Avoidance vs. EH.

Quote:

---

Originally Posted by Aggathon

Can you link me your tank's armories that are "stacking stam"? Because so far the only people "stacking stam" that people have showed me are people that use really good stam trinkets (which I've stated before is more than acceptable for changing between fights, in fact in this expansion I think that trinket switching is going to be the most important thing between fights) and maybe have 1 or 2 straight stam gems, but the rest are avoidance/mastery gems and matching socket bonuses.

The only tanks I've seen that people just can't keep up (specifically this was halfus) were tanks that just stacked stam. Like I was saying before, if you do need more raw HP for a specific fight, switch trinkets and if it's stopping your progress, then stack stam, but so far from what I've seen (granted papapaint has a much much better view of this than I do) the HP requirements are not really THAT high. They're high for some fights, but it doesn't seem like stacking stam out da butt is necessary, at least not for warriors.

Edit: I armory stalked your guild. http://us.battle.net/wow/en/characte...jubie/advanced <- I have no idea if that is one of your main tanks, but this person is gemmed exactly how I would gem that same gear. That epic stam trinket is amazing, and the other one is okay, there are better and it should be switched out depending on the fight and other trinkets available.

Edit2: http://us.battle.net/wow/en/characte...kaeps/advanced <- same with that guy, that's how you gem and reforge a warrior.

---

Neither toon is in our main raids, those are both farming alt raids--we'd be having a hell of a time if our main tank were stacked in half blues :p

http://us.battle.net/wow/en/characte...loree/advanced

There's our MT, matching socket bonuses with Stam hybrid gems, running double stam trinkets. However, the question at hand is whether or not Avoidance is more effective than EH, not whether or not socket bonuses provide enough incentive for players to match them. With health pools so hilariously large at the moment, I do think it does make slightly more sense to grand slightly less stat weight to stamina than WotLK. Not, however, to the extent of preferring avoidance to EH.

The only argument I've seen in this thread in favor of avoidance over EH are these mythical problems of healer mana and healers being unable to sustain a tank under heavy damage, both of which are patently untrue. At early gear levels, it can be difficult and mana intensive to keep up a single target, but just about any healing will be difficult and mana intensive. In full 359, it's relatively cheap to keep up a single tank without concern for mana or being unable to handle incoming damage vs. healing output.

Quote:

---

Originally Posted by Papapaint

http://us.battle.net/wow/en/characte...loree/advancedI do think it does make slightly more sense to grand slightly less stat weight to stamina than WotLK. Not, however, to the extent of preferring avoidance to EH.

---

Do you consider mastery avoidance? Because I'm not sure I'm seeing any argue for dodge/parry gearing as the primary gearing option.

Mastery is not EH until you have passive unhittable, and even then if you gear for that magic damage makes you spikier than a blue tank on malchezar.

Fascinating Papapaint. Duly noted.

I am kinda interested in how block weighs into that though. For a warrior once you're at 72.4% dodge+block+parry, if shield block is up (which it should be when you need it) then block is essentially another layer of armor since it's all multiplicative now and is for mathematical purposes, EHP.

How gearing curves work, it seems like in earlier content it's better to stack mastery, so maybe the ideal gearing path for a warrior would be more like mastery and avoidance/stam hybrid gems up to 72.4%, then stack stam once you get into heroics.

Quote:

---

Originally Posted by gacktt

Mastery is not EH until you have passive unhittable, and even then if you gear for that magic damage makes you spikier than a blue tank on malchezar.

---

Understood. But it's not avoidance either. It's RNG mitigation. 72.4+ is completely reachable, and 90+ before miss seems to reachable in full heroics.

Arguing that gemming mastery is avoidance isn't accurate as it will never be avoidance, and it's also confusing because it's a sliding scale. At soft-cap, it is gauranteed mitigation with SB, and next tier, it will likely be guaranteed passive unless they change the amount returned or add mastery DR.

To group mastery with dodge and parry muddies the water quite a lot.

Even so, if I recall from the previous WotLK disscussions, there was a point avoidance-centric 102.4% gearing led to greater survivability compared to EH, and that was 85%+ generally. At that point a double attack is highly improbable, and a 4 or 5 steak is very limited. But it was a theoretical point as the gear didn't allow it to be reached.

