Comments on: Meters Love Mages

By: hooher tod

hooher tod — Mon, 05 Sep 2011 12:39:11 +0000

Yes there should realize the reader to RSS my feed to RSS commentary, quite simply

By: magebreaker

magebreaker — Tue, 12 Jan 2010 16:58:09 +0000

I play a FFB mage and this thread has intrigued me some i’m curious as to what the explination for why my dps is greatly inflated compared to most classes. If i understand correctly with arcane builds the primary factor is that there are no dots to keep u in the fight. Since FFB places a dot on evey mob it hits that is not fire immune, why am i pulling 2k 2.5k in my crappy 187 to 200 blues against hunters and rogues in 200 to 245 epics are getting about the same. while aoeing mobs i tend to do better than hunters as well but i assume this is becuase i keep flamestrike and blizzard going with living bombs sprinkle in for good mesure. i also usualy get beat in damage done overall per fight but i wait for warriors and druids to establish solid threat before i start and i notice most classes can’t wait the 5 secs that takes before unloading and the small amounts i get beat by can be attributed to that. or am i just pugged into alot of 5 mans with crappy players?

By: Frostheim

Frostheim — Sat, 05 Dec 2009 21:14:43 +0000

We’ve figured this one out. Recount calculates dps by taking damage done divided by activity time. The problem is that during long casts recount will pause activity time – thus the mages look like they weren’t attacking as much, even though they were.

I definitely consider this a flaw with recount that should be fixed. If you are standing there casting non-stop, your activity time should not stop — but that’s exactly what happens now. Looks like they just need to bump the activity pause timer by a half second or so.

By: Minastaz

Minastaz — Sat, 05 Dec 2009 20:54:58 +0000

I just want to point out that the ~1700 dps the mage “out-dpsed” you by has nothing to do with the 8.5% versus the 8.0% reported for your dps.

That percentage actually has nothing to do with the dps number given in that report. I’m of the opinion that dps shouldn’t be shown on the damage done chart at all.

The math works out properly when you compare the 8.5% for the mage as a percent of the damage done.

If you divide 2,928,966 by 0.085, you’ll get roughly 34,000,000 damage done. I’m rounding to the nearest million, because without the full damage done report and adding up the damage, I can’t be sure of the accuracy past 34 million. Also, the 8.5% is a rounded number anyway, without enough significant numbers to be more accurate thn that.

So to continue, what happens when you divide any number on that list by the percentage listed? You get roughly 34 million.

That pally sitting at 7.5%? 2,571,290 / 0.075 = ~34,000,000.
If you wanted to see by what percentage the ~1700 dps is higher than yours (as reported by Recount), you’d have to go the DPS chart, instead of damage done. The percentage reported there should be different. Though I’m not sure how it would calculate the percentage. Adding up the raid dps and using that as the denominator? Or just rank all of the dps below the top reported as a percentage of the max? I’m not entirely sure any sort of percentage is useful to a raid for troubleshooting.

By: Arcanemage

Arcanemage — Mon, 30 Nov 2009 00:05:23 +0000

My main is arcane mage and here’s the explanation:
As arcane mage you don’t have any DoTs ticking on bosses, ever.
Hunters use stings and whatever (i only have lowbie hunter alt so ^^), as well as fire mages for example. These DoTs keep their caster in combat, meaning that Recount / Skada will record the hunter/firemage continuously damaging the boss. However, as arcane mages don’t have any DoTs, Recount/Skada doesn’t count moving, going to fridge for a drink, anything between the actual damaging spells cast.

Solution: Don’t look at dps-meter at all.
Raidbosses die when their hp reaches zero. Therefore “damage done”-meter is what you want to be interested in. However keep in mind that in certain boss fights certain classes have certain “bitchduties” which have effect on their damage done.

TLDR:
Hunters have dots –>recount thinks: in combat all the time = lower dps
Arcane Mages don’t have dots –>recount thinks: not in combat all the time = high dps

By: Armin

Armin — Tue, 24 Nov 2009 18:35:00 +0000

Ina ddition to this, I haven’t foudn time to test it myself, but if this 2,5 secs timing would be an issue, we would also see very high dps with hunter autoshot, as that is usualy in teh 2,5 to 3 sec intervals as well. Also it should be possible to test with instants. A MM has 4 instants, so with 3 secods (2x GCD) delays in between we should be able to see if recount has too short timings. I doubt it, but science will show it.

Or, I could simply look at the code

By: Armin

Armin — Tue, 24 Nov 2009 04:58:40 +0000

“There would be a simple fix for Recount and that would be only turn off dps calcs if not in combat or no damage reported in combat log for 3/4 secs”

Recount stops dps when not in combat. It is excluded. And if Neven is right, the 3/4 not in combat log would actually make it worse with a 2,5 second delay between arcane mages:-)

In fact, the solution should – of Neven is right – be that recount only stops dps for periods of longer than 2,5 secs inactive.

