DPS tested in LFR instead of sims

[Ferrair]

You can queue infinitely to test it over and over so i figure LFR is better than sims because of the raid environment, granted its LFR but you get what i mean.

I used arcano 860/bough 865, Conf/MI/FO/UM/CiS. I used Koralon and the chest Belo'vir, no legendary bracers/ring, no food/flask, only lust.

LFR Nythendra i did 418k after i adjusted my muscle memory to use Scorch at 30%. Pre patch i would be doing about 300k? 320k? I don't remember.

LFR Guarm i did 394k, which is a huge improvement from 7.1.

I feel like we crit more, FBlast was available very often, MI and Combustion is easy to pair since they share the same cooldown. I don't quite like Cinderstorm though, I will try Meteor and update this post.

Rotation I used is the same opener except no RoP so once Combustion wore off it's the usual fireball/fireblast/pyro. I love using MI, had to move a couple times at the start of a fight but since i was using MI, i just moved, no problem.

I'm 877 ilvl btw.

[Rad1um]

Testing in LFR, there are so much more variables than there is in SIM... Whats your ping like, how's your computer performance compared to the next person, RNG?, Did someone force you to move from a good spot , etc etc.

[Charge me Doctor]

LFR is not better than sims. You seriously trying to compare 1 parse against 10k parses that simulation does?

[Ferrair]

mm... ok so i simmed myself and i should do 450k with that build.

Guarm i did 394k. Single target, sort of like patchwerk i guess.

How would you figure out the difference in dps and what you did wrong. Cause if you can teach me to use simc to improve my dps, great. Else it's just showing me numbers that show my theoretical highest dps.

At least in LFR, random crap happens like a bad ground aoe, people forcing me to move in H Odyn, stacking for mind control, ping etc. etc. but at least you get a rough baseline of what you can achieve instead of simc telling you "450k".

That's my logic anyway.

[Rylalai]

But isn't almost every LFR fight just more or less a patchwork fight since most abilities that force you to move or change positioning are just not there or tuned way down so that it's no problem at all to just not move? At least that was my experience while running LFR.

[Emeraldon]

But isn't almost every LFR fight just more or less a patchwork fight since most abilities that force you to move or change positioning are just not there or tuned way down so that it's no problem at all to just not move? At least that was my experience while running LFR.

That in addition to a ton of other factors.

LFR is not a good way to test at all.

[ssviolett]

Even by simming your DPS you achieve nothing. Does it matter if your DPS grew from 300k to 400k when other ranged classes DPS grew from 300k to 600k? It does not matter how high your DPS is when you cannot get a spot in your raid.

[Deleted]

Even by simming your DPS you achieve nothing. Does it matter if your DPS grew from 300k to 400k when other ranged classes DPS grew from 300k to 600k? It does not matter how high your DPS is when you cannot get a spot in your raid.

EXACTLY.

Sure, we doe about 80k more now, but everyone else does at least 150k more.

I did 470k on nythendra and there were 7 others on top with over 600k.

Great buff, yeah.

[genzic8787]

i sim myself on xav in lfr.. i change up trinkets and abilitys and see what felt better and smoother and how it affects my dps for when i raid with guild.. took me 4 diff setting but i am now happy with my toons set up (lost alot of crit this patch but im told its ok???) numbers r stronger then pre patch though so thats positive..

[D3athsting]

895 haste/mastery chest should be better than 880 xavius chest right?

[Icoblablubb]

EXACTLY.

Sure, we doe about 80k more now, but everyone else does at least 150k more.

I did 470k on nythendra and there were 7 others on top with over 600k.

Great buff, yeah.

Mages keep crying while WW monk got like 40-50k be happy about 80k dude

[netherflame]

LFR is not better than sims. You seriously trying to compare 1 parse against 10k parses that simulation does?

No but the point is that you are actually playing the character and having to do this not just have a computer stand there and roll through everything with perfect precision.

