Druid Changes Patch 3.1 - Page 10

seminarca · 02/14/09, 9:54 AM

One major benefit of Savage Defense that I haven't seen mentioned yet is that weapon upgrades will no longer be a case of "oh look, more threat and agi/stam". Rather, the AP to shield value relation will make sure that higher ilvl weapons having higher base DPS (and hence higher Feral AP) will provide tangible mitigation qualities beyond the slightly inflated agility. It will make dropping [Origin of Nightmares] in Ulduar that much easier. Who knows, SD might even make [Journey's End] a better mitigation weapon (despite the lower agi/stam). 342 AP higher means ~86 extra damage absorbed .. better than passive 714 armor?

Druidas · 02/14/09, 10:20 AM

bluetracker, nightcrawler message:

Assume SotF would lose about half its current armor, putting it at 11/22/33% if I recall correctly and subject to additional testing.

Oiysters · 02/14/09, 12:40 PM

Originally Posted by seminarca

One major benefit of Savage Defense that I haven't seen mentioned yet is that weapon upgrades will no longer be a case of "oh look, more threat and agi/stam". Rather, the AP to shield value relation will make sure that higher ilvl weapons having higher base DPS (and hence higher Feral AP) will provide tangible mitigation qualities beyond the slightly inflated agility. It will make dropping [Origin of Nightmares] in Ulduar that much easier. Who knows, SD might even make [Journey's End] a better mitigation weapon (despite the lower agi/stam). 342 AP higher means ~86 extra damage absorbed .. better than passive 714 armor?

A full AP/Strength gemming and enchanting strat for tanking becomes viable?

halmmar · 02/14/09, 1:42 PM

Originally Posted by Oiysters

A full AP/Strength gemming and enchanting strat for tanking becomes viable?

A stamina gem - >400 health
A srength gem ->8 more damage blocked per proc

If you go for >1 minute without being overhealed, the strength gem is better. It's just there to disguise a big, fat nerf (putting us back at Tier-4 armor values). Every single patch lately seems to have dropped us a tier of highly needed armour.

Protto · 02/14/09, 2:11 PM

Replenish – to avoid confusion, this talent has been renamed “Revitalize.” It now also works with Wild Growth.

Now that Revitalize now effects Wild Growth, it might actually be worth looking at; I just can't figure out which talent points you would sacrifice, and if it is worth that sacrifice. I was thinking Tranquil Spirit or even Living Seed if you don't find yourself using Regrowth very often.

Drashian · 02/16/09, 3:44 PM

Originally Posted by seminarca

One major benefit of Savage Defense that I haven't seen mentioned yet is that weapon upgrades will no longer be a case of "oh look, more threat and agi/stam". Rather, the AP to shield value relation will make sure that higher ilvl weapons having higher base DPS (and hence higher Feral AP) will provide tangible mitigation qualities beyond the slightly inflated agility. It will make dropping [Origin of Nightmares] in Ulduar that much easier. Who knows, SD might even make [Journey's End] a better mitigation weapon (despite the lower agi/stam). 342 AP higher means ~86 extra damage absorbed .. better than passive 714 armor?

Interesting question, but it doesn't look like it. In the example of a boss that hits for 50k on 0 armor, assuming Sanctuary or Grace in addition to our armor and Protector of the Pack, 714 armor should be (unless my math was off):

+230 damage absorbed per hit at 30k armor
+156 damage absorbed per hit at 40k armor
+113 damage absorbed per hit at 50k armor (ignoring the cap here obviously)

On a weaker target with 25k raw damage:

+115 at 30k
+78 at 40k
+56 at 50k

Since our armor values will be much closer to the lower end after this change, the AP doesn't pull ahead until very low incoming damage (the 86 absorb from AP is only when SD is up, which is sub-100% by quite a bit), and it probably never really makes up for the lost stamina and dodge. From a personal standpoint I'd probably use Journey's End heavily on trash, where DPS output is more important than mitigation anyway, but never tanking a boss.

