Hand

Defense gives more avoidance than dodge point for point.

EDIT: the first time I did the math, I came out with dodge on top. Then I came out with defense on top. I just used the wowwiki numbers and came out with dodge on top again...weird. I actually have no idea.

 

Berthold

I used def rating to compare the full avoid and the full stamina with a mixed choice (4 def 6 sta). And avoidance-wise they are both very close together:
1 def rating = 0.051% avoidance
1 dodge rating = 0.053% avoidance

 

Berthold

Unfortunately I am quite busys right now in real life, but maybe I can find some spare time in the next holiday season for this. But because of the need for some kind of artificial intelligence to model the healing done the chances are not very high.
And for the second point I think you are exaggerating a bit. If you take a look at the "fast hitting boss" graph you definitively see that an avoidance tank has a way higher survivability with a fixed healing setup, so that means when viewing from the other direction:
with a flexible healing setup and static survivability you can guarantee the same survivability with less healing.

 

Chimera

So you chose it because there's no (crafted) dodge/stam hybrid gem? Makes sense.

Any chance of you running the sim with Illidan numbers? Or if not, making the code available to download so that I can do it myself?

 

Xerophyte

I'm sorry, but that implication just isn't true. What you're saying is equivalent to stating that because a > x is true for some value of x, it must be true for all values of x. Which it's not. In this case, it does not hold that because a particular gear strategy is effective for a particular healing strategy it must be effective for all healing strategies, or part of the most effective combination of gear strategy and healing strategy.


To illustrate the point further with an artificial example:

The boss hits for 100/second

Tank A: 50% avoidance, 200 health
Tank B: 0% avoidance, 400 health

Strategy A: Heal for 90, 1s cast time, 100 mana.
Strategy B: Heal 200, 2s cast time, 100 mana.

If using strategy A then tank A clearly has the superior gear, tank B will die every time.
If using strategy B then tank B clearly has the superior gear, tank A will die every time.

The correct choice is tank B with strategy B, however if all you examined was strategy A you would not deduce this. Obviously, actual in game situations are much more complex than this but trying to reason from a single healing strategy will still not lead to correct conclusions. You need to take both into account. Which gets bloody complicated, but there's no handy way around that: if you don't do it your conclusion will invariably be invalid.

 

Berthold

The code is available in some of my posts above. I am going to simulate the fight this afternoon, but I would be surprised not to see avoidance as the winner.
The 21k hp/50% avoidance tank is without sockets I assume? or else the numbers would seem very low (fully flasked I reach 20k/50% and we are not even remotely close to Illidan).

 

Berthold

I agree on your point that this is not a proof. But by using a mixed healing strategy (which will be found in many raid setups) I dare to say that the suggested gear choice is qualitatively the right result for most of the (realistic) healing strategies.
To do it right you would have to reengineer the WoW Code, and model every possibility to setup the raid, player behavior, gear choices, and then search for the optimum in these myriads of variations. Have fun with that
My result won't be the perfect choice for all healing strategies (as you proofed), but a good heuristic is better than no advice.

 

Berthold

@Chimera:

Illidan Stormrage - Attacktable:
0-69 avoidance (should count as a dualwield mob like the prince)
69-84 crushing
85-100 normal hit

He will hit for (BaseDmg + 0..1000) in the offhand and can hit double of it with the mainhand. 2.0 second swingtimer.

From Nihilums kill video I concluded ~30 sec CD for Flame Crash and Drain Soul which will both hit the MT once for 5k (but are not allowed to occur at the same time). Without the specials Illidan would be a joke tankwise*G*

every odd second mainhand hits, every even second offhand hits.



mainhand+offhand at the same time



So there is a slight advantage of using avoidance.

 

Chimera

thanks for running my numbers.
You're absolutely right. Really the only time that the main tank will die is if they mess up and get shear. Mother might be a more interesting fight from the point of view that the tank can actually die, but I already know that stamina will win on that fight.

not that I think it will reverse the trend in your graphs, but shouldn't the table look like this:
0-69 avoidance
70-94 block
95-100 crushing

 

• Ren

Illidan doesn't crush.

 

Chimera

illidan doesn't crush our tanks with full gear setup because they have about 105% avoidance + block chance. (due to him dual wielding)

are you quite certian that he doesn't crush period? we haven't actually tried a druid tank on him because of shear, but I had assumed that becoming immune to crushing was one of the advantages of avoidance over stamina on that fight.

have you actually tried a tank with less than 100% avoidance + block chance on illidan?

 

Teez

He doesn't crush, still. With less than 102.6% combined avoidance he can apply shear, though - which is counteracted by hitting Shield Block when he starts casting Shear. You just don't shield block the rest of the time to make sure you have the cooldown ready and sufficient charges to block out the Shear debuff.

edit: In other words: becoming immune to crushing is favorable on this fight because it allows you to completely disregard Shear.

Warrior - Shaman - Priest

 

Chimera

white attacks crush. special attacks do not.

His white attacks have ~24% base miss chance. His special attacks have ~5% base miss chance.

It is therefor much easier to become crushing immune than shear immune.

