Parsing methodology

[Gutterr]

The DPS posts got me thinking so I went to the test server to work on some parses.  I was only interested in the base damage from my different weapons.  My methodology is as follows:

I clicked off all buffs but my mount   
I used a lvl 85 test dummy
I 30 minute parses
I didn't cast
I didn't have my warder up
I stood in front of the test dummy as I whacked on it.

If my methodology is flawed, please let me know.  This is my first attempt, so I may be missing a crucial component.  I know the arena dummy isn't perfect.  It wasn't my intention to get a true parse against a real world mob, simply to compare weapons.  I did do a quick test on the Riverdale mob referred to elsewhere on these boards.  It died fairly quickly so I set out to the arena.

A quick rundown of the parses are as follows:

Rottrued's Enforcer in MH & Fabled in OH = 783 DPS
Fabled in MH & Rott in OH = 973/974 DPS (I ran two separate 30 min parses)
Fabled in MH & Jagged Alloy Repeato-Shiv in OH = 951 DPS
JARS in MH & Rott in OH = 627 DPS

From what I have read on these boards, the Fabled becomes obsolete with Crystallos weapons. Given this data, is this really the case?  The actual damage from the Rott in the MH was more, but the pet procs more than made up the difference. Granted this is against a test dummy and I didn't use any discs, AAs, or spire.

I normally parse our raid bosses.  We are currently floundering in MMM, but we are farming Rottrued, BMK3, Bimbilicous and some of the easier MMM mobs.  I parsed at 4k DPS for the first time last night on Rottrued, but I normally hit in the 3-3.5k region (had shm, zerker, rogue, mnk in my group).

So I guess my questions are as follows (I'm not a science guy by any stretch of the imagination, but I'm trying to approach this as logically and "scientifically" as possible):

1) Did I use a sound methodology for getting a base parse to compare weapons?
2) Are there any other methodologies that I can use when parsing?  (longer parse time, keep some buffs, etc)
3) When running parses with spires, discs, warders, etc do you have a set number of times that you fire each a set number of times or is ther a different approach? 
3a) It would seem that using spells, discs, spires, etc would introduce variables that would be hard to account for in comparing weapons.  Am I over thinking this?
4) At what point are there better weapons that will make up for the lost damage from the Fabled pets?
4a) Is the loss of the pets made up for by disc damage with better weapons even if base damage is lower?  (I've seen this reasoning given on the boards, but has anyone done the parsing to compare actual difference in DPS?)

Thanks in advance for your input. 

[Khauruk]

Having no haste up is going to severely gimp the non fabled SPFs.  Sev-damn-erely.

By seat of the pants calculation, you should be around 900dps w/ shiv/rotty @ 8% overhaste.  SPF won't go up as high, relatively speaking.  It will outparse it still, but not so severely.

Yes, bandolier in shiv/rotty when you're on raid mobs.  Pets won't work as well against the high AC mobs, the push does add up, it's irritating for other players, and they're shitty while using BA/EF/GBA/group warder/warcry/sham click/etc, as none of these hugely multiplicative damage boosters affect the proc pets one iota.  No parsing is needed, to be honest. If you want though, take the pure damage portion of fabled/rotty (non-pet), and use a calculator to factor in some discs.  That setup will be blown away compared to rotty/shiv.

Next parse:
Hit test copy once you have raid buffs, then parse fabled/rotty against shiv/rotty.  Use a lvl90 dummy, as they are all gimp, and the lvl90 more closely reflects real mobs.

[Nusa]

Augs: If the weapon doesn't have your best +damage aug in it, you aren't testing it at the actual ratio one would use it at. The value of the aug is inversely related to the weapon speed (the faster the weapon, the greater the benefit). Improves melee results.

Worn Equipment: Accuracy, Combat Effects, Strikethrough, Cleave, Ferocity, +attack, +stats, etc. If any of that isn't maxed, you'll get a different result than me. Improves mainly melee results.

AA's: All the passive offensive AA's for yourself. If you're missing any, you'll get a different result than me. Both melee and proc results improved.

Buffs: At least some standard self-buffs, since you can always count on having those. Go for raid buffs if that's the scenario you want to test. Improves mainly melee.

As you can see, most of that stuff improves melee performance while only a few things (some AA's) improve proc performance, so your existing have-nothing parses are heavily weighted toward proc performance.

Or saying that another way, a poorly geared low-AA toon who primarily solo's or groups may indeed find the proc damage far more valuable than someone at the top of the scale....with exactly the same weapons!

[Denti]

For baseline parsing you need to parse much longer than 30 minutes. At least 3 to 5 hours each configuration is needed to even out the extremely streaky random number generator that is used within everquest.

I would suggest to get your usual buffs including haste (which is crucial), added atk does not really help as the test parse dummy has near to no AC in its standard configuration. However, when posting a parse state which buffs you did use (and basicly which melee AAs you have).

Another point for pure synthetical weapon comparison is that you should parse from behind to avoid dodge, riposte, parry and block. Those can be mitigated some by your mod2s, however it would be better to avoid that problem alltogether.

