[Balance] Mists of Pandaria 5.2, Guide and Discussion [5.3 updates] - Page 10

Slippykins · 10/16/12, 3:08 PM

Originally Posted by Narwhal6

Can someone go over what the posted on the previous page was saying about Moonkins and Starfall use? I had no idea what he was trying to say.

Essentially all he said was: when you have around 30 seconds left on your cooldowns (read: Incarnation), and you're entering a Lunar eclipse so your starfall CD is reset, to optimise DPS you would hold onto that cooldown instead of using it during that eclipse, go through your cycle, and then pop Starfall one cast from the next Lunar (your cooldowns should be ready). Then you pop Incarnation, Starfall's still going and getting the extra CD damage, then you pop your next one as you would when you start the fight (triple Starfall). It's just delaying casting Starfall to do the beginning-of-fight burst again when cooldowns are refreshing.

Stommped · 10/17/12, 12:38 PM

Am I the only one who feels like Incarnation should be off the GCD? It works like a shapeshift which is why it is, but is essentially just a DPS cool down, which normally don't take a GCD to activate.

Ois� n · 10/17/12, 2:13 PM

With the now obvious nerf to Mooncat (for me, the most fun part of our rotation, dissapointed), as far as Tier 6 goes, where do we find ourselves between HotW and NV? Isit still a dps increase to switch up into Mooncat rotation (for me it was Rake, Mangle to 5, FB, switch to Moonkin (will be in solar) MF and SF, Instant SS if proc and then repeat). Or, isit now better to go NV?

Looking at HotW, if the static 6% Int is still effective and > NV for constant DPS, is there something else we can do with the CD, rotation wise? Maybe just dropping in Cat to Rake?

Also with DoC and 4t14 levels, are we talking HT before every Eclipsed DoT and only refresh DoT for the Eclipse your currently in(Except during Incarnation)? during that I assume it would be something like (going from Opening):-

HT before pull>SFall>Wrath>Incarnation>SF and MF>SF+SS to 0 energy>HT>CA>MF>SF+SS>MF again before CA+Inc wears off (meaning it will still be buffed from the second HT before CA).

After this we just HT before every other Eclipsed DoT (as the buff should last us through the Eclipses so our SF in solar and MF in Lunar gets buffed through one HT). Is that the right way to do?

♦ Tecton · 10/17/12, 8:26 PM

Originally Posted by Stommped

Am I the only one who feels like Incarnation should be off the GCD? It works like a shapeshift which is why it is, but is essentially just a DPS cool down, which normally don't take a GCD to activate.

It really irritates me too, especially since our other two cooldowns are off the GCD. Just feels off, like an interruption.

Zantaz · 10/19/12, 4:08 AM

To Oisan: I tried HotW for part of our raid yesterday and the numbers I could get with it were vastly inferior to NV plus my other c/ds. I tried switching to Moonkin during low Energy and staying in Cat form the whole time both but, try as I might, I couldn't beat my NV rotation. It's always possible that I'm still not using HotW optimally but, for now, I'm getting much better dps with NV and for a lot less effort.

I know getting the 2-set bonus will have skewed this a bit as well but I don't think by quite so much as I was seeing last night.

Karepow · 10/19/12, 7:28 PM

I don't know about anyone else here but I feel HotW is stronger in general than NV. Not for all fights but for the majority.

Ois� n · 10/20/12, 6:36 AM

When I used HotW before the nerf I could get it constantly higher than NV, even on elegon just by switching around the way I started the fight (started in Solar so I would have a buffed SF during my Mooncat phase) and it always worked well, the rest of the bosses, I would do a set opening and use Mooncat when I hit Solar. Was very good in Will where I could also get Opportunity Strikes during this phase (dancing while maintaining Mooncat rotation and switching may not of been the easiest but the DPS showed for itself).

Now however, I can get it on par sometimes to NV but it's just not worth the effort when we have a Click and Forget CD. Doing some testing on DoC as well, I get more DPS through only using a 1 Eclipsed DoT rotation even without 4t14 set bonus (can't get numbers to beat NV though). Anyone else finding this?