[Sliske]

SimC interations - 10,000 vs training dummy (patchwerk)

LFR testing variables:

- Did you lag.
- Did you space out for a second bored from your 10th LFR that day.
- Did the shit tank lose aggro.
- Did your procs just be shit for that pull.
- Did you have 4 other insane dps that crushed the boss giving you 1 major cooldown instead of 2 or 3 and a fight time of 3:20 or whatever
- Did you have no dps at all, and you did 80% of the dps
- Did you die from aggro
- Did half the dps die to nothing


LFR is the worst place to test DPS by far. I'd actually take training dummy dps over that.

Accurately testing your dps is much, much more about other players than your play. Playing your spec perfectly is easy. Convincing everyone else to play perfectly is not. If not everyone is playing perfectly, then your raids dps potential is not met. If your raids dps potential is met, your cooldown timers and proc numbers are off. If those are off, your dps is inaccurate and meaningless.

- - - Updated - - -

No but the point is that you are actually playing the character and having to do this not just have a computer stand there and roll through everything with perfect precision.

Its 10,000 iterations with an average, dude.

[netherflame]

SimC interations - 10,000 vs training dummy (patchwerk)

LFR testing variables:

- Did you lag.
- Did you space out for a second bored from your 10th LFR that day.
- Did the shit tank lose aggro.
- Did your procs just be shit for that pull.
- Did you have 4 other insane dps that crushed the boss giving you 1 major cooldown instead of 2 or 3 and a fight time of 3:20 or whatever
- Did you have no dps at all, and you did 80% of the dps
- Did you die from aggro
- Did half the dps die to nothing


LFR is the worst place to test DPS by far. I'd actually take training dummy dps over that.

Accurately testing your dps is much, much more about other players than your play. Playing your spec perfectly is easy. Convincing everyone else to play perfectly is not. If not everyone is playing perfectly, then your raids dps potential is not met. If your raids dps potential is met, your cooldown timers and proc numbers are off. If those are off, your dps is inaccurate and meaningless.

- - - Updated - - -



Its 10,000 iterations with an average, dude.

All its doing is checking the math and standard deviations not checking how you are as a player when you have to deal with things, again as a player.

[D3athsting]

I expect a nerfbat for many classes next week (or even before next reset)

[Sliske]

All its doing is checking the math and standard deviations not checking how you are as a player when you have to deal with things, again as a player.

If its a patchwerk encounter, there is no 'things' to deal with.

There really are only 2 kinds of simcraft parses that are useful - patchwerk and perma cleave (something like zandalari council in TOT).

The reason you run simc is for statweights and brute forcing the answer of what RPPM talent or trinket is better.

And yes, of coursee its 'checking the math and standard deviations' - thats the whole point! If you have 3 trinkets all RPPM, you wnat to know which one is best, so you run 10k iterations. Likewise, if you want to know which talent is best, you run 10k iterations.

Then, you take all of this compiled information (your stat weights, your optimal patchwerk talents and gear, your optimal cleave talents and gear) and add something called NUANCE. With the power of nuance, you can remember shit like "hmm, well, Iron Reaver is kind of patchwerk, but theres also a bit of cleave, and some down time.. so maybe X is better here than not"

And thus you have rank 1 parses

testing dps in LFR like come on bro. Have you ever parsed or gotten rank 1 parse? You will know then that the by far biggest contributing factor to your dps (other than whether you're shit or not) is fight length. LFR has no standard fight length.

Your guild might be killing The Butcher in 3:24 every week which as a fury warrior would be disgusting back in WoD, or it might be killing it in 2:40 which would be terrible for a fury warrior parsing. The difference 20 seconds of fight length makes is rank 1 100th percentile or rank 150 95th percentile

[Bombercloner]

For me, the benefit of SimC is knowing which combination of gear I have yields the highest theoretical dps.

[netherflame]

If its a patchwerk encounter, there is no 'things' to deal with.

There really are only 2 kinds of simcraft parses that are useful - patchwerk and perma cleave (something like zandalari council in TOT).

The reason you run simc is for statweights and brute forcing the answer of what RPPM talent or trinket is better.