[Khauruk]

Depending on your goal, Denti is right.  It's not as much about time as sample size though - depending on how close the situations you're looking at are, you need a certain sample size though (not even based upon RNG, just statistics), to obtain a certain confidence level.  Two gear setups w/ a 5AC difference would take a very long parse, versus two setups with a 500AC difference, for example.  If you're looking for a 99% confidence vs. 95% (the typical), you're going to need to parse much longer as well.

Comparative parsing, a simple question of which is better, can often be done in a far shorter time.  The odds of your 627dps parse ending up greater than the 973 parse are veeeery miniscule, for example.

I would do a longer parse specifically w/ the SPFs though, since your sample # of procs is just as important as melee swings, and the sample size is much smaller.  Though your numbers came out very nicely close.

[Gutterr]

Thanks for the input.  I will work on a parse using this information.

Much appreciated.

[trixbro]

Gutterr, I would be very interested in the results from your parses as I am in a similiar situation weapon wise and wondering if the benefit from rot/shiv is big enough to bid on or wait for a further upgrade.
thanks

[Mazame]

Another thing to think about is how your play style is and group make up.

When you have a group made up of melee focus ( war cry / shammy epic / Bard ) then the higher damager weapons are going to come into play.

When your in a group with ( Druid epic / chanter arua ) Then your weapons wont matter much. I say this because  when your casting your melee attack are not going off. I got 4 DD and 2-3 Dots loaded + wolf pet and so when I am burning my DPS is not comming from weapons it coming from the spells. and when you keep the spells gray you can see a drop in the melee damage you put out. With this said how proc work is based on time so the few swings you in get between casting can proc and by doing so is like adding another dot.

I run the Fabled and don't see my self giving it up any time soon.  But I feel it some thing you have to take a look at from all side and based on what your trying to gain. the truth is were not going to be the top DPS  ( Mages / Wizards  out do us spell wise  Rog / ber are going to do it melee wise. )  bst are a fun class to play that give you a good mix of everything and if you enjoy it that why your here. but if your looking to be the top then 90% of the time it not going to happen unless your out gear / out aa those you playing with or if the are slacking.

[Razimir]

When you have a group made up of melee focus ( war cry / shammy epic / Bard ) then the higher damager weapons are going to come into play.

Bard over haste affects on proc pets too.

-Raz

[Gutterr]

Well I finished up some additional parses.  I was raid buffed for all parses.  I compared the fabled and the shiv.  There were four total parses, two with discs, two without.  The parses without discs were done over a 30 minute period.  The parses with discs were only done over 10 minutes.  (I'm lazy and didn't want to hit Foray and Feral Swipe for an hour)

Without discs:

fabled/rott -1519
shiv/rott - 1124

With discs:

fabled/rott - 3137
(my total damage was 2726, pets were 430) - I know they don't add up correctly, blame the parser
(of the 2726 - maul 1357, crush 1182, hit 177, dir dam 9)
shiv/rott - 2674
(of the 2674 - pierce 1090, crush 1055, maul 367, hit 139, dir dam 21)

When I used the discs, I activated BA, EF, and Spirit Frenzy.  My buffs were locked, so they all stayed up.  I used First Spire 6 minutes into each parse.  I hit Foray and Feral Swipe every 30 seconds (10min, 9:30, 9, 8:30...).  I know the utilization of BA, EF and SF won't mirror real world situations.  The intent to to measure the overall DPS.  The burst in a real world situation might be better with better weapons, but that doesn't appear to be the case here.  The weapons are fairly similar the way I have them auged (fabled 40/18, shiv 44/18).

[Khauruk]

Your offhand damage in the fabled parse (disc) is 12% greater than in the shiv disc parse.  That is ginormous variance.  It needs to be significantly longer.  Offhand numbers should be within a couple % of each other to be useful.  You also show your personal damage w/ the worse weapon set as bring greater than with the better set, also removing any credibility from it.

EF is also overriding any effect from BA.  You don't disc like that in fights, I hope.  EF, then BA.  They must be separate to be non-useless.

[Gutterr]

I did some further parsing with the fabled and Shiv.  Raid buffed.  Here are my results:

Shiv + Rott
31547sec
837 DPS:
413 crush
399 pierce
24 DD

Fabled + Rott
32701sec
1377 DPS:
410 crush
361 punch
9 DD
595 pet

No discs at all.

[Camikazi]

Yea I parsed FSPF against the MM 56/19 piercer with a 4 dmg aug making it 60/19 and no discs FSPF still won, with discs and Intensity running Spear was about equal to FSPF. It seems Extended Swarm working helps FSPF a good bit, DPS on the pets went from around 400 DPS before change to 535 or so after change. That bump makes it impossible to upgrade for groupers, since the 56/19 from the MM was too much for now, and difficult for some raiders to get rid of as well.

[Gutterr]

Rott + Fabled
35158sec
1087 DPS:
430 crush
333 punch
19 DD
305 pet

Rott + Shiv
15623sec
834 DPS:
434 crush
373 pierce
26 DD

Last ones.  Similar DPS, but 300 loss in switching Fabled to offhand.  Hope this helps if anyone had questions about parses.