Warriors and Paladins Block About the Same in Build 8905

[Garath.Gorefiend]

Attention:
This has been updated to use a different summary system than it originally did. It now summarizes as (Total Effective Blocks / Total Unavoided Hits) * Your Block Value, rather than dividing by all swings.

Alright, so the simulator can be much more accurate, I changed how things are calculated. An "Effective Block" is now your block value (including from Strength) before any talents or buffs. Period.

The simulator now accepts your points in Shield Specialization/Mastery (up to 30% block value), the Glyph of Blocking (10% block value), the TBC block value meta (10% block value) and the WotLK block value meta (5% block value). In order to be accurate, it requires all this information because the modifiers don't multiply directly onto each other. Based on these modifiers, it generates an effective block. For Paladins, this is generally going to be 1.3. That's 1.3 x Your Block Value, and is what you virtually always block for. For Warriors, it tracks the modifiers and adds them up correctly, as well as doubles them correctly for the correct amount of mitigated damage.

So, let's confirm that will work. In-game, on my warrior, I confirmed that the following blocks mitigated the corresponding amounts of damage:

Standard Block: 581
Critical Standard Block: 1162 (Double)
Shield Block Standard Block: 1028 (100% increased block value)
Shield Block Critical Block: 2056 (100% increased block value + double)

I have 3/3 Shield Mastery, so 581/(1.3) is 447. This is my block value before talents and buffs. The simulator says:

Standard Block: x1.3
Critical Standard Block: x2.6
Shield Block Standard Block: x2.3
Shield Block Critical Block: x4.6

If you go through those and do those multiplications with my block value before talents (447), it comes out to the values that the game gave me. The Effective Block x Your Block Value = How much you block for. So, it appears the simulator is accurate. However, I also changed how it displays the totals. It will now display it as (Effective Blocks / Total Damaging Swings) x Your Block Value. What this does is average how much you block per swing, so that there can be an equal comparison between tanks.

So, now I'll do the exact same trials as before, so we can know who blocks more in Build 9085.

Trial 1: Minimum Avoidance, One Target
For this one, I'm using the minimal likely avoidance. The only avoidance and block %s come from talents and crit immunity through defense. The tanks are unbuffed. I also assume a single target with a 2.0 attack speed, slowed to 2.4 by Thunder Clap / JotJ, that is three levels above the tank.

Links to Repeat:
Warrior - Paladin

Averaged over 10,000 swings:
The Warrior mitigates 1.106 x Block Value each unavoided swing.
The Paladin mitigates 0.95 x Block Value each unavoided swing.

Trial 2: 50% Avoidance, One Target
For this, I upped their dodge and parry chances by 5% each, for a total of 50% pure avoidance.

Links to Repeat:
Warrior - Paladin

Averaged over 10,000 swings:
The Warrior mitigates 1.18 x Block Value each unavoided swing.
The Paladin mitigates 1.1 x Block Value each unavoided swing.

Trial 3: 60% Avoidance, One Target
For this, I upped their dodge and parry chances by another 5% each, for a total of 60% pure avoidance.

Links to Repeat:
Warrior - Paladin

Averaged over 10,000 swings:
The Warrior mitigates 1.3 x Block Value each unavoided swing.
The Paladin mitigates 1.3 x Block Value each unavoided swing.

This is very interesting. I had to run the numbers twice just to be sure I wasn't running the wrong class.

Trial 4: 70% Avoidance, One Target
For this, I upped their dodge and parry chances by yet another 5% each, for a total of 70% pure avoidance.

Links to Repeat:
Warrior - Paladin

Averaged over 10,000 swings:
The Warrior mitigates 1.48 x Block Value each unavoided swing.
The Paladin mitigates 1.3 x Block Value each swing.

This makes sense. Once it becomes easier and easier to block without Holy Shield, the superior scaling of the blocks that are left favors of the warrior.

Trial 5: 60% Avoidance, One Target, Dual Wielder
Same as Trial 3, except the attacker has an attack speed of 1.2, twice as fast as before.

Links to Repeat:
Warrior - Paladin

Averaged over 10,000 swings:
The Warrior mitigates 1.3 x Block Value each unavoided swing.
The Paladin mitigates 1.3 x Block Value each unavoided swing.

About the same as Trial 3, due to Shield Block having no charges and Holy Shield not being eaten up.

Conclusion (TL;DR)
It appears that the Warrior blocks more at lower levels of avoidance (where his 5% from talents and 100% to block Shield Block excel). The Paladin catches up and overtakes the Warrior at around 50% (likely, fully-buffed in starting tiers), they tie at 60%, and the Warrior overtakes the Paladin at 70%, when the frequency of blocking overall is lower, so the scaling of the Warrior's blocks wins.

Oddly enough, it actually appears to be kind of balanced. The Warrior excels in less likely conditions, but by a wider margin, and the Paladin excels by a smaller margin, but in the more likely conditions. Thoughts?

Modus
The error your making here is an interesting one, but I noted it myself in my simulations.

If you only mess around with dodge and parry, especially because of diminishing returns, you definitely end up with what your seeing.

Try a similar example where you maintain avoidance and change the block rating of the tanks.

Effectively, what it boils down to is that as block rating takes a larger portion of the unavoided attacks, the warrior mechanic very quickly scales up.

Paladins get to 1 and stay there

As you increase Pure Avoidance, and maintain block rating, the warrior block mechanic actually deteriorates to a point until the block rating has reached the block threshold point which starts to scale it back over 1 again.

