[Resto] Cataclysm (Old/4.1) - Page 88

Erdluf · 06/13/11, 10:10 AM

I'll try some napkin math.

A not-quite-current TreeCalcs (5/11 PTR) says 450 mastery is a 2.5% boost to HPS.

Suppose I'm not really a tank healer, but my "priority list" during some critical period is

1) Maintain Lb (on average I refresh 1.5s early).
2) WG every 7s (this and the following are all on-average)
3) SM every 16s
4) 1 each Rg, HT, and Remove Corruption, every 20s (or any other combination of 11 GCD's/minute)
5) Leftover time goes to Rejuv.

At 2005 Haste, Rejuv is 36.9% of your time. At 2455 Haste it is 39.7% of your time. That is about a 7.7% increase in the number of Rejuv's you cast. If Rj was an "average EH/CT" spell in our rotation, that would be about a 2.8% EHPS increase. However I ignored NG (drops that 2.8% to 2.1%), and Rj has below-average effective-heal per cast-time (Rj has high overheal, and it is much weaker than WG, SM+Efflo and Lb).

Conclusion: At 2005 Haste, given the choice between additional 450 Mastery or 450 additional Haste, take the Mastery for typical raid-healing situations. I'd recommend revisiting that conclusion for any fight that you'd ToL+Lb for the critical phase(s).

Anaram · 06/13/11, 5:55 PM

Originally Posted by Videl

It still stands that a 10.5s lifebloom would be nicer to work with than 9.5s lifebloom. Given reforging there's no reason you would just happen to be close enough to make it worth picking up though. It's hard to believe it could be worth 445 (~3%) mastery.

Ignore, bad math.

It's not a choice between 9.5 and 10.5 seconds. It's a choice between 10.0 and 10.8* seconds when you are dealing with a lifebloom refreshed. When you refresh Lifebloom it's impossible to get a duration under 10 seconds.

* Assuming ~20% haste.

Videl · 06/13/11, 9:10 PM

There's a few reasons why it wouldn't really be 9.5-10.5 seconds. At any rate, .6 seconds is more or less close enough. I had also settled on 2.5% from the mastery after figuring for additive bonuses. I suspect applying average values is just the wrong approach here. Collision avoidance is going to be a significant part of the gain from it which I doubt is modelled there. For example, if you want to assume you're casting WG on cooldown except where it conflicts with your lifebloom refresh, then reducing the number of lifebloom refreshes in the fight by ~6% gets you 2 more WG in a 5 minute fight (replacing rejuvs). More importantly, certain spell combos between refreshes are going to allow you an extra cast while others are not. When you're using a combo that allows an extra cast, the throughput gain will far exceed the mastery. When you're not, there will be only a very tiny gain. In a max hps scenario it looks like you'd probably be able to squeeze in an extra rejuv from the haste. I don't see a way to come up with a real answer short of some kind of simulation where a bot druid attempts to heal a fight with a damage profile that changes through the fight in a realistic way. The numbers at least seem to suggest that it is a possibility and something each person could try out to see if it makes them more effective.

What's the justification for that 10 second thing? As a hard and fast rule that's inconsistent with the way other dots/hots perform. You would get an extra half a tick on average but depending on when you refreshed that wouldn't mandate that the lifebloom was over 10 seconds. I did some casual testing in game with the combat log turned on and swapping gear to put myself just below a haste threshold it only took me a few tries to generate a cast with less than 10 seconds between SPELL_AURA_APPLIED_DOSE and SPELL_AURA_REMOVED. I'm not sure how latency figures into the timestamps in the combat log though.

Pyrates · 06/14/11, 2:15 AM

No combat log fiddling needed: If you refresh a hot (of hastened runtime t) x seconds before it runs out, the refreshed hot has a runtime of x+t seconds, which can be arbitrarily close to 9.5 as well as 11.4 (but never really reach those; ignoring negative haste effects; the exact numbers are 200/21 < t+x < 240/21). This just assumes you refresh when there's only one tick left. If you want to keep on refreshing, you'll want to use an average value for x, and then on average you'll have to refresh t seconds after your last refresh. This only goes on to show that you don't want to have a haste value significantly between two breakpoints, though.