Edit to add: See seminarca's post below for more accurate numbers on the Savage Defense side of this. Conclusion stays the same, it's still unlikely to make up for a loss of armor and stam and agility, but it's closer than it looked. An ilvl 239 DPS staff would probably give Origin of Nightmares a run for its money.

Mijae · 02/16/09, 4:39 PM

Originally Posted by Drashian

Since our armor values will be much closer to the lower end after this change, the AP doesn't pull ahead until very low incoming damage (the 86 absorb from AP is only when SD is up, which is sub-100% by quite a bit), and it probably never really makes up for the lost stamina and dodge. From a personal standpoint I'd probably use Journey's End heavily on trash, where DPS output is more important than mitigation anyway, but never tanking a boss.

I already do this, which is why I use Massacre instead of Berserking on my Journey's End. So, it really doesn't change that aspect of gearing. However, there very well could be bosses that require higher threat or dps in Ulduar. Remember back to all the weapon debates for tanking Brutallus? It also increases it's value for use when content is on farm.

Drashian · 02/16/09, 5:09 PM

Originally Posted by Mijae

I already do this, which is why I use Massacre instead of Berserking on my Journey's End. So, it really doesn't change that aspect of gearing. However, there very well could be bosses that require higher threat or dps in Ulduar. Remember back to all the weapon debates for tanking Brutallus? It also increases it's value for use when content is on farm.

Point taken. Origin of Nightmares is my primary staff on trash right now, but only because I don't yet have access to Journey's End. It's definitely more viable as a threat-sensitive boss tanking weapon with this change, which will be a great help as we encounter that kind of fight.

Originally Posted by Oiysters

A full AP/Strength gemming and enchanting strat for tanking becomes viable?

We're not quite to that point. Along the same lines as AP vs armor in my last post, while every feral-relevant stat except armor penetration and stamina gives some amount of mitigation in 3.1, none of the are major enough to gem for. In fact Savage Defense boosts agility even higher as the best mitigation stat for us to gem for, because it affects both dodge and "block" now.

Ahura · 02/16/09, 6:30 PM

Having read this thread and may other threads on the topic, it just seems there isn't enough information to really evaluate the situation.

Here's my thoughts, concerns and questions so far:
1. A major factor to be figured out is whether or not the shield is consumed by DoTs and certain AEs (like the Consecrate type, not cleaves). SD's usefulness/effectiveness relies on this point being verified one way or another. It would suck is basic poison, bleed, and similar effects consume the shield.
2. Stacks. Like most other posters, I think the effect won't stack. It's just too good to be true. I hope it does stack with things like PW:S though.
3. Many raiding druid tanks don't pick up iLOTP. Does anybody think this change might make the talent more attractive now that critting is more desirable?
4. Rage issues. I wonder how they will implement the ability as to whether or not the absorbed damage generates rage. I do hope so, as this may cause rage problems on soft-hitting AE groups, especially with everybody and their moms doing AE damage now.
5. Hidden CD. Most theorycraft posts assume there is no hidden CD. The possible existence of one puts the validity of said calculations under a big question mark. Honestly, I have a feeling there is going to be a small built-in CD (6-8 sec) but that's just what I think. Full SD uptime on AE trash doesn't seem like something that would make it to live IMHO.
6. Agi da king now? Does this change make Agi an even more important stat to stack for all obvious reasons?
7. Uptime vs % absorbed. Many of the posts I've seen talk about the uptime for this ability. Shouldn't we be looking at the average damage (or %) absorbed over a period of time, instead of uptime? I'm a bit confused there.
8. I was hoping that Blizzard would cut back on the whole AE raiding BS. Currently aside from boss fights, it's pretty much all about grouping everything up and AEing the mobs down. If you can aim a big circle, you can raid. It takes no skill and is neither challenging nor fun. This change makes me feel like AE spammage isn't going anywhere with Ulduar. And that concerns me.

