Introduction to Melee Formations

The old adage, which we've all heard, goes like this:

If the tank dies, it's the healer's fault.
If the healer dies, it's the tank's fault.
If the DPS dies, it's their own damned fault.

I take exception to this adage, and think it conveys a bad attitude and is disrespectful to damage dealing classes. If the tank dies because he's wearing spellpower gear and is not talented into mitigation or avoidance, it is not the healer's fault.

This is just one example where the adage fails. We need to be honest about survivability.

If the point of the game were merely to stay alive, then a tank and four healers would be the optimal group composition. However,

The first duty of a tank is not to stay alive, but to keep their party alive and buy them time to bring down the quarry. Staying alive is merely a habit in keeping with that primary goal.

Every member of the party is responsible for accomplishing the overall goal, which is to bring down the quarry. The responsibilities of the different roles can and do overlap.

For example, if the tank dies, it might be the damage dealer's fault. Whenever a mob parries a frontal melee attack, they get a ~40% haste buff on their next white damage attack. This means if you're standing shoulder-to-shoulder with three melee damage dealers, and none of them are expertise capped to remove the 12-15% parry chance, you will be taking a lot more damage than usual.

Yes, it is one of the melee DPS's responsibilities not to contribute damage to the tank. This further disproves the adage above.

If one of your damage dealers is a Fury Warrior with 0 Expertise skill, and you're trash tanking four mobs, each of their whirlwind attacks (assuming a 13.5% parry chance on each of the four mobs) are ~44% likely to proc the haste debuff on one of the mobs.

This is why you, the tank, should always be on the other side of a mob from your melee damage dealers. Always. Mobs can't parry attacks from behind them, so the damage dealers will be much more effective if you coordinate to keep the mob's backside pointing at the party (unless it's a dragon).

Once we accept that the tank and the melee damage dealers should always be on opposite sides of a mob, the tank becomes partially responsible for where the melee damage dealers are standing. When the tank moves, the mob moves, and when the mob moves, the DPS must also move, in order to stay behind or to the side of it. The question for the tank becomes where shall I place my melee DPS?

Acknowledging that some fights are just plain messy, accidents happen, and nobody is perfect, we should still do our best as tanks to keep our melee damage dealers safe. For starters, we should not put our melee DPS in any place where:

1. They can body pull.
2. Their abilities can pull.
3. They can be flung into a mob or other damage source.

If you do any of these things, then you're forcing the DPS to choose between not engaging in the fight until the tank moves somewhere else, engaging and likely getting themselves killed, standing shoulder-to-shoulder with the tank and maybe getting him killed, or not using the abilities that define their class and role. It's a no-win situation for the melee damage dealers, and it does not contribute to the primary objective.

So, how can you, as a tank, avoid some of these common pitfalls? The answer is by planning and practicing melee formations. Over the course of the next couple of weeks, I will begin a series of short articles describing various melee formations--general positioning strategies for party members that are not encounter specific. The goal will be to outline the advantages and disadvantages of each formation, and to provide videos that illustrate them in use with melee damage dealers. Because other websites such as TankSpot have encounter-specific raid videos, I will focus on formations for general use.

Melee Formations

My promise to you

This is a "meta" entry; an editorial post about this publication.

I was disheartened to read one of my favorite druid blogs today, which strayed completely from the type of content that made me want to follow it and link to it, but instead wandered into a thoughtless, unhelpful, uninformed, unneeded and most unkind mire of off-topic and offensive "tripe" (to borrow from one of the comments). The post first upset me, but then enticed me to step back and try to salvage something of worth from my initial reaction.

Many years ago, a dear friend advised me to "THINK" before publishing any written work, and to be certain that the written piece was:

• Thoughtful
• Helpful
• Informed
• Needed
• Kind
  

I've tried to live and to write by that standard ever since.

In light of that fact, the point of this blog is not to "entertain" a daily audience or to "stimulate" any readers that find their way here during their search for relevant content on playing a druid. Our readers don't need mere stimulation--we have all of Azeroth to excite our senses.

The point of this publication is to provide relevant and useful information for druid tanks. My promise to you, the reader, is that I will not waste your time by forgetting this. When I have nothing useful to contribute, I will not write. When, as is the case today, I have more entries in 'draft' state than published, it is because the entries can all benefit from revision.

I look forward to seeing more on-topic, useful content elsewhere in our community of writers, and am happy to wait for quality. Finally, I hope that the wayward writers I look up to can find their bearings and get back to basics soon.

Chance to Crit vs Multiple Mobs

Assuming we never miss, and are never blocked, parried, or dodged, how often will we be able to proc Savage Defense while using Swipe versus multiple mobs?

To find our chance to crit on a Swipe during a trash tanking scenario versus numerous mobs, we need to understand how the combined probability of our base crit chance stacks up.

Put another way, a 33% chance to critical strike on an attack doesn't yield a 99% chance to critical strike when attacking 3 mobs simultaneously. The actual number is closer to 69.92%, because of how combined probability works.

The formula for calculating crit chance versus multiple mobs is actually the following recurrence relation:



Where n = number of mobs, and c = chance to crit versus one mob.

This formula yields the following values for a range of base critical strike chance:



See also: spreadsheet for critical strike chance versus multiple mobs.

Expertise and Crit: The Beer Analogy

Okay, yesterday's post on Expertise might have been too cerebral. I like analogies. Here's a fun one.

The right mix of expertise and crit is like a well drawn pint. If you have too much crit but have ignored the rest of your attack table, you're effectively tanking with this:



Have a fun tanking weekend, everyone!

How +hit can affect your chance to crit in a single-roll system

+Hit does affect Crit in the single roll system. You can have 50% chance to crit, but if the total block + dodge + parry + miss chance on your attack table versus a given mob is 55%, then 5% of your crit chance is wasted.

