It's a well know fact that using sheild block prevents crushing blows from landing on a tank (see thread http://forums.elitistjerks.com/viewtopic.php?id=6726 for a quick summary of how physical attacks work).

A given attack from a level 63 raidboss has a 15% chance to be crushing. Spamming sheild block usually reduces the crushing blow rate to around 10% or a 1/3 decrease in crushing blows. However, if a player was able to stack enough block/dodge/parry/defense put their combined block/dodge/parry/boss miss over 85%, crushing blows would start getting pushed off the hit table.

With current itemization this is now possible.

http://ctprofiles.net/1833640

The profile contains the best evasion gear I can find, totaling 45% block, 20.6% dodge, 16% parry and 125 defense (resulting in [125-15]*0.04 = 4.4% increased miss, coupled with the boss' base 5% miss gives 9.4% miss). This leaves the player with 91% base evasion. Throw on the Hakkar buff (5% dodge and 10% stats), elixir of greater agility and kings and you are up to 99.1% (I'm assuming my agility goes [152+25]*1.1*1.1 from these buffs, I've never tested it so I don't know for sure).

Juju Escape can also add 5% dodge for 10 seconds on a 1min cooldown in the absense of the hakkar buff.

I'm pretty sure what this means is if you were able to collect this set of gear and buffs, not only would you be uncrittable, you would be nearly uncrushable.

It will be interesting once tanks have full Dreadnought (even though your profile only has some of it) and see what sort of damage they are taking.

In practice it doesn't quite work this way, though. If you hit shield block with 15% base block chance and end up with 90% block, 20% parry, and 20% dodge, you can still get crit/crushed. We've had long discussions of this and I have yet to see a real explanation of how these mechanics truly interact. The key point is that you need 100% block to avoid crits/crushing. Dodge+Parry+Block > 100% will not do it.

Its like the old "a missed hit can crit" debate. I suspect they roll a dice for all avoidance independant of eachother. so chance to avoid (parry or dodge) an attack would be P(parry) + P(dodge) - P(parry and dodge).

As I understood it, the "block" roll modifier is the last table and is applied *after* the hit/miss/dodge/parry table... So on that basis, you'd require dodge/parry > 100% to exclude crushes/crits purely with evasion gear.

-edit- Rodent was faster, but yes.. could be that each avoidance is a seperate roll table also, perhaps testable with a Rogue/evasion?

Have you read Satrina's guide on Hits, misses and crits reguarding this?

http://evilempireguild.org/guides/

There is a lot of really good information on the site about the mechanics of melee combat in WoW, including a compelling arguement that what you are saying is exactly not what WoW does to calculate combat.

It seems odd that Dodge+parry+block >100% would not prevent crushing and block >100% would. Especially when you consider crit > 100 (recklessness) does not eliminate miss/glancing/dodge.

I guess I'll have to test it out tonight to see if I can get a mob to crush me while I have sheild block up, but less then 100% total block.

Edit: I'm also sure that I've dodged/parried my way through a sheild block without blocking once. However, I don't think I had >100% block at that time.

Sorry for being unclear before, but in my mind there is no debate about "missed hit can crit".

http://forums-en.wow-europe.com/thre...rchive-en&t=15

You can't crit a miss or miss a crit. There is, however, a priority order in which different abilites are added to the hit table.
http://forums.elitistjerks.com/viewtopic.php?id=6726

This is the information that I was working off when I assumed dodge+parry+block+miss>100 would prevent crushing.

Even your information says you are wrong though surely? That info says dodge/parry is one roll and block/crush/crit/glance/hit are applied on subsequent rolls.

No, it doesn't. It says that's the priority order for filling the table, no more, no less. The m/d/p being 1/1b/1c is because no one's been able to determine the priority between the three of them.

Yea, sorry about that. I copied more of the quote to make it more clear. The reasons they are number in that way is it is uncertain of which comes first miss, dodge or parry. The reason that no one knows is because no one has every pushed their combined miss/dodge/parry > 100% to see which one starts occuring less.

