Originally Posted by Shha
15% ArP = 184 ArP rating
5% Crit = 228 Crit rating.
Given how average SEP for ArP is 1.1 or so (assuming average ulduar gear and not capped), and crit at 0.85 , 15% ArP is better. But slightly :0
|
I know that the ratings could be compared on SEP basis, and that this is gear dependant. Point stands that 15% ArP (max 7.5% dmg increase) is certainly not clearly better than 5% crit, no matter how you look at it.
And as for SEP in general. I've got the gut feeling that crit as such is undervalued by the spread sheet approach. I just cannot pinpoint it to the cause. But because "gut feelings" are not an acceptable argument on this board, I'd like to explain a bit, so bear with me.
In the "bottom up approach" of most spreadsheets (and simulators btw) one starts with your stats, knowing exactly how they influence your styles and approximate a realistic combat rotation or priority system (including lag, execution phase, bloodlust periods, triket proc behavior etc). If everything goes right (and I am not saying it does not!), at the end one knows exactly what effect a certain stat point increase would have on the total damage. Once this is done, getting to the concept of a comparable metric (like SEP) is trivial.
Conceptually different to this would be the "top down approach". One starts from the end (lets say an evening full of wipes on Mimiron hard mode), analyses carefully the collected data and extrapolates from there what effect an additional 1% crit/hit/ArP od some given value of STR would have had on the outcome. Again, to do this one has to know exactly how combat mechanics work, how styles correlate to the different stats (e.g. DW does not profit from ArP in any way, but does profit from Crit/STR and to a lesser extent hit/haste).
Keep in mind though that while this approach does have its drawbacks like no clear distinction between boss and corollary adds and being more prone to luck (whereas spreadsheets and to a lesser extent simulators work with exact long term averages) it does have its clear advantages too: no need to approximate your play style, no need do model difficult mechanics like stacking trinkets with skill cooldowns in burn phases, execute phase mechanics, rage model etc. All of this is condensed into the final data where you see how much your average execute has hitted for, what the average execute crit was and what portion of your total damage execute has been.
To do this properly you need some time though. And having some discipline in collecting enough viable data is paramount.
To cut a long story short, I have been doint this occasionally and naively and found based on this data crit to be more valuable in comparison to the spreadsheet results. Crits on average and in total effect (i.e. considering Deep Wounds contribution) do up to 3 times the normal damage of the corresponding hit. As such adjusting the data by 1% more crit would have produced more than 1% extra total damage (as long as effective crit rates are not absurdly high). Similiarly I found STR to be somewhat overvalued. That was based on damage style distribution (white, BT, WW, Slam, DW, HEroic ...) of your total damage, average raid buffed AP (including trinket procs) and factoring in latent AP contribution in form of weapon range and style bonuses. As such the effective AP (read: the virtual total AP that is contributing to
all of your total damage) has reached some abnormal values. And STR does add in absolute terms to that, contrary to crit, ArP and the lot which enhance your total damage in a relative way.So there is a dimished effect of STR relative to the other stats when seen in the context of ever growing AP values.
Don't mistake me. I do appreciate spreadsheets and the like. I do use them and I think they are of tremendous help. It's just that I seem to observate some divergent results which I cannot explain when I'm looking on my damage from the other way (top down). In short I do not take spreadsheed results and suggestions for the gospel of truth (and I do not think that the creator wants them to be seen as such).
Thoughts anyone?
I apologize for the lenghty post and the anecdotal vibe. This topic is just somewhat difficult to describe in short terms.