I think at this point in time, the theorycrafts are simply too specific and make way too many assumptions. I think overall, the change would make our tanking a bit less spikey (attenuates the spikes). From a game design stand-point, I find it very dull and unimaginative.

seminarca · 02/16/09, 8:36 PM

Originally Posted by Drashian

Interesting question, but it doesn't look like it. In the example of a boss that hits for 50k on 0 armor, assuming Sanctuary or Grace in addition to our armor and Protector of the Pack, 714 armor should be (unless my math was off):

Those are some good numbers and I agree with your conclusion. However in the interest of accuracy, if we subject the AP difference to the same kind of rigor, we get:

[Origin of Nightmares]: (2085 + 154) * 1.06 (Protector of the Pack) * 1.2 (Predatory Strikes) * 1.1 (Unleashed Rage/Abomination's Might/Trueshot Aura) = 3132.8088
[Journey's End]: (2353 + 228) * 1.06 * 1.2 * 1.1 = 3611.3352

Which is an AP delta of 478.5264, leading to a Savage Defense delta of 119.6316. While it doesn't quite change the conclusion in view of these 2 specific weapons, there should be a threshold after which we may prefer another weapon for bosses as well.

(The numbers improve slightly for [Origin of Nightmares] if you consider 2% extra armor from [Austere Earthsiege Diamond] and 26 armor extra from the Agility difference, but that's fairly minor .. 754.28 armor instead of 714).

Deathwing · 02/17/09, 12:15 AM

Originally Posted by Ahura

Having read this thread and may other threads on the topic, it just seems there isn't enough information to really evaluate the situation.

Here's my thoughts, concerns and questions so far:
6. Agi da king now? Does this change make Agi an even more important stat to stack for all obvious reasons?

I think at this point in time, the theorycrafts are simply too specific and make way too many assumptions. I think overall, the change would make our tanking a bit less spikey (attenuates the spikes). From a game design stand-point, I find it very dull and unimaginative.

On easy content, probably. If they don't touch the sta modifiers(and maybe even if they do), on tough content, stamina is and will be the best stat. Time To Live is the most important stat, not average damage mitigated. There are very few encounters where the tank has died to average damage instead of burst.

Yes, stacking stamina will require you to be healed more, but it will also increase the efficiency of the people healing you, while at the same time giving them much more latitude. See: VW tanking 3 drake sarth with 78k HP.

Which segues nicely into my main concern about savage defense: it reduces our raw time to live vs physical damage. I'm sure they'll figure out how to balance the average mitigation so it will be a small difference. But, this shield reduces(drastically? no idea) our ability to stay alive in the worst conditions, which is arguably a tank's most important stat and replaces it with mitigation that requires offensive activation. Even if the mitigations come out near equal, I don't see the TTL aspect balancing out.

pepesebranco · 02/17/09, 6:10 AM

I noticed one thing about the little formula I derived above for the probability that the shield is up just before a swing on us. For quick reference here's the formula and definitions.

Originally Posted by pepesebranco

Definitions:

c: Basically what percentage of our attacks will be a crit. This is not the same number as in the character window.

a: Our total avoidance.

bs: Time in between two attacks on us.

ms: Time in between two of our attacks.

AP: Our attack power

p: Probability that when the Boss/mob swings at us, the shield is active.

$p=\frac{1-(1-c)^{\frac{bs}{ms}}}{1-a(1-c)^{\frac{bs}{ms}}}.$

This is against one mob. Now say that we have n copies of the same mob attacking us and we swipe. We can model this as a single mob with a higher attack speed of bs/n and that our crit chance has increased (since swipe can crit individually at each mob). What would be the chance that swipes crits at least for one mob? It's one minus the chance that it does not crit at any so:

$1-(1-c)^n.$

So we replace
$bs \to \frac{bs}{n},$
$c \to 1-(1-c)^n.$

What's the net result for p? It doesn't change!!! The effect of increased mob speed and increased crit chance exactly cancel each other.

