Berthold

After hearing lots of "arguments" for and against socketing avoidance / stamina I decided to write my own simulation tool to see what gem gives the highest survivability.
Results first:

Stamina wins so my gut feeling was right, hehe.

The tool simulates a boss fight against hardhitting Bosses like Gruul / Magtheridon.
The Boss has a 2.0 second swing timer with 6500 BaseDmg (after Armor) and cleaves for 1.5*BaseDmg every 11 seconds.
The combattable is just plain avoid or hit (unmodified at 0.5 avoidance).
Heal is modelled as 2 full Stacks of Lifebloom ticking every second for 1.8k, a 4.5k greaterheal every 3 seconds and a flash heal every 2 seconds. Both of the last two heals have a 10% chance of not being cast due to AoE Silence / Brainlag whatever.
The MT has 18.2k HP and the fight continues for 10 minutes.
It is evaluated over 200.000 tries for each socketing option(15 slots assumed), and the output is the percentage of wipes.

Any comments would be appreciated.



And thanks again ej for hosting such a nice forum, I gained a lot of insights reading it.

Bye Berthold

p.s. If anyone is interested in the code I can send the VC++ project.

 

Cantic

The general consensus is that once you hit the defense cap, stack stam since you will naturaly gain avoidance on your gear. Most can hit the defense cap through gear and talents alone so +12 gems would be the way to go, as your graph demonstrates =).

 

• Navaash

please model patchwerk next tia

Bye Berthold

 

Braque

It is well known that hard hitting (spike damage) encounters favour high hit-point tanks over high-avoidance tanks. A more interesting question would be to find under what conditions a high avoidance/mitigation tank comes out on top.

I suspect it is under multi mob tanking conditions, where many potentially fast hitting mobs are being tanked, and the risk is not that the tank is spiked down while healers have mana, but that the healers run out of mana.

There could come a point (perhaps one that TBC gear/encounters will never reach) where more stamina is irrelevant, and more avoidance/mitigation is what makes a fight feasible.

In theory.

Nice graph though!

 

◊ Anias

The only fight where avoidance will trump stamina (in the ratio given to the two stats in game of 8 dodge rating vs 12 stam) is a fight with a severe mellee applied debuff. Think pierce armor, mortal strike, or something similiar. In such a fight, the time spent outside of the debuff for the avoidance tank could potentially outweigh the lower effective life total.

One of the reasons not often highlighted that stamina comes out ahead as often as it does in TBC is that 12 stam is more effective life than 8 dodge rating, for the same itemization budget, at 70. Blame the pvp change, but in terms of itemization points, you gain more from stam than the other (non armor) stats. Stupid? Yes. Reality? Yes.

 

Berthold

As for the melee debuff I would support your thesis. But even with +8 Stamina the graph is lower than the 8 def rating until 7100 basedmg. From then on they are more or less the same,

 

◊ Anias

That's a result of your initial values =P

 

Trouble

Class Mechanics and Theorycrafting

 

Hand

The problem with this (and the problem with most people that theorycraft mass stam > avoidance always) is that you assume the only way to wipe is your main tank dying to spike damage. The less damage the tank is taking, the more heals you have free'd up for whatever else is happening in the fight. Most of the time (from my experiences) you wipe for reasons other than a tank death, usually something to do with a non tank taking some amount of damage, and not getting healed quick enough before he dies. If the MT is sitting at 100%, that rogue that dips down to 5% and is about to get killed is a lot more likely to receive a power word shield/ns heal/flash heal etc than if your MT is taking very consistent damage. Additionally, assuming your healers are perfect and mindlessly spamming you for the entire fight is also an incorrect assumption, at least for my guild. Just because you can theoretically survive a "maximum burst" from a boss doesn't mean you're going to get the amount of heals you expected in said time period, and doesn't mean the next hit won't potentially finish you off.