Fun Fact: I also play a level 80 Fury warrior.

Consider the following completely imaginary attack table for a dual-wielding L80 warrior fighting a raid boss, with 50% crit chance and 0 expertise or hit. Assume the target mob is wearing a shield and has a 12.5% chance to parry and a 6.5% chance to dodge.

Hit:     n/a
Crit:    01-36
Block:   36-55
Dodge:   56-62
Parry:   62-73
Miss:    73-100

The warrior will damage the enemy 36% of the time he attempts to attack it, and 100% of that damage will be from crits. 14% of his chance to crit is completely and utterly wasted.
Crit fills up your chance to hit, but does not itself push block, dodge, miss or parry off of your attack table. So, without adequate +hit and expertise, your crit chance can actually suffer.

Let me repeat that: in a single roll combat system, you need +expertise and +hit in order to maximize your chance to crit. A 4 cup can of whoop-ass can't fit into a 2 cup Mug O'Hurt.

Now, let's put this into perspective for bear tanks. We automatically hit much more often than a dual wielding melee class would. We don't dual wield, so spare ourselves an additional +19% chance to miss. Awesome. We also talent into our first 10 expertise rating. Expertise pushes both parry and dodge off of the attack table, so is twice as valuable for threat generation (per 1% of effect) as hit rating, up until we hit the "soft cap" for expertise and completely push the 6.5% dodge chance off of our attack table. After that, expertise is exactly as valuable as +hit in terms of increasing the size of our crit envelope. However, it's still more valuable for tanks than hit, because it also improves our Effective Health.

Expertise contributes to our effective health? Yes.

Reaching the hard-cap for expertise so that your attacks can also not be parried by raid bosses is important for a tank, because when a mob parries a frontal melee attack, it gets a speed buff, reducing the time to its next white damage attack.

This is sometimes called the "Tank Jib". Think of it like a backdoor crit.

Even if you never suffer a critical strike, this speed buff can give the mob a big DPS increase and tax your healer, particularly if the mob is the slow swinging, big damage type. This is also a major reason why your melee DPS should never ever ever stand shoulder to shoulder with you unless they are hard-capped for expertise, meaning none of their attacks on raid bosses can be dodged or parried. They should stand behind the mob, because attacks from behind cannot be parried.

The bottom line for feral tanks is this: If you haven't pushed parries completely off of your attack table, you can benefit from more expertise. Unless 100% of your white damage attacks are critical strikes, and you are hard-capped for expertise, you don't need more +hit. Hitting hard cap on expertise will help your threat generation and, especially in 3.1 (if Savage Defense goes live), will also contribute to your mitigation. If you overemphasize crit and ignore the rest of your attack table, you can undermine your threat generation and survivability. In 3.1, if you overemphasize the rest of your attack table and ignore crit, you do the same.

TankPoints 2.8.2 Update / Workaround

This meta post is an update to TankPoints Addon v2.8.2 - Helpful Hints for Feral Druids.

As I threatened to do, I cracked open the lua files in the TankPoints addon, and recognized that Whitetooth is using the Ace3 framework, with a new architecture that can integrate TankPoints into his other wonderful addon, RatingBuster.

The good news is that there is a version of RatingBuster out there that correctly accounts for the changes to Feral Druid armor contribution from items made in 3.0.8. After I updated RatingBuster, TankPoints began reporting the correct values as well.

Painless! My goggles go off, once again, to the magnanimous Whitetooth for this elegant design.

As an aside, I refrain from using Rawr or Pawn or any other addon that lets me set my own numerical weights to the different combat ratings, because I don't want hypothetical numbers and values. I want the correct values, specifically the reverse engineered values that Whitetooth has produced through empirical testing, and I want my talents, glyphs, current form, diminishing returns, and level to all be accurately accounted for when determinining the impact a stat change will have on me at go-time. I also want transparency in the methods used to perform the reverse engineering.

This optimal data is precisely what RatingBuster and TankPoints bring to the table.

TankPoints Addon v2.8.2 - Helpful Hints for Feral Druids

I nearly had a heart attack on Sunday when I got the Hungering Greatstaff from a quest in Taunka'le village in Borean Tundra.

My TankPoints Addon suggested that I'd be getting an additional 4798 tank points from it, mostly stemming from the nearly 5191 armor. /boggle. Here's the screenshot:



The prospect of 5191 more armor from a weapon was, needless to say, mind boggling, and I beckoned Miss Direction over to have a look at it. She did so, and with a pat on the shoulder, kindly reminded me that the 370% additional armor contribution from items that Druids in v3.0.9 gain from Survival of the Fittest doesn't count if the armor is on your weapon. Only from cloth and leather items.

Just exactly like it says on the talent text.

Head firmly reattached to my neck, I decided just then to make a short post about this minor TankPoints bug, along with another tip that I hope will save my fellow Feral Tanks some headaches later on:

Tip for TankPoints and RatingBuster users:

When you're using TankPoints to evaluate Bear gear, remember to switch into Bear or Dire Bear form. The addon calculates the effective health values based on your current form, so you might miss out on a good piece of tanking gear if you evaluate it while in Cat or Caster form. The same rule applies if you're shopping for gear at the auction house.

Here are a couple more screenshots evaluating the exact same staff while in cat and caster forms, to illustrate the point:








The moral of the story is to know your addons and keep alert--they're a useful and helpful metric, but are not infallible. Don't follow any metric off of a cliff. All of that aside, I absolutely love both TankPoints and RatingBuster, and use them every day.

To make myself feel better after this, I slunk over to Zul'Drak and, with the help of a guildmate, retrieved the Staff of the Sorrowful Chieftain.