Tuscarora · 10/20/12, 10:00 AM

Originally Posted by Slippykins

What does this mean overall? Essentially, all I'm saying is that the two meta gems are equivalent. Just as we reforge in and out of hit/crit/haste, making them all equal in "weight", so too can we gem in and out of int/crit. Therefore, we have a new "cap" - namely, the point at which Int = 2x Crit. This occurs (assuming perfect haste/hit cap) at roughly 16161 Int and 7500 Crit, where int = 4.5 and crit = 2.25.

Ultimately there is not a single point where the marginal value of Int rating equals double the marginal value of crit rating but a set of points where that condition holds. For each value of remaining stats (since mastery and haste likely influence the position/shape of that line) the picture looks like this:

In theory, that frontier can be computed from the WC model (as a function of mastery and haste) but it's probably more practical to just compute the weights for your situation and reforge/regem accordingly.

Originally Posted by Slippykins

You may think that these tiny numbers don't mean much, but they do shine a lot of light on how dynamic our stats are. There are two things that I find most interesting: 1. haste drops significantly below mastery once at the BiS stat levels, and 2. crit's weight increases as more crit is added. Now, the latter part is purely theoretical, as adding additional crit means sacrificing the other stats, therefore most likely reducing the value of crit anyway, but I thought it was just interesting to note. The former part is irrelevant, as we cut back on excess haste and put it into crit, which is above mastery anyway.

Somewhat off-topic: from a theoretical point of view, the phenomenon is not strange. In fact, you would expect marginal values (or "stat weights", but I don't like that term) of stats to routinely equalize under optimal gearing/reforging/gemming. Let $f(x_1, x_2, \dots, x_n)$ be the DPS function (where x_i is the amount of rating of stat number i) and let $g(x_1, x_2, \dots, x_n) = c$ be the reforging/gemming constraint. If you could reforge any amount (including to negative amounts) of a stat, optimal gearing satisfies:

$\nabla f + \lambda \nabla g = 0$ (1)

$\frac{\partial f}{\partial x_i} = -\lambda \frac{\partial g}{\partial x_i}$

For secondary stats (and Int at half value), $\frac{\partial g}{\partial x_i} = 1$ so all stat weights / marginal values equal the Lagrange multiplier λ with negative sign and thus are all equal.

The reason this doesn't usually happen in practice is that reforging is limited, so the global maximum of the DPS function does not meet the equations above but lies instead in the frontier of the constraint and thus we can talk about "absolute stat priorities" (a paradigm that only makes sense when the global maximum is not reachable with reforging)

* Note that all local extrema satisfy (1) (they are stationary points) but not all stationary points of f (the ones that satisfy (1)) are local extrema (saddle points and all of that)

Crucial288 · 10/20/12, 12:25 PM

I dabbled today in challenge modes for the first time.

First thing I noticed, was that my haste value dropped a bunch due to gear normalization, meaning I lost the haste breakpoint.

My question is in two parts:
1) How much haste do I need without normalization, to still hit the haste breakpoint after entering a CM?
2) Is it even worth going for the haste BP for them? There is certainly a lot of multidotting and quite minimal boss fighting - would it be smarter to just go full crit for them? Would this mean moonkins need to reforge all their gear when going from a raid to a CM every time?

• Hamlet · 10/20/12, 1:00 PM

We don't know the formulas yet, as far as I know. I've tried to test a couple times, but the behavior was confusing and possibly buggy--I'm going to test more when I have a few more >463 items to examine (which may be hard due to not raiding). I will say--I think socketed items may still be disproportionately good, so you have some 463 and >463 items in a slot, try using the socketed one for now (can always confirm just by actually adding up the stats and comparing of course).

Right now I think the easiest way to deal with haste breakpoints is to walk right into a challenge zone (you can do this solo and stand at the entrance) and look at your stats. I'd swap gear on/off a couple times just to watch for potential inconsistent/buggy behavior and make sure you're still over.

Slippykins · 10/20/12, 4:34 PM

Originally Posted by Tuscarora

Ultimately there is not a single point where the marginal value of Int rating equals double the marginal value of crit rating but a set of points where that condition holds. For each value of remaining stats (since mastery and haste likely influence the position/shape of that line) the picture looks like this:

In theory, that frontier can be computed from the WC model (as a function of mastery and haste) but it's probably more practical to just compute the weights for your situation and reforge/regem accordingly.

