1. Your point regarding block contributing more mitigation as we increase DR is misleading IMO. In all cases in your example you block 500. Thus in all cases your contribution to overall damage reduction is 500/10000 = 5% or the equivalent of 5% more damage reduction.

2. Your point on item points and the fact that getting more and more armor to increase your DR is prohibitively expensive when compared to puting points in another stat like block is a good one. Imagine if you had 70% DR and could get 5% more from block. It would take alot more item points to go from 70 to 75 via armor.

3. Remeber that Block becomes more important as the average hit value goes down. So two mobs both hitting for 5000dps on you, one hits for 5000 with an attack speed of 1 sec while another hits for 10000 with an attack speed of 2 sec- blocking 500 on the former gives 10% mititagion while only 5% on the latter even though the incoming dps is the same.

4. But you have to put point #3 in context of your block % and your shield block talent. Because as the hits get faster your shield block talent will cover a lower percentage of hits and your normal block rate will come into play. Its easily graphable assuming perfect use of shield block talent and given a block rate number that as hits get faster, the contribution to total mitigation from block will decrease. But also remember when i say hit speed i really mean "non-avoided hits" since we have to remove misseed, parried or dodged hits before doing any of these calcs.

5. Upshot of #4 is really that there is a sweet spot for what incoming attack speed allows for block to contribute the most to mitigation. On slow, big hits, the most imp thing about block will be shield block talent preventing crushings and the spike damage they create. In these cases a separate gearset which de-empasizes block number/% and focuses more on stam/avoidance/armor would be in order.

6. remember that after block is accounted for you then need to remove 10% more of the damage for defensive stance as well if you are building a model.

7. Check the warrior T4 or T5 set bonuses. One of them adds 100 block to one block after using shield block talent. That itself will give us a very nice mitigation boost in many cases.

Here is a nice thread where i got into a discussion around shield block and some other warrior vs druid tanking stuff.
http://beta.worldofwarcraft.com/thre...06834&sid=1#32

Here is an excerpt pasted from there where i wrote about shield block:

2. One way to incorporate Shield block into Melee DR calculations would be as such. You need to create a model and then modify the model based on your assumptions and maybe come up with average numbers tho they might be misleading and should be backed up with an explanation of your train of thought.
2a). The first thing to note is say a warrior has 25% Dodge, 25% parry, and 25% Shield block %. What percent of hits are blocked assuming he never uses the shield block ability? The answer is not 25% because 50% of Attacks are either dodged or parried. So if there are 100 attacks, 50 are dodge/parried, and of the remaining 50, 25 are blocked. So its actually 25/50 = 50%. And actually the mobs missrate needs to be incorporated so it would be 25/(50-number of misses). With a 5% miss rate this would be 25/(50-5) = 55.5% of hits are blocked. This is a much different number from the 25% base block rate so its important to use this line of reasoning. Finally the active use of the shield block talent not only decreases crits/crushings, but it will buff this 55.5% number. By how much is more complicated to figure out, tho a reasonable model combined with field tests could be done. And note that as you change the Avoidance (dodge%+parry%+mobs missrate) number you have to recalculate the effect of shield block each time.

2b). Once you have an estimated % of hits blocked, you need to estimate its effect based on the size and speed of incoming melee hits before applying damage reduction. Shield block will have a larger contribution to smaller hits and a smaller one to larger hits. So given 50% avoidance, 25% block, block number of 300, 60% DR (before def stance), and incoming hits of 10k before mitigation, you would have: 10k * .4 = 4000 hits before block minus 300 = 3700 for each blocked hit. So hits are reduced by 300/4000 = 7.5%, Mitigation for hits is increased by (50%*300)/10,000 = 1.5% (these 2 numbers are most relevant when considering spike damage reduction IMO, esp when considering the crit mitigation of shield block), and Overall effective mitigation is increased by 1.5%* 50% = .75%. These numbers change a good amount as you decrease the size of the average hit or change the avoidance or DR numbers. While the .75% number seems small it should be mentioned along side the other ones to give context.

No, it's not misleading, Todes is 100% correct. Since block is applied *after* armour, it's % mitigation goes up as armour goes up.

For instance, with a block value of 500, and a 10k hit, with 0% and 50% mitigation from armour:

0% mitigation means you take 10000, then block 500 to take 9500, which is (500/10000)*100 = 5% mitigation from block.
50% mitigation means you take 5000, then block 500 to make 4500, which is (500/5000)*100 = 10% mitigation from block.