It should also be known that defense is the best avoidance stat, better than dodge and parry point for point, which I assume is why he used defense. Has nothing to do with the defense cap, even if you're capped its still your best stat avoidance-wise.

All that being said, I still stack stam in every slot, but its mainly because of --
"One of the reasons not often highlighted that stamina comes out ahead as often as it does in TBC is that 12 stam is more effective life than 8 dodge rating, for the same itemization budget, at 70. Blame the pvp change, but in terms of itemization points, you gain more from stam than the other (non armor) stats. Stupid? Yes. Reality? Yes."

 

PsiVen

If the MT is sitting at 100% and one of his healers switches to heal that rogue, what happens when he stops dodging? The problem goes both ways, and realistically you can't afford to change the number of heals incoming, only the percentage that get canceled before they land.

Btw, this study could be improved by looking at socket bonuses; e.g. how 5 def 7 stam x2 + 6 stam socket bonus fares against 2x 15 stam (so 10d 20s vs 30s) or the blue versions.

 

• Astrylian

These numbers look excellent to me, and give you a very good idea of how to gem your gear, and choose gear...if you have 50% avoidance, 18.2k HP, and are constantly being healed by healers who have infinite mana.

 

◊ Lord BEEF

I'm sure he could model it for different levels of gear and fight lengths on request. Stop giving him such a hard time

Check out my friend's bitchin' Lord of the Rings Art

 

• Xaviera

To possibly save you an infraction since your post was cool, you may want to edit all of the posts you've made thus far to have full proper caps and punctuation, etc.

Then you may want to change the aim of your post to be "Graphing Tank Damage / Life tables" or something, and then ask for input and suggestions on situations to graph visually and post here for us to see.

For example, do the same type of stat comparison for a fight that has a fast hitting boss for like ~3000, etc.

 

Caligula

To go along with what Xaviera said, you could turn the VC++ project into a program where the user could enter their own variables (Health, avoidance, incoming heals, etc.) and that way you wouldn't have to do it for each person that wants a different set of data.

It might actually prove to be a very useful theorycrafting tool.

 

◊ Maniq

Slight derail:

Edited the title and moved the thread to somewhere more appropriate.

As to all the nae-sayers, shut up and stop shitting up the thread.

Thanks!

 

Soulshade

As for how it compares to real game, it is just an estimate. His estimates on healing are not too far off, since dedicated tank healers are usually used. While raid deaths can cause wipes, most raids will not use 3 healers on a raid and will have healers around to top the raid off and give the MT extra heals.

The tool used for these calculations is just a rough estimate, it doesnt account for any raid damage, such as cave ins, shatter, fires, blossoms, etc etc, that could also lead to more tank damage. It also does not factor in 4th or 5th healers on the tank. Its just a rough estimate tool, designed to find straight forward survivability. You also have the factors of crushings and the parry mechanic causing faster swings. If you were looking at survivability versus a boss attacking at 1.0 attack speed, hitting just as hard, then yes I suppose avoidance might do better.

Really best bet for seeing avoidance in action is to create a program to test multimob tanking, seeing how stamina and avoidance does versus 3-5 mobs at once.

 

Hand

I strongly disagree with this statement, if you can reduce your incoming DPS due to more avoidance/armor, then why can't you reduce the amount of HPS you assign to yourself? I've done it a lot in the past, and I'll continue to do it as my gear improves. I used to die on naj'entus with 2 healers on me, now one person can solo me and their mana bar doesn't even move, so they can usually throw out lots of extra heals on the raid as needed, as an example.

Back on topic, re-reading my post now a few hours after I made it, I just want to clarify that I actually am glad you made this tool and I think its useful, I wasn't trying to bash on you. I was just trying to point out that when choosing your gear/sockets/etc, this isn't the only thing to be taken into consideration. I also agree with the suggestion that it would be really neat if you could have the boss damage/speed and player avoidance and all that as inputs, so people could mess around with it as they please. That would be nice to have.

