Haste vs. AP gem question (not what you think!)

[Dbol]

I've played enhance since Sunwell and before this played as resto since vanilla. I have never run into damage problems and even in TOC/TOGC I was always in the top 5 dps for our guild. Fast forward to ICC when I finally get enough frost badges to get the trinket and I change out all of my gems to haste. Reason I did this is becuase it simmed for much higher dps than my AP gems did. Since then, even though I've upgraded almost all of my gear, my damage is pretty low. I know sims are just a tool and represent your potential more than your actual performance. But my gear is simming out to about 9800 and on miminal movement fights, I'm hitting about 8600.

I use S&A, always have, so I don't think it's a rotation thing. The one thing I'm thinking is that since I've switched from AP to Haste my dps isn't the same. I'm thinking that since I've gone from about 200ish haste pre-ICC raids to about 700 haste now (even more with procs) I think with the tighter rotation I'm just not able to keep up.

I realize I'm not giving all the variables here to properly diagnose my issue but I've always been able to fix any dps issues in the past and I'm wondering if switching back to full AP or even haste/ap gems could possibly help. Rawr shows a dps loss but I'm wondering if actually game play would show otherwise because of my seemingly slower reaction time.

[Butu]

Currently this is pretty much impossible to test with Rawr/EnhSim since neither of them support reaction time. If you set your latency to your real latency + your assumed reaction time (in your case probably around 300-500ms) you might get some results, but as the sim thinks its actual latency the results will be skewed.

There is one simulator though which supports reaction times. Team Robot Simulator and Gear Comparison Tool . You could try to mess around with the reaction time setting + changing your gemming to see if you get any interesting results. The sim is by far from finished yet though, so use it with caution.

If all else fails you have no option but to do the thing you should never do to evaluate your dps, test full ap and haste in real boss fights, but since you already (at least subconsiously) think which one is better for you you might end up underperforming when using haste, and it would take several tries before you could make any real assumptions.

[Barawr]

Currently this is pretty much impossible to test with Rawr/EnhSim since neither of them support reaction time. If you set your latency to your real latency + your assumed reaction time (in your case probably around 300-500ms) you might get some results, but as the sim thinks its actual latency the results will be skewed.

I think you are a bit off here. There does not need to be a separate field for both latency and reaction time since they are one in the same. They both affect your dps in the same manner just one is server related and the other is brain-to-finger related. If you want to get an idea of what your actual reaction time you can search for several online tests which will give a good approximation. Do the test several times for consistency then add that value to the average of what your latency in game is. This alone can change the sim results drastically if the values you currently use are too low.

P.S. I checked out the site you linked and it seems pretty sweet. It is giving me numbers consistent with what I see in EnhSim as well.

[Butu]

I think you are a bit off here. There does not need to be a separate field for both latency and reaction time since they are one in the same. They both affect your dps in the same manner just one is server related and the other is brain-to-finger related. If you want to get an idea of what your actual reaction time you can search for several online tests which will give a good approximation. Do the test several times for consistency then add that value to the average of what your latency in game is. This alone can change the sim results drastically if the values you currently use are too low.

P.S. I checked out the site you linked and it seems pretty sweet. It is giving me numbers consistent with what I see in EnhSim as well.

There definately is a difference.

Example. Latency 200ms reaction time 300ms
You get MW_5. Server sends you information which you get at T=200ms. Then you need 300ms to react, so you use MW_5 at T=500ms. 200ms for the data to get back to the server, T=700ms. Now If you had set your latency to 500ms the sim would think the server got the data at T=1000ms, because 500ms for the data to get to you, 500ms to send it back. T = 1000ms. There is also a long post about how the sim is overvaluating MW_5 because there is no difference in reaction time and latency in the sim, kudos to Rouncer for the research he has done on the subject http://enhsim.codeplex.com/WorkItem/...orkItemId=6109

Edit: Also sometimes your reaction time for some abilities could be near 0 if you are already mashing the button. If you set your latency to your reaction time + your real latency you'll see the problem that emerges; many abilities gets unnecessary latency.

Edit2: Seems like there is reaction time in Rawr. Never noticed this myself before.

[booyakasha]

with all that delicious haste you have shorter gcd's plus we have more buttons to press now.

I'd suggest mashing buttons harder and earlier to get the full advantage.

[Deleted]

Its an interesting point.

More haste gives reduced GCD.

Server lag increases the time it take for a spell to appear to cast and makes GCD seem smaller

Would be interesting to sim high server lag with haste, to see how much it effects things.

And yeah my server lags big time during raid hours.

[Butu]

Its an interesting point.

More haste gives reduced GCD.

Server lag increases the time it take for a spell to appear to cast and makes GCD seem smaller

Would be interesting to sim high server lag with haste, to see how much it effects things.

And yeah my server lags big time during raid hours.

On the other hand haste is the best stat to get to increase your white attack dps, and since it is unaffected by lag haste it not so bad with a laggy connection as some people think.

[Alexandrra2012]

What's your current priority list you are using?

[Barawr]

There definately is a difference.