I think this is a matter of semantics. I believe the original point was that in the 10k/500 blocked example, the block provides 5% more mitigation when you consider it as an extension of the armor mitigation:

0% mitigation from armor + 500 off block = 5% mitigation net from armor + block
50% mitigation from armor + 500 off block = 55% mitigation net from armor + block

In other words you can add the blocked percentage of the raw hit to the armor mitigation, if you consider them both to be the same 'layer' of mitigation. Whether or not that interpretation is more useful is more questionable, though.

Yes exactly. To me it is easier to compare apples to apples this way and is more useful in the context of opportunity cost analysis (i.e. how much more armor would you have to stack to get the same effect, even though the effect is applied in a different way/layer). The same thing can be done with the application of defensive stance, except that with defensive stance the effect is normalized in a way since it diminishes with increased base DR.

The fact that you can look at block as a base 5% contribution irrespective of armor is true.

The problem that you run into is that you begin to evaluate any percent as being equally valuable at any point in the continuum. Thats a bit of an esoteric statement so let me illustrate with an example. Its a bit absurd in terms of realistic in game occurrence but I feel it highlights how a percent can significantly impact total mitigation.

1) Moving from 0% mitigation to 1% mitigation. What is the net effect in damage reduction on a 10k blow?

10,000 - (10,000 * 0)= 10,000
10,000 -(10,000 * .1)= 9,900
9900/10000= 99% (or 1% additional damage mitigated)

2) Now lets look at moving from 98% mitigation to 99% mitigation

10,000 - (10,000 * .98)= 200
10,000- (10,000 * .99)= 100
100/200=50% (or 50% additional damage mitigated)

This is precisely why armor was thought to have diminishing returns. It takes more and more armor to get less and less of a percent mitigated because as you approach 100% the value of an additional percent becomes greater and greater in terms of "real" mitigation.

Additionally, I illustrated block in the manner I did to emphasize its strong positive correlation with increasing other mitigation stats. No other mitigation stat has a positive correlation with increasing other mitigations statistics. Parry and dodge both suffer from being multiplicatively combined so you tend to lose %mitigation as you push these stats. Armor has no bearing on either parry or dodge and consequently can be classified as having zero corellation. Only block receives a benefit from the other two statistics in that Armor increases the % blocked in terms of damage mitigated, while parry and dodge help increase the percentage of blocked attacks assuming the use of shield block.

Edit: Fixed the example.

I think the take away is that the net damage (after armor and other mitigation) from the hit, is what really matters when you're computing block value's overall contribution to mitigation.

I could argue that the one mob hitting for 1000 dps but attacking second has a very different profile vs a 1000 dps mob that attacks every 3 seconds. One mob will have it's dps severely cut vs a high block value / high block rate warrior, the other one will not.

Block will never be a consistent % mitigation value, so no, you can't easily model it. Unlike armor, it is not going to be consistent from hit to hit, as there will always be some variation in the other variables (particularly when you consider 15% armor gain effects from healing spells, which won't always be up).

I would argue that regardless of how you're trying to model it, block is becoming a bigger part of warrior mitigation and that it is a significant number today (upwards of 5%). Druids have greater armor level to provide mitigation, warriors have a combo of armor + block.

Looking out total avoidance rates from Patchwerk I would think that Dodge + Parry + Miss chance are all on the same hit-table and thus together give you your total complete avoidance. I'm not 100% sure how block is then factored in. (Looking at average runs and such is hard since I tend to use shield block frequently and that pushes the number around.)

Blocking is an interesting stat indeed. This thread has been very fullfilling to see. I also doubt that Blizzard will give us more than 50ish percent avoidance in total since avoidance doesn't have the same scaling function as armor does. Going from 50% -> 75% avoidance doesn't cost as much itemvalue as going from 50% to 75% armor does. Hence the possibility to use that.

I've put together an imaginary profile of +block value gear(http://ctprofiles.net/4984186). It looks like the best you can do unbuffed is 399 (without activating a trinket)... that's assuming ctprofiles doesn't calculate 30% block value talent correctly, which I don't think it does.
TBC introduces new slots with +block value like back and finger, which should allow us to go much higher.

I was a great fan of custom maps back in Warcraft 3, designing a few, testing/balancing others and trying out hundreds. One of the most common mistakes I came across was not taking mitigation into a context:
An ability which reduces 10% damage per level might seem balanced at first glance and the imbalance won't be striking if it only has a few levels. If the players get the ability to take it to 50% (or in some of the worse cases, 70%) the imbalance becomes striking though and you start considering diminishing returns. Instead of 10% per level, the relative reduction in damage for each point was like this: 10%, 11.1%, 12.5%, 14.3%, 16.7%. As you can see, growing exponentially better to the point where players were getting virtually immortal.

