and prove your point, at least bloody compare the same damned things...same period of time, same number (and type) of CAs in that time, same period under AS and CA, etcetera.

And if you ever DO bother trying an honest test, I'll tell you right now what results you'll get: It'll turn out that I'm right, insofar as AS damage over X period and proc number spawned by AS swings.

My own tests have actually EXCEEDED my expectations, with 7 minutes each, 63 and 62 CAs respectively (had one more DW CA fail to land), and 8 total and 18 total procs, respectively (granted, 3 of 8 with 2H were CA spawned, and only 2 of 18 with DW were)...but 7 minutes isn't long enough a statistical sample, so I was intending to run 30 minuts pure AS against level 38 carrion dreg groups tonight, and then another 30 minutes of each with a set number of CAs per encounter.
Message Edited by GurgTheBashur on 09-26-2005 04:34 PM

But the point is, after you accused him of stacking, we find that the figures show if anything, the stacking was in FAVOR of DW. And I have shown even tho this is true, 2H still procs SIGNIFICANTLY more often than DW does. My point is that you accused him of doing something and then, perhaps unknowingly, did it yourself (altho it did nothing to help your cause), because those of us who understand math realize that total procs means little, it's % Proc rate that matters, which, no matter how long you extended the DW parse, would remain about the same.

As for whether that was a single weapon, or if he DW. Well, now you are just calling a liar, not much I can say to that. It is of course interesting to note you support the numbers when it suits your purposes, and you attack them when it does not.
Message Edited by Sokolov on 09-27-2005 05:10 AM

Notice that Gurg originally stated other people's numbers said he is right. NOW he is saying the numbers are stacked and that if they were true numbers he'd be right. Makes me giggle, this one.

Actually, given that he was parsing proc rates, and not DPS. His addition of DPS figures in the last post was additional info not included in his original calculations. Of COURSE the parser did DPS too. But we were interested in proc rates. Why? They tell us a good picture if the weapons ARE equal.

And it is a perfectly good comparison, when we are concerned with proc rates, not to include a time element so long as the sample sizes aren't extremely small or disporportionate. You seem to be able to do math, so why do you suggest Proc % Rates derived from such a sample are not comparable? They certainly are.
Message Edited by Sokolov on 09-27-2005 05:11 AM

Nobody is attempting to show absolute DPS here. We have been discussing the relative difference in DPS between a DW setup and a 2H one. You do not need to actually measure time, nor damage to do that, because the in-game examine windows do that for you (that's their job).

Please elaborate on this statement. Please tell me how, when trying to show the difference in proc rate between two different setups, is it unfair to measure the proc rate in those setups and compare the rates?
And please tell me why it matters that there was a different ratio of CAs to auto-swing in the tests. The proc rate (chance for a proc to fire when you hit) does not change if you are buffed versus unbuffed, in fabled gear or naked, mashing buttons or making coffee. It only changes if you change weapon speed or type, thus all you need to measure is the number of procs, and the number of hits. The only thing you gain by making either, or both of the tests longer is a value closer actual rate in each setup.

Riiight.

You do not need to measure the same period of time to show which procs more frequently - that's what we have multiplication and division for.
You do not need to measure auto-swing damage as the weapons have equivalent damage ratings in the examine info (DW is half the 2H).
You do not need to measure proc damage, as that is in the examine.
You do not need to run seperate tests for auto-swing and combat arts, because I parsed the logfile to seperate the auto-swing and combat arts, and how many procs fired from each - you can use subtraction to get the auto-swing rate alone.
You do not need to use the same number or type of combat arts in the tests, because combat art damage is constant when changing weapons.
The only thing that was in debate is what that "5% chance" actually means, and that's what the test showed.