Just like before, block (mastery) doesn't fit into avoidance and doesn't quite fit into EH (yet). I think it's important, for keeping the discussion clear and on point, to think of them in that content. Dodge/Parry = avoidance. EH is damage reduction and stamina. And Block is a stepchild, but one that fits more closely with EH, espcially in times of SB usage, than avoidance.

So aren't we back to Burst Time then?

With Block being considered as avoidance below Unhittable for 2/3 of the time and being mitigation (and crit block being avoidance) 1/3 of the time, assuming max SB uptime.
(This would be true for nearly any fight except Chimaeron, where you would be able to SB every Double Strike, thus always counting it as mitigation.)

Even when for the time one is below Unhittable, block (or mastery) is not avoidance. You GET dmg when you block or crit block. It has to be healed. A blocked attack needs some healing afterwards. An avoided hit does not.

Especially if you think about the model of Burst Time you linked, this is important. Because the assumption is, that one can be healed back to full, if one avoids at last once. That's not the case for a block that is following a full hit in Cata heroic raids. At least not everywhere and all the time.

Btw: Mastery also gives a warrior crit block. Which is not to neglect. At the mastery values we are discussing one will at least crit block over 50% of the time (more like 60%). Those crit blocks may be somewhat nearer to the assumptions made about Avoidance in the Burst Dmg Model.

However, I don't think any old model works for todays block-masteries before Unhittable.

One important part is, that healing actually has changed, too. Heal-sizes are reduced compared at least to the Wrath model. And Blizz has introduced a relatively stringent approach to heal sizes and their mana costs. Ok, most healers have some kind of free or low cost instant heals to add, but overall quick heals are relatively big but extremely expensive. No healer can afford to use them all the time. So one new element is to at least think about the pattern of dmg that has to be healed. Sure everybody will get dmg spikes that need at least one or two fast heals afterwards. But that's mostly because of special boss activities that are just there and can be predicted. The important part is how the dmg income will and should look for the rest of the time. Because that's the time where the healers need to get some time to heal the tank back up hopefully with less expensive heals and maybe even get a break so that they can do at least something else than chain-spamming the tank (popping a regen CD, throwing something at the group, reposition themselves, refreshing stuff on the tank, ...)

The old models don't include the new healing model. But it's relevant and most of us include it in our gearing decisions. It's not that mana is too scarce or whatever. It's less that heals are not as big compared to healthpools anymore. The important new element is that the amount to be healed is not the dedicating element for mana costs, but the time you have to heal this amount. If you have enough time to heal it with the slow cheap heals, every healer in raid gear will even regen mana while doing so. If the heals have to be faster they get more expensive for the same amount. I've only healed some bosses in normal mode. But I don't think that any healer is able to use the burst heals constantly with the gear one has for heroic modes for the whole course of one encounter. (Papapaint should correct me if I'm wrong here.)

So tank models that expect burst healing are less appropriate for current raids. And that's one part of the Burst Time model. Old avoidance is about reducing the amount that needs to be healed. EH is about surviving on one's own as long as possible with healers doing anything and hopefully something to get all those tank health back up again.

The old TTL model works better. It's about what a tank is able to do without getting heals. You can add heals and see how long the TTL is with different kinds of heals. And that's something where block shines much more than either Avoidance or old EH.

Agg, I have historically loved Burst Time but I've paused to think about its implications nowadays and am not 100% certain it's quite right without some minor adjustments to the model.

a) Block doesn't fit in well with Burst Time and is somewhat undervalued compared to its practical value. It would need to be considered in a bit better of a way somehow. (From the Burst Time mentality, Block is simply weaker than avoidance. In reality, it's pretty awesome.)

b) I'm not certain that the model of 'you get topped off between every avoid' is quite right given the size of heals nowadays, at least in 10-man raids. I'm certain it takes longer than one swing timer to top off a low-health tank nowadays in most cases. (I could be wrong about this, but that is just my observation.)

I still think Burst Time is a good statistical model--it just needs some adjustments to account for new mechanics.

Given the nature of incoming damage and healing, I'm simply inclined to take the 'minimum HP threshold then stack the most efficient mitigation' approach--which is basically what all Druids do indirectly via Agility stacking and what most Warriors and Paladins are doing with Mastery stacking.

At the end of the day, this may get tougher when gems get bigger and we have more slots, but honestly it's hardly even something we have to think about beyond trinket slots anymore. The difference between stamina gems and avoidance gems is not a whole lot of health, and a number of enchant slots are one or the other, rather than choosing between both. So, it comes down to carrying around a couple different trinkets and swapping when needed.