Somewhat off-topic: from a theoretical point of view, the phenomenon is not strange. In fact, you would expect marginal values (or "stat weights", but I don't like that term) of stats to routinely equalize under optimal gearing/reforging/gemming. Let $f(x_1, x_2, \dots, x_n)$ be the DPS function (where x_i is the amount of rating of stat number i) and let $g(x_1, x_2, \dots, x_n) = c$ be the reforging/gemming constraint. If you could reforge any amount (including to negative amounts) of a stat, optimal gearing satisfies:

$\nabla f + \lambda \nabla g = 0$ (1)

$\frac{\partial f}{\partial x_i} = -\lambda \frac{\partial g}{\partial x_i}$

For secondary stats (and Int at half value), $\frac{\partial g}{\partial x_i} = 1$ so all stat weights / marginal values equal the Langrange multiplier λ with negative sign and thus are all equal.

The reason this doesn't usually happen in practice is that reforging is limited, so the global maximum of the DPS function does not meet the equations above but lies instead in the frontier of the constraint and thus we can talk of "absolute stat priorities" (a paradigm that only makes sense when the global maximum is not reachable with reforging)

* Note that all local extrema satisfy (1) (they are stationary points) but not all stationary points of f (the ones that satisfy (1)) are local extrema (saddles points and all of that)

I love you. I was just guessing at what was happening, but you math'd it and that makes me all happy inside.

And the phenomenon was strange to me because I had no theoretical data to back up it.

Off-topic: It looks like you're mostly talking in multi-variable calculus if I'm not mistaken, and unfortunately I wasn't able to take that class last semester (international students get put on the bottom of the list), so instead this semester I'm taking DE's/LA. I've seen grad(F) and saddle points thrown around a bit, but nothing on Langrange.

lazybear · 10/20/12, 5:24 PM

Got a couple questions does anyone know if [Light of the Cosmos] is better than [Blossom of Pure Snow] i would assume Blossom is especially if we line up the crit for reapplication of dots/eclipse. Next is Jade Spirit actually better than windsong given than atm we seem to be able to have 2-3 of the buffs up at the same time (i.e 1500 mastery + haste). As the enchants are just becoming available be nice to know if its actually worth it.
Cheers

Karepow · 10/21/12, 8:31 PM

Originally Posted by lazybear

Got a couple questions does anyone know if [Light of the Cosmos] is better than [Blossom of Pure Snow] i would assume Blossom is especially if we line up the crit for reapplication of dots/eclipse. Next is Jade Spirit actually better than windsong given than atm we seem to be able to have 2-3 of the buffs up at the same time (i.e 1500 mastery + haste). As the enchants are just becoming available be nice to know if its actually worth it.
Cheers

I'd actually say that Windsong is more powerful at the moment since having all 3 instances of it active at the same time works out to be about 1.5x stronger than the 1650 Int proc from Jade Spirit (because of stat weights). I've not done the exact maths behind it but that's just an estimation.

Also I think critical strike updates dynamically. Meaning you don't have to re-apply DoTs when you activate the trinket. Would appreciate it if anyone could confirm this.

Aarondf2003 · 10/21/12, 8:36 PM

I had the Jade Spirit enchant for all 6 bosses this week and had between 20% and 23.5% uptime on every fight (WoL).

While Windsong provides secondary stats as opposed to Intellect, my kill attempt logs show a wide range of uptimes but definitely much stronger percents (in the range of 50-86% uptime when combining Haste, Crit and Mastery procs).

On paper, they're moderately close. But, we have to take into account some +1500 Mastery procs are almost entirely wasted if you just left an eclipse. Especially if you have to move, prolonging your out-of-eclipse time.

I think the real deciding factor is how capable you are of capitalizing on multiple, simultaneous Windsong procs. Proc addons and/or high awareness will really allow you to take advantage of the unique effect of multiple Windsongs.

• Hamlet · 10/22/12, 1:02 AM

That issue with the mastery proc is illusory. We had this same discussion all the time with Theralion's Mirror--mastery is below average outside Eclipse and above average inside Eclipse. On average, proc-based mastery gives its average value (oddly enough).

Utilizing procs well by paying attention to them for your DoT refreshes is neither here nor there; you can do that with either enchant. Also none of these procs are really that huge like trinket procs/potions.