To reiterate:

For 2H:
"procs per minute of auto-swing" = "chance to proc in the examine" / 100 * 60 / 3
"chance to proc per hit" = "chance to proc in the examine" * delay / 3
"procs per 100 combat arts" = "chance to proc in the examine" * delay / 3

For 2x DW:
"procs per minute of auto-swing" = "chance to proc in the examine" / 100 * 60 / 3
"chance to proc per hit" = "chance to proc in the examine" * delay / 3 / 2
"procs per 100 combat arts" = "chance to proc in the examine" * delay / 3 / 2

For 1x DW:
"procs per minute of auto-swing" = "chance to proc in the examine" / 100 * 60 / 3 / 2
"chance to proc per hit" = "chance to proc in the examine" * delay / 3 / 2
"procs per 100 combat arts" = "chance to proc in the examine" * delay / 3 / 2

For 1H:
"procs per minute of auto-swing" = "chance to proc in the examine" / 100 * 60 / 4
"chance to proc per hit" = "chance to proc in the examine" * delay / 4
"procs per 100 combat arts" = "chance to proc in the examine" * delay / 4

[ Edit : Coloured the 2H vs 2xDW cases green to highlight where they are the same, and red to highlight where the 2H is better ]

And that "test" is a) meaningless without stating what weapons you are using, and b) lies/fake.

Please try harder.
Message Edited by Pin StNeedles on 09-27-2005 06:53 PM

If you buy the "advanced features" stuff you can directly load images and then xref them here

Solokov...maybe I'm not being clar...I dunno, I thought I was being crystal clear, personally.

YES, the 2H procs "more often" in comparison to the number of times it hits than a DW does...but, given the same time period, the DW HITS MORE OFTEN.

What I'm saying, in simple terms is: it may be true that the 2H procs 1 in 4 hits, and the DW pair, together, procs it 1 in 16 hits BUT when the DW hits 16 times in the time it takes the 2H to hit 4 times, they proc exactly the same number of times in a given time period.

Getting it now? THIS is why the time period is important. It's not "how many swings you have to make to get a proc" it's "how long do you have to swing on AS to get one"

Thus proc rates DON'T give a good indicator if the weapons are equal or not, unless you figure them in procs per second of combat, which he did not, has not, and seemingly refuses to do.

And when I tried to balance time periods to demonstrate that the procs per time in combat WERE comperable, you accused me of giving an advantage to DW off his CA calculations, when it wasn't true...if ANYTHING, HE gave that "advantage" by lanuching more CAs per seconds...though I don't think I'd count that as an "advantage", per se, since we already know that, due to the proc-per-swing ratio, the CAs give advantage to the 2H, in comparison, and the CAs take the same time regardless of weapon...so "stacking" CAs "in favor of" the DW is actually stacking the ODDS in favor of the 2H.

I dunno if you're just not making the connection in what I've been trying to say this whole time, or what, but from the get-go, I've been saying "if you spend the same time in "melee" (auto-swing) combat with comperable DW and 2H weapons"...and I've REPEATEDLY pointed out you'd have to be blind, or terminally blonde, at the very least, to NOT recognize that on a "per swing" basis, the 2H has the DW beat to the four winds...but the ADVANTAGE is in the fact that DW hits "faster"...meaning, it hits more often in the same period of time, procs more often in the same period of time (even if it procs less on a per-hit chance), and so on...which balances the DPS readings, if it doesn't stack it in favor of the DW.

Now I admit that this totall disregards the impact of time under stun, casting time for CAs, and so on...but I don't spend even 1/10th of my time under stun, MOST my CA's as a berserker are instant or less than my delay WITH DW to cast, etcetera...so the only REAL impact that either of those factors has is in favor of 2H, but is NOT "slowing down" the DW...there are 4 that actually "slow down" a DW...whirlwind, BA, Stunning Cry, and Cold Retribution. All of which are on a timer of at LEAST 1 minute, two of which I don't normally use in regular combat.
SO let's say I use ALL FOUR in a particular encounter...since the average cast time for them is 3 seconds, that's 12 seconds...1/5 of a minute eaten up by those CA casts. I'm STILL spending 48 seconds of 60 under auto-swing during that minute, which means I've STILL got the "procs per second" advantage of the DW working for me, if I'm DW.