As a quick follow-up, the biggest argument I see right now for stacking avoidance is this:

There is a lot of raid damage nowadays in many fights.

It seems less of a tank's job to be a mana sponge/burst damage avoider as much as it is to minimize the amount of time healers have to spend healing them. More fights seem to be about managing healer time (rather than mana, although that's a consideration as well) than it is about avoiding massive attacks on the tank.

Certainly a few fights (and I presume hardmodes, although I haven't done many myself) are tough on tank healing, but erring on the side of damage reduction in the majority of cases seems to be working pretty well for me. (Which is good, because otherwise Druids would be pretty crap.)

Shouldn't the debate shift away from avoidance vs. EH towards "high volatility" vs. "low volatility" gearing?

1. Low Volatility Gearing: One gearing strategy would be to stack nothing but mastery. This would decrease the overall variation in damage you are taking over time. This strategy would benefit healing styles that provide a steady, predictable flow of health to the tank.

2. High Volatility Gearing: The other gearing option is to stack dodge/parry + stamina. This choice would increase the overall variation in damage you receive. There will be long periods where the tank takes no damage at all, and others when he is constantly being hit. A larger health pool is needed to help mitigate the bad spikes. But the down side of volatility can also be countered with skill: wise use of trinkets, abilities, etc. This strat seems to be more suitable for healing styles that provide burst healing and then regen, or for those who have multiple healing assignments.

Given that there is very little difference in overall stat points between dodge/parry/mastery according to Rawr, I think either strat can work. But personally, I like the second strategy. It is like hedging a high-risk stock. With skill, you can minimize the downside, and the upside can be used to save the healers mana and stress. I think there needs to be more people speaking up for dodge/parry rather than just the "all mastery, all the time" status quo.

The thing is, even though there is little statistical difference, Mastery and Parry do come out on top in terms of total damage reduction at moderate gear levels anyway. Almost any profile optimized for Mitigation in Rawr.ProtWarr, for instance, will be at least soft-capped during Shield Block for 100%--and typically be advised to stack primarily Mastery with perhaps a bit of Parry reforging.

For example, optmizing a very well-geared Warrior with heroic gear for mitigation ends up with stats something like 2.5k Mastery, 2.5k Parry, and 1.8k Dodge. This yields 35% avoidance and 52% block--adding to ~88% chance to either avoid or block without Shield Block active.

(And obviously 100% with Shield Block active. This averages to 92% of attacks being either avoided or blocked with normal Shield Block usage, and 24% of all incoming attacks on average being Critical Blocks. This means around 54% of attacks are either avoided or critically blocked.)

So, the "high volatility" avoidance is generally worse on both accounts. Parry and Mastery both have nice synergy and with the increasing returns of Mastery make for a solid combination for Warriors.

In regard to Druids, there is little to be be considering either. Agility increases both forms of mitigation (shield absorbs via Crit and AP in addition to Dodge) and as such is a well-rounded gemming approach. Otherwise, all ratings-based contributions are done via a mix of static absorb-based increasers like Crit/Mastery, with the lowest value being reforged to Dodge. As Dodge extends the potential duration/uptime of absorbs, there is some synergy there that works together better than going one direction or the other.

Paladins seem to have Mastery as so efficient that it's hard to see them not going that direction. DKs I don't really know enough about the balance of mechanics to say.

So, really, the theory you mention is probably interesting--but in practical terms, the low volatility options are almost always coming out on top either way.

This debate comes down to gemming options, really. You can't reforge to Stam, and people seem to generally agree that at this point in the gearing spectrum, mastery is better than dodge/parry. That leaves Gems and Trinkets. Trinkets can and should be switched specifically for each fight, so that's kind of a moot point. Gems are harder to do this with.

I have 12 total gem slots, which could net me as much as 720 stam if I ignored socket bonuses (I'm counting the cog slots in my helm like normal gem slots). That's close to 12k HP. If I went for socket bonuses, I'd get around 570 Stam, or ~9k HP.

If I went all mastery, I'd get 608 mastery rating, which is somewhere around 3% block. If I gem for socket bonuses, as I do now, I have something like 448 mastery rating, which is probably around 2.5% block.

So the swing to stamina only nets you around 10k-ish more HP, while the swing to Mastery nets you around 3ish% more block. That's what we're really fighting over here. Trinkets change based on fight.