Thats just the order in which each variable will overwrite the others i.e. 100% dodge 'beats' 100% crit.

Edit - Too slow :(

It does.. because Block is a separate roll.. So you can't use Base Block with Dodge/Parry to exclude crushes/crits because they are on adjacent rolls and aren't capable of totalling 100% on one roll.

There's significant evidence against the "block is a seperate roll" theory.

Sure.. but he's using the information above saying that it is :)

And for what it's worth, in 18 months or whatever of tanking, I've not seen myself crushed/crit when SB is up and I've had 25% base.. but that's not quantitative.

Do you have a link to the evidence contrary to the separate roll theory?

Now that is an interesting mechanic possibility to explain why block>100% prevents crushing and dodge/parry/Shield Block >100 does not. My normal base block is 20%, so with sheild block up I should be seeing in the neighbourhood of 5% crushing blows if this is true.

Also, if this is the reason Sheild Block doesn't eliminate crushing blows when you have <25% base block it would be good for the total passive evation build because it implies that the "weird" mechanic is with the Shield Block ability itself and not with the rest of how the outcome of an attack is determined.

Does anyone happen to have a combat parser that only records data while shield block is up?

I think the reason most believe that it is not two seprate rolls (one for block, and then one for parry/dodge/miss, etc) is that if it is, from a backend stand point that is all the more rolling that the server has to do each combat cycle.

Im sure that the servers are powerful and that 2 checks to the rng(); function is not a huge hit, but if you multiply that over all the block capable combat rolls that take place in a given day that is a huge increase to the amount of rolling.

I allways thought that block, because it has less signifigant avoidance would have the least prioirty.

(Parry is 100% avoidance plus a minus to the swing timer.
Dodge is 100% avoidance
Miss is 100% avoidance.
Block is some damage mitgated.)

Not being a warrior main I am not sure, but I assume that you can still parry and dodge when sheild block = 100% correct, so that means the other two must have higher priority so the block does not push them off the chart.

So from a design standpoint, if I increase my block chance, all I want is to have hits/crits/crushes pushed off the chart, in that order. But I dont want to push off some of the better avoidance I have. I would not want a high block chance to mean I parry less for example because that would make block be hurting my mitigation.

So with that design in mind, I think what makes the most sense is to have block back fill, not cause a seprate roll, but fill the table first.

So you start with a blank table
and fill in block:
0-25 block
26-100 Enemy does something

Grab the rest of the defenders values:
0-25 block
26-40 dodge
41-60 parry
61 -100 enemy does something

Grab the attackers values and fill in from the top.

0-25 block
26-40 dodge
41-60 parry
61- 80 Hit
81-85 - Cirt
86-100 - Crush

Now thats the sample table with no overwriting. If we increase block to 100% via sheild block, we dont want to push dodge or parry off the chart. but we still fill in the same way.

Grab Block
0-100 block

Fill in defenders values:
1-65 Block
66-80 - Parry
81-100 - Dodge

(same system, block fills the bottom, then fill in the top, however since parry and dodge have higher priority than block they overwrite the higher end of it.) The attacker cant fill in any values since the defenders values have higher priority, thus no crushing)

If its less than 100% block, say 95%

Grab Block
0-95 block
96-100 - Enemy does something

Fill in defenders values:
(If we assume that enemy does something has greater priority than dodge and parry)
0-60 - Block.
61-75 - Parry
76-95 dodge
96-100 - Enemy does something

Fill in attackers values the highest priority being crush, since it is the most powerful
0-60 - Block.
61-75 - Parry
76-95 dodge
96-100 - crush

I think this is how it works, and based on what I have seen this makes the most since from a design standpoint (of having the most important values present in the tabel and not having silly things like hits > crush, or Block > parry) and having the server only run one calculation per combat exchange.

However I dont have hard parsed numbers to back this up, Im not sure how without a very large sample size one would be able to test this.