Pzychotix · 02/17/09, 6:40 AM

Originally Posted by Deathwing

Which segues nicely into my main concern about savage defense: it reduces our raw time to live vs physical damage. I'm sure they'll figure out how to balance the average mitigation so it will be a small difference. But, this shield reduces(drastically? no idea) our ability to stay alive in the worst conditions, which is arguably a tank's most important stat and replaces it with mitigation that requires offensive activation. Even if the mitigations come out near equal, I don't see the TTL aspect balancing out.

The thing about it is that in "worst case scenario TTL", bears already had it best, having high armor to compensate for lower avoidance/no block (both of which don't count into TTL). I don't see SD nerfing us to the extent such that we would suddenly be below all the other tanks; in all probability, it would knock us down to somewhere just above the pack, due to our high innate stam scaling rather than just mitigation.

Erdluf · 02/17/09, 8:27 AM

Originally Posted by pepesebranco

hence, the probability that the shield is up before a particular Boss swing is:
$p=\sum_{k=1}^{\infty} (1-a)a^{k-1}[1-(1-c)^{\frac{bs}{ms}k}]=\frac{1-(1-c)^{\frac{bs}{ms}}}{1-a(1-c)^{\frac{bs}{ms}}}.$

Suppose I have zero avoidance, 100% crit, and I swing half as often as the mob.

I should have the shield up 50% of the time, but this formula says I will have it up all the time.

Edit:

Let sw = expected number of my swings (your k(bs/ms)). Let p1 be your expected shield probability for zero avoidance (1-(1-c)^sw). If we change that to a discrete form where (sw=3.4) means that there is a 60% chance that I make three hits, and a 40% chance that I make four hits, then my likelyhood of having the shield up, p2, is

p2 = (1-((1-c)^floor(sw))*(1-mod(sw,1)*c))

c:	40%	
sw	p1	p2
0.01	0.5%	0.4%
0.1	5.0%	4.0%
1.5	53.5%	52.0%
4.5	90.0%	89.6%

c:	80%	
sw	p1	p2
0.01	1.6%	0.8%
0.1	14.9%	8.0%
1.5	91.1%	88.0%
4.5	99.9%	99.9%

So particulary for fast-hitting mobs (small sw) and high crit rates, your formula is too optimistic. With fast hitting mobs (or lots of mobs) the shield will be less effective.

Edit2: Against lots of mobs, the probability that my last swing had at least one crit approaches 100%. The probability that my shield is still up against any particular attack approaches 0%. 0% * 100% = 0%. The shield loses effectiveness as the number of mobs increases.

pepesebranco · 02/17/09, 11:23 AM

Originally Posted by Erdluf

Suppose I have zero avoidance, 100% crit, and I swing half as often as the mob.

I should have the shield up 50% of the time, but this formula says I will have it up all the time.

Edit:

Let sw = expected number of my swings (your k(bs/ms)). Let p1 be your expected shield probability for zero avoidance (1-(1-c)^sw). If we change that to a discrete form where (sw=3.4) means that there is a 60% chance that I make three hits, and a 40% chance that I make four hits, then my likelyhood of having the shield up, p2, is

p2 = (1-((1-c)^floor(sw))*(1-mod(sw,1)*c))

c:	40%	
sw	p1	p2
0.01	0.5%	0.4%
0.1	5.0%	4.0%
1.5	53.5%	52.0%
4.5	90.0%	89.6%

c:	80%	
sw	p1	p2
0.01	1.6%	0.8%
0.1	14.9%	8.0%
1.5	91.1%	88.0%
4.5	99.9%	99.9%

So particulary for fast-hitting mobs (small sw) and high crit rates, your formula is too optimistic. With fast hitting mobs (or lots of mobs) the shield will be less effective.

Edit2: Against lots of mobs, the probability that my last swing had at least one crit approaches 100%. The probability that my shield is still up against any particular attack approaches 0%. 0% * 100% = 0%. The shield loses effectiveness as the number of mobs increases.

Going directly to the point: I completely agree with you! My post about the shield probability being the same regardless the number of mobs is wrong. I need to think about this some more.

. Basically my approximation:

The second approximation is that between any n swings of the Boss I have swinged n*bs/ms times.

doesn't work in the multi (or very fast hitting) mob case.