I believe the threshold point has something to do with the magic block number generated due to the shield block talent, but am not sure and frankly havent had time to do that much on it lately.

The spreadsheets here are interesting, but only seem to model a single attacker. The simulators we have are more accurate.

Basically though, it is highly dependant on gear choices. If when you are gearing, you notice more impact from diminishing returns, then it is time to switch to gemming and enchanting for a different stat.

The advantage of the warrior shield block mechanic is that it actually overcomes much of the issue of stacking defense. All calculations in TBC showed that once you had become uncrittable that it was better to stack dodge for a warrior. This is no longer true.

Given block values and the new block mechanics, it is actually better for the warrior to prioritize Defense rating. Try it you will see what I mean.

[moduspwnens]

Yeah, I know what you mean. I was trying to err on the side of the Paladin, and it seemed clear that giving the tanks more block would only make the Paladin look worse. It'd probably be fair to say that these gear stats are in the Paladin's favor.

[snowwight]

Necromancer!

Yeah, I actually made a handy graph about that very issue Garath, versus 1 mobs and 10 mobs:
http://spreadsheets.google.com/pub?key= ... tput=image

Of course, I was modeling the old Armed to the Teeth as well. Basically we (obviously) don't scale with block rating at all on 1 mob. But as you increase block rating for warriors, you get this crit block effect - as they block more, the more compounded the 1.3*1.3 they get beats our meager 1.3.

EDIT: Forum software doesn't like google's lack of extension.

It might be interesting to plot the derivative of the scaling with block rating.

[moduspwnens]

Necromancer!

Yeah, I actually made a handy graph about that very issue Garath, versus 1 mobs and 10 mobs:


Of course, I was modeling the old Armed to the Teeth as well. Basically we (obviously) don't scale with block rating at all on 1 mob. But as you increase block rating for warriors, you get this crit block effect - as they block more, the more compounded the 1.3*1.3 they get beats our meager 1.3.

Yeah. I posted a possible fix over on the Beta forums here:
http://forums.worldofwarcraft.com/threa ... sid=2000#1

...and that'll be my last stab at the issue until I do some more in-depth simulations.

[Garath.Gorefiend]

I wonder if warriors have tweaked yet to the idea that prioritizing defense in WotLK is going to be win...

[Elsie]

Necromancer!

Yeah, I actually made a handy graph about that very issue Garath, versus 1 mobs and 10 mobs:


Of course, I was modeling the old Armed to the Teeth as well. Basically we (obviously) don't scale with block rating at all on 1 mob. But as you increase block rating for warriors, you get this crit block effect - as they block more, the more compounded the 1.3*1.3 they get beats our meager 1.3.

Yeah. I posted a possible fix over on the Beta forums here:
http://forums.worldofwarcraft.com/threa ... sid=2000#1

...and that'll be my last stab at the issue until I do some more in-depth simulations.

That might be better worded to "Your block value is increase by an amount equal to your chance to block after 50%."
If the lowest amount of avoidance you can get is 40%, and we gain roughly 10% from gear, then it's saying about the same.

As someone else pointed out, not everyone knows about attack table math.

[Snake-Aes]

It's flawed in that it may or not be more worth to stack block% or real avoidance.

It may or not be balanced/too weak/ too powerful, but it definitely doesn't have a good feeling on the modeling side.

[Elsie]

It's flawed in that it may or not be more worth to stack block% or real avoidance.

You're still confined by gear available. The only real thing you can change for your statement is tier or offtier, avoidance trinket or BV/BR trinkets. I'm not against Block trinkets being good for paladins.

Item Value just wouldn't let the situation get bad imo. Also lets our tier gear remain very different.

[Snake-Aes]

That it does.

[Jedah]

I like turtles.

Zombie Kid ftw.

edit: forgot to make this post useful - I'm glad to see the gap isn't as wide as you once suspected. I hope this holds true in practice.

[Garath.Gorefiend]

It's flawed in that it may or not be more worth to stack block% or real avoidance.

It may or not be balanced/too weak/ too powerful, but it definitely doesn't have a good feeling on the modeling side.

Wasnt talking about stacking block so much. Was restating the value of defense.

Because defense adds BR, it may well become more valuable once diminishing returns kicks in. It can definitely become more valuable if there is no diminishing returns on miss from defense rating.

However, given WotLK Block Values and diminishing returns, there is a point where stacking Block rating actually does become beneficial.

I could do the math for you...but am feeling lazy

Really wish I knew if there was diminishing return on Miss and, if so, what the Cap is.

[Bundy]

I realize this is kind of old, but I was reading it and was wondering how the warriors glyph worked into this. The one where you get 10% BV after a Shield Slam.

Considering Shield slam is a 8 sec CD and the proc lasts 10 secs I think its safe to say it would be up all the time.

So would it be,

Standard BV 1.0
BV with talents 1.3
BV w/talent + glyph 1.4
all above + block crit 2.8
Shield block still 2.3 ? or 2.4 ?
Shield block crit 4.6 ? or 4.8 ?

If I had to guess id say its still 2.3 and 4.6, but somehow that doesnt seem right. And would that extra 10% be from base BV or would it add from the already 30% extra meaning youd get 1.43 instead of just 1.4 ?

I agree that palys blocks should block for a little more than they currently do, but not as much as a warrior. I dont think they should have more overall health than warriors. Your kings buff makes up for any lack of health and then some.

Granted everyones getting kings most likely in a raid situation, but you have to look at it from an idividual point of view as well. In a 5 man there might not be a paly to buff a warrior tank with kings.