Personally, I'd rather have the 3%ish more block than the 10kish more HP, for general purpose. I might switch in stam trinkets for some bosses, but generally speaking, I prefer the mastery.

It's especially worth considering that as you push upward on Heroics and get higher ilevel gear, the value of Stamina gems drops pretty quickly. That 10k HP is only 5% of the health of a buffed Warrior with two Stamina trinkets in mostly heroic gear--which can easily hit 200k from what I've seen.

At that point, 16-17 gems of Mastery starts being very powerful due to the increasing returns, so it's a harder case to make that it's worth dumping a ton of avoidance for 5% health.

It also doesn't hurt that the Stamina trinkets themselves have tons of avoidance and/or Mastery on them anyway via the effects--so even if you love avoidance, the "Stamina" trinkets are pretty nice.

While this may be an argument in the future, it really isn't a meaningful choice right now.

@Bakabon
You cannot assume that only tanks gearing for avoidance over mastery would use trinktes and CDs with their skill.

Actually I see no point at the moment to gear for avoidance over mastery for block tanks. At least while speaking about raid or even heroic raid content. There are no such healing styles like you assume. You cannot burst heal and regen in Cata like you could before. There is no mp5-rule anymore and burst healing costs much more mana than constant healing with lower amounts.

There are a few points where avoidance is slightly better than Mastery (primarily Parry, due to Hold the Line reinforcing Mastery indirectly) in the reforging department. As a general rule, spamming Mastery works fine though.

Sure, but in most cases you get this parry from reforging other values where mastery is on option, anyway.

@Katzazi
I'm not trying to imply anything about the skill of differently geared tanks. What I'm saying is that volatility has an upside and a downside. On-use abilities can be used to counteract the downside, just like hedging can protect against the downside of a volatile stock. The upside is that during the non-hit streaks healers can either save mana or be free to focus on other targets.

@Kojiyama
Given that as you say there is little statistical difference, I think what is more important is the type of healers you have and their awareness of your gearing strategy rather than a hard-fast rule of stacking mastery. It seems to me that given the variation in healing mechanics, some healers will be better at healing "high volatility" tanks, while others would be better with the "low volatility" variety. It also of course depends on the encounter...

You can adjust Block to be relative to Avoidance. Because avoidance is already a chance on occurance, you can model Block to be 30% (or 60% for Crit Block) of it's chance as 'full' avoidance. e.g. 30% 'normal' block chance would be equal to 9% avoidance.

While healing has changed, it is not as drastic as some claim/wish it to be.
Although you can not mindlessly spam heals, your healers will usually spend certain timeframes of a fight expecting a certain amount of damage. This will allow them to predict an average HPS requirement and during those timeframes, heal accordingly. (Best example is Maloriak)
On top of that, if you adjust the Burst Time model to assume you start from a certain %HP or HP deficit, you only need to adjust h (the tank HP) and the model remains relatively valid.

The problem with TTL (and TankPoints, which is a similar model) is that it is based on ending in your untimely demise.
A perfect model takes into account:
1) (incoming) HPS >= (incoming) DPS over an entire fight
2) In any given timeframe t -> t + dt, where HPS < DPS, you need dt < HP(t) / ( DPS - HPS )
3) At any time when you drop below a certain x %HP, HPS is increased until you reach a 'safe' y %HP as a result of healers responding to your situation.
4) If you are being healed by a Shaman, an increase in HP reduces the healing you receive at a same amount of deficit HP.

Let's take (1) as a given, assuming both you and your healer(s) are appropriately geared for the encounter (with the exception of 10man Hard Modes, which is statisticly impossible half the time). We can also ignore (4) or atleast link it to the Shaman's gearing rather than the tank's.
That leaves us with (2) and (3).
We can state that during those timeframes in (2), one dodge/parry extends the frame with atleast (at 0 HPS) 1 swing duration.
Similarly, Block has a 30% chance and Crit Block has a 60% chance of extending this frame with the same length as above.
At the same time, HP extends this time frame by HP/DPS seconds.
Adding the chance of such timeframe happening, Burst Time is still a close approximation to this situation, with the lack of HPS as it's major flaw. Because Block both reduces HPS needed as well as smooth out damage to prevent (3), it's probably the best option at nearly any point in gearing. When looking at EHP vs full avoidances, there needs to be a balanced model, where again Burst Time is technicly the best model we currently have.

If someone has a better model, I'm more than interested, would create one myself if I had the time.