However, it will be a simple test to see how that works. No pet, no sting and use instant shots with variable pauzes. It will be easy to see if there is indeed a hidden time-periode recount uses to detect non-dps.

By: Bindura

Bindura — Mon, 23 Nov 2009 05:07:12 +0000

I have to laugh a the people who immeditaly rebutt what Frost has said without much thought. The comment dps meters need IQ tests to install applies to you (my opinion). Frost is close in stating that Badass mage out dpsed him by 300 dps. Very simple calculations show that the boss fight lasted in the order of 515 seconds. (It is a mystery to me why Frost’s dps indicates he was active for the full length of the fight; unless he was brez’ed before his dots were off the boss) This is the mode (a form of average from statistics) for the time dps was on the boss.

Recount has counted Badass mage as being active for 416 seconds. I am sure that he did not run around blinking during the boss fight or /sleep. Looking at what Cocky mage said (based of his haste and recount stopping for 0.4 of a sec per cast) that is approximately 250 casts averageing around 11.8k damage each. (Does this sound about right for an Arcane blast). This completely ignores arcane missiles being used during arcane barrage which would push the average damage done for arcane blast up.

If we normalise Badass mage’s dps against the time that dps was on the boss it comes out at 5687.3 which is 380 dps more than Frost NOT 1700 which is what this post is about, recount reporting wrong numbers and these are the numbers that get you into pugs alot of the time.

I have a fun way of getting a great dps reading prior to pugging on a Saturday night. Run H HoS and on the final boss burn rapid fire readiness then rapid fire again. I can break 5k dps on that boss…and I average 4k in 10 man. I then have a great link saved in recount for Ony 25man…lf ranged dps 4k+ (haven’t done it but the thought did cross my mind). There would be a simple fix for Recount and that would be only turn off dps calcs if not in combat or no damage reported in combat log for 3/4 secs. Everone would suffer a really minor dps loss because of the extra time recount records at the end of a boss fight; but the difference would be negligible compared to RNG variables in the order of 1% loss across the board for a fight of this length. For shorter boss fights the error would be more prevolent. However Badass mage’s dps has been inflated by 25%; recount needs to be fixed (even if we all have to suffer a small dps loss for it). Or we will have to upload logs to World of Logs etc and link that instead and use the equivalent dps which is what Frost would like from recount.

On a side note top ret pallies dps is too low. Normalising his dps shows him at 5090.9

By: xavatar

xavatar — Sun, 22 Nov 2009 15:40:51 +0000

ive been having problems with recount for a while, notice that your dps does not significantly improve when lock n load procs and u slam x3 exp shot into the target and when u crit with killshot for say 17k thats 17k damage in 1 second, still the meter ‘lies’. i feel its because hunters r bound to the 1.5 second beat so damage per second is not good for our class maybe theres a way to measure ur damage in damage per 1.5 secs? or hope blizz put us on a 1sec beat like most spellcasters who can quicken their attacks with talents and haste thus reducing casting time, on dots…i have experienced the recount prob with dots on the training dummys, and its very annoying, u put everything u got into the thing and when u stop dps and passive pet thus leaving combat, the dots still tick at the low dps they do by themselves thus bringing ur average dps way down. i gave up on damage meters for dps a long time ago i do what frostheim does and go on damage done which is the best way to measure ur damage, if u still wanna calculate dps u gotta do it yourself, DBM will tell you how long the boss fight was, just divide your damage done by the seconds fought, mostly i get a different score to recount, plus if i wanna beat a mage in dps i ‘forget’ to drop imp hunters mark and md the bastard lolz only joking! all i can say about the whole dps thing is tell those who are ignorant of ur skillz that no matter what recount etc says, any raid is gimped without a hunter fullstop.

By: Noodleguitar

Noodleguitar — Sat, 21 Nov 2009 23:54:10 +0000

I have also been wondering a while how it is possible that I came first in damage done with 6.8k DPS, and then a mage followed second with 7.1k DPS.

I’m pretty sure recount is programmed to stop adjusting your DPS whenever you don’t have something hurting an enemy (Dots, a moron pet etc.). But in the case where for example you drop dead -because someone that is not you made a mistake- and Serpent Sting is still on the boss you see your DPS plummet by sometimes over 500.

It seems logical that mages wind up higher up in DPS as they don’t have DoTs. I’m guessing recount doesn’t see any difference between a Hunter being in FD with no DoTs and pet on the boss and a mage Cycloning himself for some mana.
This also fits with movement indeed, where Hunters are treated as standing still by recount and mages are treated as not being in the raid at all and so their DPS stops changing.