 

Berthold

Thanks Maniq for moving the thread to the right forum.

Before answering to the feedback i would like to make clear that the whole purpose of this tool was to get some "proof" on the common sense, that for raidprogress stacking stamina is the best choice for hardhitting Bosses. But now on to your replys

Currently there is 1 paladin and 1 priest assigned to the MT spamming heals. The druids can keep up the 2 stacks while crosshealing the raid. So that leaves enough room for additional raid or offtank healers. It is also modelled that the paladin / priest is not able to cast the spell (with 10% chance) due to crosshealing / lag whatever - which is a realistic assumption in my eyes.

On the point of mana consumption i have to agree that currently they have an unlimited mana pool - unfortunately this would be very complex to model, maybe someone has a good idea how to do this, but (in my raiding experience) you die more often because of dead tanks, the enrage timer, someone not doing his job, than oom healers.

And for the concern of Astrylian - i don't have the numbers in my head, but i'd guess that t4-t6 roughly have the same avoidance (def+dodge+parry), so that leaves only the 18.2k hp MT, which is at first look a static number.
But its more important to take a look at the ratio of (incoming heal - incoming damage / maximum HP), which is why the basedamage is flexible.
Increasing the basedamage is comparable to lowering the maintank's hitpoints or reducing the incoming heal. This is why the model and its results should be applicable to Magtheridon as well as any other hardhitting Boss in WoW Classic or future addons.

And for further "research" i ll try the multi-mob tanking scenario and a fast hitting Boss. Lets see how that works out.

p.s. as soon as i have enough time i'll add some options to change parameters but currently it's a quick'n'dirty command line tool that outputs a text file. But if you can make yourself comfortable with VisualStudio here is the link to the sources: www.iseurope.de/theheap/wow/TankCalc.zip

p.p.s. If my posts are lacking punctuation etc. please bear in mind that i am not a native speaker

 

vorda

A ton of people here arent, just know that its not 'i' but 'I' and your post is fine.

Also, the mana aspect is pointless, cant remember the last time when we wiped because healers were OOM. (as you stated of course, just backing you up )

 

Berthold

1 second swing timer , 5 second cleave


So here avoidance is clearly in favor. Multimobtanking still pending.

 

david0925

Your table shows a optimiziation of gear for what happens "most of the time"

Tanks stack a lot of stamina for worst case scenarios, which happen "some of the time"

Are you wrong? no. But its really a viewpoint difference

 

Berthold

That's why I am not looking at how much expected damage a boss has to deal in order to kill me, which is totally different subject, but simulate a whole boss fight. The results show me that using stamina gear when encountering a hardhitting boss with basedmg of 7000 I will die in 15% of the tries and if I use avoidance gear I will die with 35%. But if one encounters a fasthitting Boss (like Prince) avoidance is the way to go (up to 20% lower chance to die at some time in the encounter).

 

Berthold

And now the multimob situation:
6 mobs with no special abilities (i.e. standard level 72 or less adds) and a 2 second swing timer


Which also favors the avoidance.

 

Chimera

Is there a particular reason you've used defense rating instead of dodge rating for your avoidance stat?

edit:

an illidan tank has about 21k hitpoints, 50% avoidance (full consumables)

illidan swings for 4-6k non crush mainhand every 2 seconds
illidan swings for 2-3k non crush offhand every 2 seconds

because of the mechanics of the fight, assume the tank is NOT crushing immune.

 

◊ Xerophyte

Ah, but additional stamina also improves healer mana efficiency as a high-health tank allows healers to better optimise their strategies. A greater margin for error allows them to cancel more overheals, rely more on efficient heals - typically lifebloom stacks - to top the tank up and chose slower, more efficient casts in general. You can also accept fewer healers assigned to a higher health tank because the drop in damage-to-heal lag that a drop in assigned healers entails (yes, even with healers who know enough to pre-cast) is less likely to be lethal, allowing for a more flexible setup with more healers covering both tank and ra