Here is Euliat's Frost spreadsheet: https://docs.google.com/document/d/1...XsL2K5ExM/edit
Here's a post explaining some of the functionality: http://us.battle.net/wow/en/forum/to...01256?page=4#6
Message him on MMO if you want more, he's quite active on these forums.
(and probably find something incorrect, because i am annoying like that ^^)
---------- Post added 2012-12-24 at 02:11 PM ----------
Well ive studied his code for about an hour or so.
I dont see any comparisons between stalling for a potential km proc and not stalling which will regen resources faster. And without running the sim to see how it handles rp regen, we arent going to know.
The basic premise is the faster you use runes, the more rp you generate, which in turn regens runes faster, which in turn regens rp faster, etc etc etc
Does generating resources slower amount to more dps because of a potential km proc?
Without the maths to verify one or the other, all arguments are invalid and nothing more than guesses.
Last edited by mmociloveu; 2012-12-24 at 02:11 PM.
We currently have plenty of room in our rotation to game KM procs off swings with no downside, Dossou is 100% correct in saying it is a dps gain, even if it is only 0.5%.
I agree with everything but your conclusion. Nobody should get a swing timer, because the best case benefit is ridiculously small and it's not worth splitting your attention for a tiny thereotical performance boost dwarfed by the RNG. People trying to time their KM-buffed attacks are doing themselves a disservice.
I don't use a swing timer personally, if I were staying 2H for much longer I would use one, as I'm sure reacting to swings would become second nature, but I will be going d/w when the rng gods give up their 1h purples! On a patchwerk style fight I can see the benefits of using one, when there's little else but your rotation to focus on, and for people who like to min-max why not squeeze that little extra?
Because there are no patchwerk fights, and trying to squeeze that little extra (and it is a little extra) means you are likely to make a mistake elsewhere. But you're the best judge of your own skill; if you really feel you can execute that perfectly, go for it. I know I can't.
If you have a swing timer of 2.5 seconds, and you are waiting the last 0.5 seconds for a potential km proc, you might have 2 x obliterates ready in the next swing timer and a freezing fog proc, meaning you cant dump these abilities, dump RP AND wait for the next swing timer for a potential KM proc.
There are so many situations where waiting is a dps loss, and i really dont believe any human on the planet can play it to perfection, every 2.5 seconds, for minutes let alone hours on end.
You're far more likley to screw something up and cause dps loss than gain in all reality. All this and not even considering isp and server lag.
You are far better off doing it like i do, which is to track how many seconds left on your next obliterate, and if KM has procced with obliterate being ready in the next 1-2 seconds, wait for that and completely disregard the swing timer.
So much easier to track, and so much more chance to reach maximum dps potential.
It truly is not worth the effort for so little gain.
Not at all.
When i was hardcore i used to rank top 20 dps for dk, and right now i am casual only have ilevel 492 WITHOUT 4 piece tier set and i have killed 2/4 rank 8 bosses in brawlers club and have top 50 ranks in some 25 normal/heroic fights. No easy feat with such low gear.
Without being elitist, i know my playstyle is in the top 0.1% of wow players and not once have i ever tracked my swing timer.
Last edited by mmociloveu; 2012-12-24 at 09:30 PM.
Well it definitely isn't for me. I know there are better players out there. Maybe some of them can execute it properly and see that 0.4% gain. It's certainly possible, it's just a lot of annoyance for a chance at a very small gain. But that doesn't mean it's good advice for the general public-- exactly the opposite.
A post along the lines of "ok, point conceded on the concrete reasons for waiting/ignoring KM, let's agree though that spending effort to make the stars align for 0.4% is likely going to hurt your performance by more than 0.4% in other areas" or something similar, would be much "wiser".
The reply at the moment is more like a petulant child being stubborn about being wrong.
---------- Post added 2012-12-24 at 09:46 PM ----------
Last edited by mercutiouk; 2012-12-24 at 09:44 PM.
Originally Posted by BoubouilleOriginally Posted by xxAkirhaxx
No, I just didn't go into great detail. I shouldn't need to get into the nitty-gritty explaining the exact numbers and reasoning in every single post on random forums.
Someone said "should I wait on obliterate" and I said "no". That was the correct answer.
Please, on your next post, include exactly what Simcraft priority list you used to differentiate between waiting and not waiting.
You keep deflecting away from the fact you have "evidence" when clearly you haven't posted it yet.
Officially Retired Until Further Notice
Duality US 4th - Recruiting Exception Players
The standard simcraft T14H profile is 0.4% over the same profile with the KM actionlists commented out.
Again, I posted this in other threads. Someone else (I thought it was you) also posted their results, and I agreed with them.
Aslong as you are keeping your runes recharging and not capping RP there is no resource loss, you will not run out of GCDs with current levels of haste, and you will not magic resources out of thin air by using all your abilities ASAP.
I cannot understand how some people fail to see this... You can run sims all you like, some things just come down to common sense, a bit like the Str/haste gem debate.
As for gear and rankings; I am similarly geared to you and also rank top 50 regularly, I was using a blue weapon and no tier until recently, I do not see what relevance this has?
My play style is similar to yours- waiting with KM procced 1-2s until the next OB, but if I wanted to improve my dps I would use a swing timer in the fashion Dossou mentioned.
Mathematically, the hard "limit" on resource generation is shown in this workbook for 2H Frost (PL is not perfect):
If sims did not waste any time or resources at all (via GCDs primarily), that's the throughput you would be looking at. Since you are not GCD-locked yet, the result from that spreadsheet should line up pretty dead close to sims given the same inputs. Also, as you can see, since you are not GCD-locked, you do have some free time to adjust abilities around, but the amount you have is very limited before you start causing loss. As you stall more, you cram more of your abilities into a smaller window of time. After a certain amount, stalling will inevitably lead to pushing abilities out of that window; this, then, is that loss. The previously linked post goes into more detail about this.
I had already done the analysis on whether or not to stall during Beta, as I have extreme control over the ability priority in my sim, and it was the only DK sim at the time; the conclusion was that you basically should not. Stalling did increase DPS via more KM-OB's, but it also decreased DPS via lost actions. The latter effect was prominent enough such that waiting more than 0.50s (up to, not blindly) resulted in an overall loss. Unfortunately, those forums are gone along with that data, and I didn't plan to be around later to need it anyway.
Since this issue has resurfaced here, on the official forums, and on EJ all rather recently, I'm going to revisit it and generate that data again. I will be using a slightly different priority than the first time I looked at this issue, however, it should be more comprehensive. I will also be testing a priority that has been floating around on the boards as "superior" despite tangible proof.
It will take time to rewrite the priority (those of you who have looked at the control I have will understand), and it will take even more time to gather all the necessary simulation data before presentation. That's not even considering holiday distractions.
In the meantime, have patience, and quit this bickering. I will try to have simulations running by Thursday, which means you won't see anything probably until Saturday or Sunday.
Last edited by SSHA778; 2012-12-25 at 01:55 AM.