And yes, of coursee its 'checking the math and standard deviations' - thats the whole point! If you have 3 trinkets all RPPM, you wnat to know which one is best, so you run 10k iterations. Likewise, if you want to know which talent is best, you run 10k iterations.

Then, you take all of this compiled information (your stat weights, your optimal patchwerk talents and gear, your optimal cleave talents and gear) and add something called NUANCE. With the power of nuance, you can remember shit like "hmm, well, Iron Reaver is kind of patchwerk, but theres also a bit of cleave, and some down time.. so maybe X is better here than not"

And thus you have rank 1 parses

testing dps in LFR like come on bro. Have you ever parsed or gotten rank 1 parse? You will know then that the by far biggest contributing factor to your dps (other than whether you're shit or not) is fight length. LFR has no standard fight length.

Your guild might be killing The Butcher in 3:24 every week which as a fury warrior would be disgusting back in WoD, or it might be killing it in 2:40 which would be terrible for a fury warrior parsing. The difference 20 seconds of fight length makes is rank 1 100th percentile or rank 150 95th percentile

Not really arguing with that side of the point, all I'm saying is that everyone focuses on parses and these patchwerk style encounters. It serves exactly the point that you said above, but we are still set with the problem of what happens when you aren't on a patchwerk fight. Actually dealing with mechanics is where we really need to look at parses because, although it changes from player to player and fight to fight, we are able to see how classes perform when other things besides pure math are taken into account.

For example, enhancement shaman parses in the top 1/3rd on simcraft but based on yesterdays warcraftlog stats its near the bottom of the bottom 1/3rd.

[Sliske]

Not really arguing with that side of the point, all I'm saying is that everyone focuses on parses and these patchwerk style encounters. It serves exactly the point that you said above, but we are still set with the problem of what happens when you aren't on a patchwerk fight. Actually dealing with mechanics is where we really need to look at parses because, although it changes from player to player and fight to fight, we are able to see how classes perform when other things besides pure math are taken into account.

For example, enhancement shaman parses in the top 1/3rd on simcraft but based on yesterdays warcraftlog stats its near the bottom of the bottom 1/3rd.

First off, again, if it is a patchwerk fight (which we occasionally have) then there ARE NO STRATS

Secondly, even if there are strats, they could be a trivial problem at best thanks to shit like ice flows or shimmer.

Third, even if the fight isn't patchwerk, it will inevitably become patchwerk thanks to gear. Think blackfuse. Think Brackenspore.

If the fight is heavy movement and heavy swapping and just generally fucking with your spec, then simc won't help, no. But neither will LFR due to unrealistic timers. LFR won't kill adds as fast as your guild etc

now if you're talking about progression, thats even easier. You will be wiping on bosses that are reasonably difficulty 50-100 times. You'll know the times like the back of your hand. 37-45 seconds roughly between this phase and that phase, ok what spell does that line up with (and so on)

SimC is merely the template or the starting point. After that your dps is about your fight length and your knowledge of it.

LFR is just poisoning the well against that.

[PinDrop]

SimC actually has a setting that can be tweaked to reflect "player skill", in a sense. Based in it's setting, lines in the APL will be skipped over at random, simulating someone missing a proc, or forgetting to refresh a buff, etc. Those are all taken into account in those 10,000 simulations as well. If you are unable to reach the theoretical numbers that sims give you, lower that setting until it's more in line with the numbers you actually pull, and your stat weights will be even more tailored to your specific needs.

It also has light/heavy movement options, adding periods in the sim where only instants can be used, for when you're getting out of the fire, etc.

If you are just loading your character into SimC and hitting "simulate", you are doing it wrong. The stat weights will still help, but they wont be accurate 'for you'.

In a perfect world, every single fight would be programmed into SimC, with timers for movement phases and untargetable/out of range boss periods, etc. for every fight. But that requires a ton of work.

It's a tool to help you understand your class more and what stats benefit your dps the most, not give you final answers. It's up to you to put it into practice in a real raid setting.