[Balance] Mists of Pandaria 5.2, Guide and Discussion [5.3 updates] - Page 26

The Grog · 02/23/13, 9:22 AM

Originally Posted by Hamlet

Yeah, I forgot to mention the big bump to the healing utility compared to 5.1 (I've mentioned it other times). Only issue is that's it's hard to evaluate for a reason I point out a lot in Resto discussions--the value of healing is heavily discounted when it comes at times that are either random or fixed (i.e. not used intelligently for healing purposes). If it comes at an easy time, even if it's effective, it just amounts to saving them a little mana. There will definitely be times when the hard healing and the hard DPS part of a fight are at the same time though (maybe people are used to watching for this already after using NV all through the first tier).

Honestly, there are times I wish NV was healing only and had no effect on your dps like Ancestral Guidance does. Too many fights where AG saves the raid, but I had to blow NV some time ago because that's when Incarnation/CA or 'zerk came up and I can't wait because of heroism/burn phase/blocking a CD cycle.

Quite frankly, the healing feels like a pvp talent and the change feels like a pvp change.

Qaajn · 02/24/13, 11:57 AM

Originally Posted by Hamlet

UVLS is going to grow between tiers. Makes sense: PTR Class and Set Bonus Issues, Part III - Forums - World of Warcraft

Otherwise the situation would be what I pointed out a little ways back: the things is too good at LFR tier since you get the same proc.

I've been pondering some about this too, but and at first I thought it was the right solution. But is it really now? Yes, it does mean that there will be a big increase to upgrade to the next tier one... But think of it from a LFR point of view. All your stats will be lower, so you will deal less DPS. On top of that, the procrate will be lower too, resulting in the proc effect of the trinket being relatively much worse for lower gear levels. Consider also that if you have access to heroic gear, your haste will likely be much higher (next breakpoint?) which would further increase the value of the trinket.

While a chance depending on the quality of the trinket makes it less viable to "sit" on LFR trinkets and spread normal/heroic trinkets to those that don't have any kind, my concern is that it will seriously devaluate it for lower item levels, likely to the point that you shouldn't use it (negative feedback - lower damage during proc and lower procrate).

Tuscarora · 02/24/13, 12:25 PM

Originally Posted by Qaajn

All your stats will be lower, so you will deal less DPS. On top of that, the procrate will be lower too, resulting in the proc being relatively much worse for lower gear levels.

Your point applies to all pieces of gear, since usually every proc or stat gives you more DPS if you have a greater amount of the remaining stats. Whether UVLS scales harder with ilvl compared to other trinkets is a quantitative question, not a qualitative one.

Qaajn · 02/24/13, 1:03 PM

Originally Posted by Tuscarora

Your point applies to all pieces of gear, since usually every proc or stat gives you more DPS if you have a greater amount of the remaining stats. Whether UVLS scales harder with ilvl compared to other trinkets is a quantitative question, not a qualitative one.

It is true that all stats scale with every other stat. But the difference here is that the proc itself doesn't give you any extra stats, the potency of the proc scale only with your other gear. The increase in DPS would be relatively the same with 463 item level as it would for 541, simplified but should not be far of.

While intellect and crit procs also are better the higher your other stats are, it surely can't be as much as the 36% increase in prochance for the trinket (LFR - Heroic). It may work out about the same increase if you consider the same gear but swap trinket, but I highly doubt a trinket proc with LFR gear will result in a remotely similar relative DPS increase as if you have Heroic gear. For normal trinkets the relative increase should be about the same.

Edit: I apologise for not being able to supply any numbers on it. Sadly it's just what logic would tell me.

• Hamlet · 02/24/13, 3:24 PM

The scaling of the proc rate in that post GC made is the same as the normal stat growth of all items, around 13% per 13 ilvls.

The proc is in effect equal to a bunch of added crit (for purposes of this discussion). Just like any other bunch of added crit, its value does increase with your other stats (as all stats), but you also want the amount of stat you're adding to increase with the ilvl of the item. Most trinkets do this by increasing the size of the proc (like all of our Int trinkets). But UVLS and Rune can't do that, so they have to boost the proc by increasing proc rate.

The only unique issue with UVLS as far as basic scaling is that the proc is worth slightly less as you get more crit, but this is probably too small an effect to matter. And in fact, since added crit increases the expected duration of the super-DoTs anyway (via Starsurge crits), it may even get disproportionately better with crit.