The only way to REMOVE that advantage is to spend enough time castig CAs that the % advantage the 2H has on a "per swing" basis "catches up" to the total number of swings the DW's will make during the time that you aren't casting the CAs. I won't deny that this is possible...but it also involves burning ALOT of power with each encounter, because you have to equalize the "delay ratio" to manage it, so to speak...in other words, you have to deliberately contrive circumstances which make the speed advantage of DW non-existant.

I already addressed this in an earlier post, but here it is AGAIN for your benefit.

Based on that parse data:

2H procs FA 63 times off 710 autoswings.
In other words, 8.87 times out of every 100 autoswings..

2H procs GS 29 times off 710 autoswings.
In other words, 4.08 times out of every 100 autoswings.

DW procs FA 14 times off 832 autoswings.
In other words, 1.68 times out of every 100 autoswings.

DW procs GS 13 times off 832 autoswings.
In other words, 1.56 times out of every 100 autoswings.

2H Proc Rate / DW Proc Rate = Amount of Extra Attacks Needed by DW to Achieve the Same Number of Procs in the Same Time Period
For FA on Autoswing:
8.87 / 1.68 = 5.28
For GS on Autoswing:
4.08 / 1.56 = 2.61

What the above calculations tell us is that for Pins' DW to have procced the same amount of FAs in the same time period (whatever that may be), DW would have had to swing 5.28 times as fast as 2H.
And, for GS, 2.61 times as fast.

Note that the above does not change based on the sample size, nor does it change based on whether the same samples entertains the same time interval. THOSE variables have a bearing on the reliability of the figures based on the central limit theorem and law of large numbers which have to do with the probability of these figures being representative of the true values they are meant to survey.

Certainly you do not expect me to believe that DW attacks are that much faster? (Especially in normal combat situations, with all the variables of haste, stuns, cast times, etc. incoporated as well)
Message Edited by Sokolov on 09-27-2005 09:57 AM

Wow. You've actually got it. "they proc exactly the same number of times in a given time period".
Just remember that point.
"they proc exactly the same number of times in a given time period".

Oh darn, I think you've lost it again
I have not considered procs per second of combat? I believe I did that in the first post I made in the thread, and the second, and every other one. The chance to proc as stated in the examine window IS procs per 3 seconds of combat (I hope you're able to divide by 3), thus is the easy part to look at and see they are equal. Proc chance per hit matters when you are considering combat arts, which is what the other half of the discussion was.

Sorry, did you just say "DW ... procs more often in the same period of time" ?
I thought you'd already agreed that "they proc exactly the same number of times in a given time period" ?

"they proc exactly the same number of times in a given time period"

"they proc exactly the same number of times in a given time period"


And now we have that clear, you can go and consider the proc rates while using combat arts.

Pristine Imbued Fulginate Greatsword Pristine Imbued Fulginate Short Sword (X2)
Damage Rating: 34.3 Damage Rating:17.2
Damage: 56-168 Damage: 13-38
Delay: 2.5 seconds Delay: 1.2 seconds
Rated Proc Chance: 5% Rated Proc Chance: 5%


Under 30 minutes of timed combat, purely AS, against grey doubleups level 31-32 golems in RoV

Pristine Imbued Fulginate Greatsword Pristine Imbued Fulginate Short Sword (X2)
691 hits (96% landed hits, just about) 2910 hits (97% landed hits, just about)
28 GS procs (4.05% actual, 4.05% per swing) 58 GS procs (1.99% actual, 3.98% per swing)
95,358 total damage from melee swings 98,940 total damage from melee swings
4,368 total damage from GS 8,990 total damage from GS
99,726 overall damage dealt 107,930 overall damage dealt
55.403~ DPS 59.961~ DPS
138 damage per melee hit average 34 damage per melee hit average