Example. Latency 200ms reaction time 300ms
You get MW_5. Server sends you information which you get at T=200ms. Then you need 300ms to react, so you use MW_5 at T=500ms. 200ms for the data to get back to the server, T=700ms. Now If you had set your latency to 500ms the sim would think the server got the data at T=1000ms, because 500ms for the data to get to you, 500ms to send it back. T = 1000ms. There is also a long post about how the sim is overvaluating MW_5 because there is no difference in reaction time and latency in the sim, kudos to Rouncer for the research he has done on the subject http://enhsim.codeplex.com/WorkItem/...orkItemId=6109

Edit: Also sometimes your reaction time for some abilities could be near 0 if you are already mashing the button. If you set your latency to your reaction time + your real latency you'll see the problem that emerges; many abilities gets unnecessary latency.

Edit2: Seems like there is reaction time in Rawr. Never noticed this myself before.

Good call. I should have thought this one out a little more based on my RL area of work ;D. I do remember seeing the options inside RAWR as well, however I'm not sure how they are modeled. Levva may know who was in charge of the code for that particular model.

[Barawr]

What's your current priority list you are using?

This thread has absolutely nothing to do with priority list. Please check the FAQ to find the list.

[detramental]

i think something similar has happened to my dps as well.

the better gear i get, and the simmed dps keeps growing, but my actual dps is lacking a bit. it is true that my overall dps has improved, but not to the extreme that enhsim outputs (granted, that's perfect situation vs. lag/reaction time/movement).

this is one reason i choose to try to balance out my gems, instead of going all +haste. i think it's just something you need to try to juggle, depending on your play style/reaction time.

[LyskaWF]

Currently this is pretty much impossible to test with Rawr/EnhSim since neither of them support reaction time.

What.... yes there is.
You can determine your approximate reaction time using simple tests which I cannot site right now, and add your server latency. There's nothing else to it other than inconsistency, which is why you should err on the higher side.

[Aozora]

It's probably a reaction thing that others are talking about, there's no way your personal reaction time should be at 300ms or so, for example if you play i mage you keep mashing buttons all the time reducing the "player" lag to almost 0. When you use an addon such as S&A it makes you slower to react since you need to wait until using an ability to see which button you should press next, the only thing you should be reacting to as an enhancement shaman should be MWx5 since that is random.


Basically you should try to plan ahead, so you don't have to react to each and every ability.

PS: Problem could be somewhere else aswell, armory link would be helpful

[Barawr]

What.... yes there is.
You can determine your approximate reaction time using simple tests which I cannot site right now, and add your server latency. There's nothing else to it other than inconsistency, which is why you should err on the higher side.

Lyska, check out the EnhSim link he posted where Rouncer goes into a lot of detail on it. It makes sense if you read over that and understand the basics of how server to server messages work. You can add your reaction time into the latency window, but it will not model your reaction time correctly.

[Melonberry]

Its an interesting point.

More haste gives reduced GCD.

Server lag increases the time it take for a spell to appear to cast and makes GCD seem smaller

Would be interesting to sim high server lag with haste, to see how much it effects things.

And yeah my server lags big time during raid hours.

Haste only will reduce the GCD on spells not attacks, small difference but it's there.

[Deleted]

Currently this is pretty much impossible to test with Rawr/EnhSim since neither of them support reaction time.

Wrong. Both support reaction time.

[Deleted]

When you use an addon such as S&A it makes you slower to react since you need to wait until using an ability to see which button you should press next, the only thing you should be reacting to as an enhancement shaman should be MWx5 since that is random.

This misunderstands how ShockAndAwe works and how the you can configure the pre-notification time. If you are waiting until the CD timer is over yes you will lose reactions if you are reacting as soon as the ability is available (by default 0.25 seconds BEFORE the GCD is up) then you get almost zero downtime. If your personal reaction time is higher change the defaults from 0.25s and it will warn you earlier about what comes up next. Be aware however that the higher you set this the more likely you are to encounter "that ability isn't ready yet" messages, so its a trade off. I used to use the bars to assist the sequencing too, I don't any more as I don't raid any more.

[Deleted]

Wrong. Both support reaction time.

you are correct in that both support reaction time, but the way it is interpreted is different.

in Rawr, you have server lag and reaction time. server lag is how long it takes for a new priority to get from the server to your computer. the time calculated is (arrival time - departing time). reaction time is how long it takes your brain to respond to the new priority and press the button. the time calculated is (button pushed time - priority noticed time). in this case, a general timeframe is 60 MS to send the data, 240 MS to process and push the button, and another 60 MS to send the button push back. the total time is 360 MS.


in Enhsim, you have 2 fields that combine to form the Reaction time. in these fields, we need to add the SUM of the 2 values above, and 30 above and below. so we fill in 60 + 240 = 300. we substract 30, 270, we add 30, 330. but in this case, the total time is 300 seconds to, and 300 seconds from, for a total of 600 seconds. in this case, the reaction time from the Rawr field is used in the equation 2 times, but it should only be used once. but because Enhsim adds these 2 values, it can't distinguish the 2 times.

I don't know how this works for DPS calculations, but unless we completely misunderstand the workings of those fields, it should have a negative effect.

[Butu]

Wrong. Both support reaction time.

I already edited my second post way before you posted to correct my error that Rawr wouldn't support reaction time, and as nzall already said EnhSim doesn't do it right as it actually considers the given value to be latency.