Qaajn · 02/24/13, 3:40 PM

Originally Posted by Hamlet

The proc is in effect equal to a bunch of added crit (for purposes of this discussion). Just like any other bunch of added crit, its value does increase with your other stats (as all stats), but you also want the amount of stat you're adding to increase with the ilvl of the item. Most trinkets do this by increasing the size of the proc (like all of our Int trinkets). But UVLS and Rune can't do that, so they have to boost the proc by increasing proc rate.

You've got more experience with this, so I guess you're right. I would still find it interesting to see a comparison of 4 sims: Two with BiS gear, where you use UULS on one of them, the third best trinket for the other sim. Then the same with the LFR gear. Then we can see if the trinket change DPS (relative) as much with the same set of gear, and if upgrading to heroic with and without is the same relative increase. Would do it myself, if the gear was in wrathcalcs

• Hamlet · 02/24/13, 4:41 PM

As far I know it should be there except for the changing proc rates, which I haven't updated yet. Not too hard to hack in different proc rates though if you know where to look (Basic Calcs tab where all the trinket PPM stuff is).

Qaajn · 02/24/13, 8:13 PM

Originally Posted by Hamlet

As far I know it should be there except for the changing proc rates, which I haven't updated yet. Not too hard to hack in different proc rates though if you know where to look (Basic Calcs tab where all the trinket PPM stuff is).

Yes, found it now. For adjusting the procrate I adjusted the UVLS in the Basic Calcs formula (60/(0.5*1.0674*G37) for heroic). For gear I used the heroic gear that was set up. Then I changed them to the LFR version. The thunderforged I swapped to other pieces, with focus on haste/crit/spirit/hit stats

Cloak: Deadly Glare Cape
Chest: Robes of the Haunted Forest
Belt: Abandoned Zandalari Moonstrap
Boots: Sandals of the Starving Eye
Ring1: Dudumu's Captive Eyeball

When I didn't use UVLS I changed it to Cha-Ye's Essence of Brilliance, as it seemed to give the best DPS. I reforged and regemmed for reaching hitcap in LFR gear. +3 ticks MF/SF was still kept

Results:

LFR with UVLS: 118238 DPS
LFR with CYEB: 114807 DPS
Difference: 2.99%

HC with UVLS: 187347 DPS
HC with CYEB: 181446 DPS
Difference: 3.25%

UVLS difference: 58.45%
CYEB difference: 58.04%

While there is a difference, it seem to be smaller then I expected.

Slippykins · 02/24/13, 10:51 PM

Thanks again for going through and putting up a target module Hamlet, once the small issues you said are fixed it should really help with modelling most of the fights from here on. One question - do you think the current model is accurate enough to base some assumptions on? For example:

Trinket	1 target	2 targets	3 targets
Cha-Ye's (crit)	184552.01	240975.92	278500.989
Hydra (crit)	184843.93	239917.79	276024.38
Cha-Ye's (hasteBP)	187146.34	247655.55	286111.73
Hydra (hasteBP)	187272.58	248100.84	286883.5

I wanted to tabulate these for two reasons: 1. to find which trinket of the two is the better option, and 2. find whether or not the haste breakpoint is worth achieving. With Hydra, you end up sacrificing ~2000 crit for haste to reach the breakpoint, whereas with Cha-Ye's you sacrifice ~3000.

If your model is accurate enough, it shows that crit gains in strength as more targets are added, but so too does the haste breakpoint. Comparing the crit builds with different trinkets, the difference between the two builds is -0.16%, 0.44% and 0.90% for 1-, 2- and 3-target fights, respectively. This shows one of two options (or a combo): that Cha-Ye's grows in strength with more targets moreso than Hydra, or the additional crit with Cha-Ye's is more favourable for multi-targets (the more likely situation).

Conversely, if we are to compare the hasteBP build vs. the crit build with Cha-Ye's, we see differences of 1.41%, 2.77% and 2.73%, depending on targets. I have a feeling that the 3-target scenario is a drop from 2-target because of the issues you were talking about with the modelling earlier. However, these differences are much more significant than the sub-1% changes we saw changing trinkets.