 

Teez

His white attacks actually don't have a 24% miss chance. Here's the deal for Shahraz & Illidan. They both dual-wield, which would imply 24% miss chance. 24% miss + give or take 8% miss from defense, lets say 30% dodge and 20% parry, and 25% block for good measure. That's 107% total. However it appears as though both Shahraz and Illidan have a +19% hit modifier giving them the base 5% chance to miss. The combat table still doesn't allow for crushes since they're pushed off to start with and replaced with hits. Thus, you end up with them *never* crushing if you have half-decent gear. The stats I listed above are incredibly easy to come by by the time you get to Illidan. However, you're right, you still need 102.6% base avoidance to never have to worry about Shear. In all likelihood, you'll need more actually, since he negates your chance to be missed and his base 5% miss chance on specials via increased +hit.

edit: you seem to have missed what i was saying about shear initially - of course it can't crush. You can avoid the debuff from ever landing though by making sure you either avoid or block it. This is accomplished by timing your shield blocks correctly, which again is possible because he won't crush you, ever, even with shield block down, which allows you to not use it whenever he's not casting Shear, to keep the cooldown free.

Warrior - Shaman - Priest

 

Chimera

Of course our tanks shield block the shear. I (apparently erroneously) thought you were talking about becoming passively immune to shear.

Is there a discussion somewhere about the rather peculiar way you described becoming crushing immune on mother/illidan?

 

• frmorrison

How about trying that sim of Illidan again, but this time with BT gems:

10 defense
5 def/7 stamina
15 stamina

 

thugthedum

so .... what's your modeling look like? how'd you do it?

I ask because I have a pet project for a rage simulator, and I'm curious to see if we are using similar techniques. Mine is fairly transparent, I think you'll find.

http://www.ro-guild.com/thug/rage.php

 

Teez

There was a thread/post explaining it somewhere, and after 15mins of searching I still can't find it. I don't know if I'm just imagining things or if my search terms are badly picked but yeah, take that for what it's worth. For all intents and purposes though, I yet have to see a combat log of a single instance of either of them crushing. Same goes for Archimonde.

Check out any WWS log, for example Facecleaver on Shahraz - and you won't find a single crush. You could say it's great Shield Block uptime, but even without blocking our tank doesn't ever get crushed.

My +hit theory is backed up by the following part in the WWS I linked - assuming, again, a 7-8% miss increase from +defense, my suggested 5% base miss + 8% def miss should be around the 13% mark, give or take. Else, they'd be sitting at 24% + 8% = 32% miss. If you expand "Normal" damage taken on Facecleaver's WWS entry, you'll see his miss% is far closer to 13% than 32%.

Granted, this sample size isn't exactly monstruous (a mere 172 swings total), but this is just one example of many - browse any Shahraz or Illidan WWS and you'll come to the same conclusion.

Warrior - Shaman - Priest

 

Teez

I just gave that a try and got some ... rather odd results. Apparently, boosting my +hit rating from 88 to 150 (I was putting in some ballpark value to account for the lack of being able to add weaponskill) made my TPS go from 875 to 824 - could it be that you're using the inverse of what you intended to there? I'm having the same thing happening if I increase mob damage, keeping everything else constant. Went from 13k damage before mitigation to 23k and lost considerable amounts of TPS - something seems a little off here :/

Warrior - Shaman - Priest

 

Chimera

We use a bear tank for Archimonde (yay fear ward) , who doesn't get crushed. he sits at about 50% dodge, so if Archimonde had a 24% miss chance against him he would still take a lot of crushing blows. I think it's pretty clear that, at least in the case of Archimonde, blizzard didn't do anything more complicated than simply turn off crushing blows.

You've convinced me that Illidan and Mother can't crush, but I'm not sure I buy your explanation as to why. It seems more likely that blizzard simply turned off crushing blows artificially, as they did in the case of Archimonde.

We're straying a little bit far from this thread's topic, so I'm going to go try to dig up the thread you mentioned rather than drawing out the conversation here.

 

• Ren

Teez's explanation (Shahraz/Illidan have +19% chance to hit) seems correct because Shear should have a chance to miss but it never does.

 

Teez

Yup, the only ability that throws that whole idea off a bit is Saber Lash - because that definitely misses me quite frequently.

Warrior - Shaman - Priest

 

Dots

Shahraz doesn't dual-wield. Gurtogg and Illidan both swing very fast and clearly only hit for half damage quite often, indicating that they do dual wield. That is not the case for Shahraz.

Crush immunity is nice, but it doesn't apply to as many bosses as some might think. The following Hyjal/BT bosses don't crush: Azgalor, Archimonde, Supremus, Shahraz, Gathios (council), Illidan.

 

Teez

Ah, thanks for clarifying that Dots - for some reason I've had that misconception stuck in my head for quite some time now about Shahraz.

Warrior - Shaman - Priest

 

Berthold

Should give the same results but I will think I will test it this afternoon.

For the Shahraz fight. Can some one give me a very quick glance at the fight concerning incoming damage/attack table/swing timers?
Then I will give it a try.