Please ignore all breakpoints that aren't WG or RJ...

One thing that doesn't get a lot of airtime, but is significant, is the existance of WG/RJ breakpoints between 916 and 2005, namely those during nature's grace (ignoring heroism). Its uptime can be close to 25%, properly maintained, and it can be quite usefull depending on the damage pattern of the encounter.

Greentouch · 06/14/11, 2:55 AM

Originally Posted by Pyrates

One thing that doesn't get a lot of airtime, but is significant, is the existance of WG/RJ breakpoints between 916 and 2005, namely those during nature's grace (ignoring heroism). Its uptime can be close to 25%, properly maintained, and it can be quite usefull depending on the damage pattern of the encounter.

Actually it did way back when we hashed out the initial TC. We came to the conclusion that its a decent point (1776 I believe), as it gives you a substantial throughput increase (6th RJ and 2 extra WG's 25% of the time) but the other 75% of the time you are pretty much wasting 850 haste.

Stack just another 229 haste and you get the passive 9th WG/Efflo AND the 6th RJ/10th WG during NG.

So yes you're right we don't talk about it very much, I'd consider it a possibility to hang out at if you can't quite make 2005 without losing more then 1-200 int (25 man), though I certainly wouldn't want to lose any INT whatsoever to get 1776.

This is also why in BIS gear (Shard of Woe) we are going for 2032, it gives us the same point during the haste use as you get during NG. So I have 6th RJ/10th WG for 25 seconds out of every minute. I make sure I have one or the other for every major raid damage.

Dendrek · 06/17/11, 5:34 PM

Hamlet, in your updated information about Harmony you say mastery is slightly better than crit because of better itemization. But if what I've read elsewhere on this thread is correct (that since it's additive, the 1.25% for 179.28 rating is actually worth 0.8% to 1% effective healing), wouldn't crit be slightly better?

Per 179.28 rating in either stat, you theoretically gain upwards of a 1% increase in throughput (the actual increase obviously varies based on the amount of the stat you already have as well as on your spell usage). However, it's my understanding (feel free to correct me if I'm wrong) that mastery is worth 0.8 to 1x the value of an equal amount of crit rating. In other words, at best mastery is equal to crit - at worst, it's much weaker. But there are other things to consider as well:

- The value of crit rating is increased by Efflorescence (if SM crits, Efflor heals for a lot more) and Living Seed (assuming the druid is talented into it). I don't consider the fact that Efflorescence can't crit as worsening the value of crit since Efflorescence also doesn't double dip into mastery, so the two stats break even in that regard. The value of crit is slightly reduced by the increased crit rate of RG when spec'd into Nature's Bounty, but I only mention this for posterity's sake since it is not a significant loss in value.

- The value of mastery rating is decreased with higher usage of Rejuvenation since RJ receives less of a percent increase from Harmony than our other spells (because it's buffed by more additive talents than our other heals are).

[edit] In retrospect, maybe I'm wrong to consider Efflorescence as a boost to the value of crit. Although it's a beneficial impact of having more crit, it wouldn't increase our hps in any significant way more than mastery would since it's really just added healing onto Swiftmend. A 1% increase in the healing of Swiftmend from mastery is equal to a 1% chance to double the healing of Swiftmend with crit in terms of hps.

Erdluf · 06/18/11, 8:37 AM

Dendrek,

Look at the latest TreeCalcs. Crit may be slightly better for a tank-healing rotation but:
- Efflorescence double-dips Mastery.
- WG has very few additive bonuses (no GoN, Imp Rejuv, ET, ...)
Those two add up to a big piece of the "normal" druid rotation, and are enough to keep Mastery ahead of Crit.

neroleira · 06/20/11, 3:55 PM

I am confused about the haste caps I recently got some new gear and it took my haste to well below the 916 haste cap, before I reforged I took a look at my rejuve and it was still hitting a 5th tic. Have the stats changed?