In WoW, it also seems important to note the context. Why not take armor into account for blocking, if we can assume every tank to have between 50 and 75% mitigation before blocks are applied?

We have mitigation, avoidance and blocks as the 3 primary factors. There's also resistance though it doesn't interact with any of the others:

The relative value of blocking grows with increased mitigation. Reducing 50% damage before blocking x means that x is twice a big a share of the incoming damage.
The relative value of mitigation grows exponentially with more mitigation. The increase from 0 to 1% is a relative increase of 1%. The increase from 50 to 51% is a relative increase of 2%. The increase from 98 to 99% is a relative increase of 50%.
The relative value of avoidance grows exponentially with more avoidance. It's similar to mitigation.
Resistances follow the same model as mitigation except it's even more backended, with normal mitigation going from 0 to 75% from resistance 0 to max (linearly, I think) and an additional avoidance factor being added at the end. At level 60 I think it's like 1% avoidance at 250 resist and 25% at 315.

Notice that for the absolute values of mitigation and avoidance, the value of each decreases as the other is increased. 40% dodge doesn't make as great an impact if you mitigate 60% of the hit anyway. 60% mitigation doesn't make as great an impact if you dodge 40% of the hits anyway. The relative benefit is the same though so this shouldn't be a concern.

Also note that while it can be beneficial to stack avoidance or resistance, mitigation doesn't work the same way. The exponential increase of mitigation is considered in the diminishing returns formula so adding an extra point of armor will always have the same relative benefit. The difference is that you can kind of stack the percentages directly for the other attributes.

While I think this covers the basic relations, there's still a lot of the context which isn't considered. To give an example:
a) A mob hits you for enough damage to oneshot you. Adding 10% mitigation would guarantee that you survive the hit. Adding 10% avoidance wouldn't have nearly the same effect.
b) The mob hits so hard that you can't get enough mitigation or the old mob hit you once and you didn't get healed in time. At this point 10% avoidance (and prayer if you're into that sort of thing) will serve you better.

A few other things which come to mind:
Relation between avoided hits and shield block charges.
Relation between mitigation and avoidance in terms of consistent rage generation.
Relation between blocking, shield block charges and mob attack speed.
Relation between avoidance and healer attention/5 sec rule (on PW, a few free seconds can save another tank at times).
Relation between block value and shield slam damage/threat.

I think it would be relevant to try and plot the relative value of Block Value and it should be reasonably easy to give a better estimate than has been done in the past but this game is too complex and situational to make perfect formulas possible for this sort of thing. Just be sure to take its relation to other damage mitigation into account or the data will be practically worthless.

My intention here was to say that due to the fact that parry and dodge are multiplicative(Chance to be hit=(1-parry%)(1-dodge%)). That as you push these stats higher you tend to encounter diminishing returns. I'm leaving miss out deliberately. Example:

At 1% parry and 1% dodge
(1-.01)(1-.01)=98.01% chance to be hit or 1.99 chance to avoid. (A minimal loss of % mitigation compared to the original starting values if they were figured additively)

At 25% parry and 25% dodge
(1-.25)(1-.25)=56.25% to be hit or 43.75% to avoid (Here you are losing 6.25% of your avoidance as opposed to an additive comparison)

At 50% parry and 50% dodge
(1-.5)(1-.5)=25% chance to be hit or a 75% chance to avoid(Here you are losing 25% as opposed to an additive comparison)

My point here was not to say that these values are bad but that they tend to suffer from diminishing returns as you push them higher in the sense that item points invested do not return the same amount of mitigation as percentiles rise. Ideally you would invest all your item points into one or the other, but this isn't a reasonable goal.

Also, I think that while you could apply the argument higher percents have higher value as I did on my block argument, I think one caveat is important on avoidance mitigation. Burst damage kills tanks. I think you can push hourly mitigation higher with higher avoidance mitigation stats but I feel you lose if the standard deviation of the damage you take rises.

Finally thank you for pointing out the fact that no block related gems exist. That was a poor oversight on my part.

It is my understanding that there is substantial evidence to refute your belief, but I could be wrong.
Edit: I think most people follow the "church of attack table" (http://www.wowwiki.com/Attack_table).

That's true if you can roll both parry and dodge at the same time, but you can't. They're additive.

The numbers I used were parsed out of CombatMonitor, but they aren't nearly as good or as extensive as the work done to prove the table theory. Lacking better evidence to the contrary I will concede the point.