Under 30 minutes of timed combat, with CA, against grey doubleups level 31-32 golems in RoV
For this bit, I did BA, WW, RB, FR, all adept I, in a cycle every time BA became available (every 45 seconds), in the same order
CA stats
RB 0.5 cast 20 second recover 133-223 damage
WW 2.0 cast 30 second recover 84-254 damage
BA 2.0 cast 45 second recover 137-413 damage
FR 0.5 cast 20 second recover 43-129 damage
As you can see, only WW and BA give any "advantage" in time to 2H by casuing the DW to "miss swings" during cast...or rather, by chaining them as fast as they could cast, I spent 5 seconds to get 4 swings for 4 hits and 8 hits respectively, plus CA damage, with a "calculated chance to proc" of 2.08% and 1.04% per hit respctively (4 hits in 5 seconds is 1.25 "delay" and 8 hits in 5 seconds is .625 "delay", so the calculation, according to Pin, is ( 1.25 / 3.0 * 0.05) and ( 0.625 / 3.0 * 0.05) )
The ACTUALS I got out of this are out of line with that calculation, though, and aligned more with the "per hit" figures from melee (which threw me off a bit, since you still hit with both weapons with a CA when dual-wielding...does it maybe only count the FIRST weapon impact as the CA powered one? This is possible, even probable, since my CA damage indicators come floating up from mobs in-game with one being "boosted" and glowing, and one being normal, with the purely CA damage rising in the middle of both, as a seperate figure, and parsing similarly shows that the CA damage is calculated, then one weapon has higher than normal damage, and the other has normal damage, with weapon-involved CAs...parses read a bit differently, and seem to include all of these in one figure as a total for the attack, but the combat/spell window seperates them as "171 damage from whirlwind" "43 slashing damage" and "32 slashing damage" respectively, for example)



Pristine Imbued Fulginate Greatsword Pristine Imbued Fulginate Short Sword (X2)
718 hits 2939 hits
39 CAs landed (40 cast) 38 CAs landed (40 cast)
30 GS procs (4.17% actual, 4.17% per swing) 59 GS procs (2.007% actual, 4.014% per swing)
99,802 total damage from melee swings 96,987 total damage from melee swings
3 GS off CA (7.69% chance off CA) 1 GS off CA (2.63% chance off CA)
4,710 total damage from GS 9,086 total damage from GS
6,992 total CA damage 6,775 total CA damage
111,504 overall damage dealt 112,848 overall damage dealt
61.946~ DPS 62.693~ DPS
139 damage per melee hit average 33 damage per melee hit average

Now, as you can see, I seperated the "weapon" damage from the "CA" damage here, which I did by comparing the spell/combat window to the parses...parses, for some reason, give a total for the attack, while the spell/combat window shows what the CA did on its own, and the slashing damage done by the weapon.
And when you do that, you can see the procs stay in line with the "per hit" percentage, not the "per swing", and that, for whatever reason, the game counts ONLY the first weapon to contact during a CA to be "CA powered" (and, apparantly, to be the only one eligible to proc during a CA). I don't know this to be true, but it makes sense, and supports the data.
Now THIS means that parses will (and do) show a higher damage per CA with a 2H than with DW, because, well, to put it baldly, when you add 139 average swing damage and 280 CA damage from a cast of BA while holding a 2H, it adds up to more than when you add 33 average weapon damage, 33 average weapon damage, and 280 CA damage from a cast holding DW...and the fact that CAs boost what damage the first weapon to hit when "powered by a cast", it just widens the gap (in other words, adding 10% damage to the "weapon swing" component of a CA for the first weapon to impact is a greater advantage when there's only one weapon, and it packs a mother of a punch....it makes the difference between " 153 + 280 = attack damage for this attack " and " 37 + 33 + 280 = attack damage for this attack".


But ALL of this STILL means that, unless you are casting your CAs often enough and fast enough to "eat up the difference" between swing speeds, the DW still proc weapon-based procs more often in the same period of time, whether it's 1 minute, or 2 hours (I didn't test with any stance-based, and OTHER gear-based procs aren't in debate).