This leads me to believe that, regardless of how many targets there are, we can't underestimate how important reaching a breakpoint is. Looking at marginal stat changes shows crit pull ahead, per point, than haste once a second target is introduced. Even though 1 point of crit may mean more than a point of haste pre-breakpoint, every target also increases how much the haste breakpoint is worth, and one would think this pattern extends past 2 or 3 targets.

And as Hamlet said - we're reaching pretty high crit values, and not every SS proc can be used. As such, the value of crit itself may be overestimated.

Anyway, just food for thought. Obviously we can't make many conclusions at this time, but if Hamlet's preliminary look at multi-target modelling is somewhat accurate, it does show noticeable trends with the increase of each point of crit, the haste breakpoint, and trinket differences.

Erdluf · 02/25/13, 12:01 AM

Originally Posted by Slippykins

... we're reaching pretty high crit values, and not every SS proc can be used. As such, the value of crit itself may be overestimated ...

Of course if you ever clip a dot (trinket proc, movement, add dies, or you just make a mistake during multi-target dotting), you've overestimated the benefit of reaching a haste breakpoint.

Still 2% is a fairly large margin, and Haste has other benefits not modeled by WC (for instance, Haste usually lowers the penalty associated with moving out of the fire). If those numbers don't change, Haste seems like a good choice.

The margin between the trinkets seems to be too small to trust.

Slippykins · 02/25/13, 12:20 AM

Originally Posted by Erdluf

Of course if you ever clip a dot (trinket proc, movement, add dies, or you just make a mistake during multi-target dotting), you've overestimated the benefit of reaching a haste breakpoint.

Still 2% is a fairly large margin, and Haste has other benefits not modeled by WC (for instance, Haste usually lowers the penalty associated with moving out of the fire). If those numbers don't change, Haste seems like a good choice.

The margin between the trinkets seems to be too small to trust.

I agree with the margin on the trinkets. Too hard to tell, given there's some inaccuracies with the model. And yeah, clipping dots with >1 tick remaining will overestimate the breakpoint, but as you approach an optimal rotation, you'll asymptotically approach what the breakpoint should be worth. Then again, it is important to note the practical implications from what is theoretically best.

I'm hoping that, once we get all the gear into WrathCalcs and it's more appropriate for the 5.2 release, there's either a larger margin between the two options, or they're basically the same (within .5%-1% margin). Either way, it should lead us to having the option to choose without repercussions, or one option being the best of the two.

fountaiin · 02/27/13, 1:20 PM

With regard to the two and four piece, I was wondering how accurate the WrathCalcs model was. I was also wondering, if the models aren't working, if anyone had any close numbers.

Right now with 513 full bis gear, it has the two piece at 4500 DPS, and the four piece at 7100.

If I sub in UVLS, the two piece jumps up to 6200. Obviously this trinket is going to make the two-piece value weird, because of the extra extensions of 100% crit dots, etc.

I've heard that with napkin math, the four piece if around 2.5%-3% damage. I've been using it on every PTR fight and it seems to be pretty strong on most of these fights. I was just hoping some math-lords could give me some reasonable numbers for the sake of comparison with other vanq users.

• Hamlet · 02/27/13, 1:44 PM

WC should be pretty good here. 2-piece is easy to compute. The strong interaction with UVLS that you saw in the sheet should be quite real. 4-piece is overestimated in the sheet, but by very little in a stationary rotation where you have very high NG uptime. 4-piece will lose a lot with heavy movement, but should still be quite a strong bonus.

Duba · 03/04/13, 4:03 PM

Sinister Primal Diamond – 1.18 base RealPPM on damage/absorb of harmful or periodic spell. No ICD.
- That base proc rate is multiplied by an additional coefficient:

Moonkin: 6.192

Wait what? How big of a boost is the meta now ?

Xavy · 03/04/13, 4:16 PM

Dont forget

- Unerring Vision of Lei-Shen – 0.50 RealPPM on damage/absorb from harmful spells and periodic spells. >No ICD<. Proc rate multiplied by 0.00 for non-caster specs. >Proc rate multiplied by 0.5 for Balance Druids<. Proc rate multiplied by 0.25 against player controlled units. Proc rate multiplied by 1/(1.15^((528-ItemLevel)/15) for the various Item Level versions of the trinket.

So.. What that means?