As for stance based procs...well, most are "when target takes damage" for the "primary" part of the "spell"...your parry and dodge chance is your parry and dodge chance...2H, or DW, this isn't changed (it drops with 1H/S, naturally, since avoidance/block goes up, and you take less damage), but there IS a component of several of them (most notably, the "Rage" chain of self-buffing stances) that is "on a successful attack, this skill gives X% chance of casting <blah> on <caster, target of attack>"...and for THESE skills, Pin's 100% right...there's a SINGLE INCIDENCE of the skill present, which means that the calculation is based on ATTACKS, not HITS...you make 2 hits per attack with DW, but that doesn't help you one little bit, because the calculation is "delay/3.0 * chance to proc" for the ATTACK, and you're doing nothing to improve this by making one attack include two hits...so in THESE cases...well, 1.2/3.0 * 0.1 = 0.04 for a 4% chance to proc FR on any given attack (two hits) with DW, or once every 30 seconds of AS combat (once every 25 swings), and 2.5/3.0 * 0.1 = 0.083~ for an 8.3% chance to proc FR on any given attack with 2H, or once in just under 30 seconds, each time (1 in 12 swings, nearabouts). Over time, this will give the 2H an advantage, since it's right on once every 30 seconds of swing time for DW, and just under that for 2H.
In terms of CAs, this is a HUGE advantage to 2H, since they MASSIVELY increase the number of attacks they make in a given time period with CAs (in my test example, they do 4 attacks in 5 seconds)







[expletive ninja'd by Faarbot], trying to make this readable was a pain

Message Edited by GurgTheBashur on 09-27-2005 11:46 AM

(Color-coded responses for readability)

First: I've noticed a flaw in your terminology that is probably causing some misunderstanding. When you're dual-wielding two weapons with the same delay, every time that delay expires, you make two swings, not one. You make a left-handed swing, and a right-handed swing, and they are independent. You might not notice because their timers are synchronized. Calling it one "attack" with two "hits" is flawed and leads to confusion. Equip weapons with different delays and damage types to better understand what is really going on.

Second: We have two stances - offensive and defensive. The defensive stance has no proc associated, and the offensive stance has an offensive AE proc. So at face value, the highlighted statement is false. Now, if I assume you meant to talk about buff-based procs in general, we have one defensive damage proc with a sub-100% rate (Hold The Line / Taunting Defense / etc.) and a few offensive procs with a sub-100% rate (Fury line, creature mastery strikes). Our berserk buffs have both offensive and defensive triggers. Most damage-dealing proc buffs from other classes are offensive in nature, except for damage shields. It is the offensively triggered proc buffs that make 2H the superior weapon choice, and there are a lot of them out there.

Finally: This really is the bottom line; none of the situational intermediate findings matter. I'm not convinced you fully understand why quite yet, but the result is the same: To achieve the highest DPS, you should equip a slow 2H weapon and use as many offensive proc buffs as possible. Like we've been saying for months.

najena.konk
Message Edited by konofo on 09-27-2005 12:36 PM

Gurg, please.
It's okay being wrong, particularly if something works in a non-intuitive way for you, but you really don't need to post fake numbers to try and prove your point.

And as some were unhappy with the duration of my DW parse, here is another, taken from 60 minutes of pure autoswing using the same 2x 1.5sec DW imbued batons as before:

4180 hits
52 gleaming strike procs

That gives 1.244% chance to proc on a successful hit, which scales to 2.49% chance to proc with each weapon in 3 seconds of autoswing, or 4.98% chance to proc with either weapon in 3 seconds of unhasted autoswing, which is almost exactly the 5% chance stated on the weapons.



Now, I don't know if you just made your numbers up, or just doubled the number of Gleaming Strike procs that you parsed, but as you clearly aren't going to admit that you're wrong on these boards, I'll post http://www.btinternet.com/~challand/DWProclog.txt



(And in case someone questions whether it's just luck that Gurg parsed double the proc rate that I did, the chance of that happening with the sample sizes used is billions to one)

Nice parse Pin...you've convinced me man