My bad, that's what I get for posting before I wake up. Has anyone run tests yet to see if the tool tip is working properly? They may have just fat fingered the tool tip math.
Originally Posted by Stengel
Some numbers to help me understand this better (you're welcome to double check this, math isn't my strongest point):
For anyone wondering what this is about: there's a discrepancy between what the Shield Block buff tooltip and what the Shield Block ability tooltip is saying (everyone seems to have based their math on the latter apparently). We don't know which one is true, but I'm trying to understand what the differences mean.
OLD MECHANICS (Shield Block critical block chance bonus = avoidance+block-75)
Let's say our warrior has these stats: 75% avoidance (5% miss, 12.5% dodge, 12.5% parry, 45% block, 30% critical block)
Against 100 attacks of 100 damage, his damage taken is on average
25 * 100 = 2500 [hits]
45 * 100 * (1-0.3*1.3) = 2745 [blocks and crit blocks]
Total dmg taken: 5245
With Shield Block it looks like this:
70 * 100 * (1-0.3*1.3) = 4270 [blocks and crit blocks]
Total dmg taken: 4270
That makes Shield Block equal to an 18.58% damage reduction in this situation.
Now let's add 10 mastery, and we get: 90% avoidance (5% miss, 12.5% dodge, 12.5% parry, 60% block, 45% crit block chance)
10 * 100 = 1000 [hits]
60 * 100 * (1-0.3*1.45) = 3390 [blocks and crit blocks]
Total dmg taken: 4390
With Shield Block up (assuming it adds 10% block chance and 15% crit block chance):
70 * 100 * (1-0.3*1.6) = 3640 [blocks and crit blocks]
Total dmg taken: 3640
That makes Shield Block equal to a 17.08% damage reduction.
NEW MECHANICS (Shield Block critical block chance bonus = avoidance+block-50)
Let's take the same warrior as in Scenario 2, but pretend that the current Shield Block buff tooltip is correct, which means that Shield Block would effectively add 10% block and 40% crit block in this situation.
70 * 100 * (1-0.3*1.85) = 3115 [blocks and crit blocks]
Total dmg taken: 3115
That makes Shield Block equal to a 29% damage reduction.
Edit: added the formulas for each mechanic and a short explanation what this is about.
Edit2: fixed the damage reduction
Last edited by Stengel; 11-29-2010 at 02:54 AM.
Okay, I went and tested this to see which formula is correct. With Shield Block up, only 6 out of 45 blocks were regular ones and the rest were criticals. That's a 86% critical block chance which is almost the 85% that I should have (29% mastery +10% Hold the Line +46% Shield Block buff tooltip).
I tanked the mobs for a while with combatlogging enabled, then loaded the relevant sections into WoWCardioRaid to make the screenshots.
This at level 80 with 96,97% avoidance+block against lvl 80 mobs.
I tested it in sethek halls once on 2 guards when i went to kill Anzu, also got way more then 50% crit blocks. Good to see your test also confirms it. Its not a big sample size but the 25% gap is too great to be just a fluctuation.
Math looks good except the percentage reduction part. If you take 10k as 100% scenario1=9,75% scenario2=7,5% scenario3=12,75%, if you take the amount of damage before shieldblock as 100% 5245 for scenario1 and 4390 for scenario 2-3 you get
scenario1: 4270/52,45=81,41% so 18,59% reduction
scenario2: 3640/43,90=82,92% so 17,08% reduction
scenario3: 3115/43,90=70,96% so 29,04% reduction
This will improve the value of mastery compared to dodge/parry, giving mastery about 0,5-0,8% more crit block during shieldblock per point depending on the diminishing returns of dodge/parry. Lets take 0,7% extra crit block, then the average block would increase by 0,3*0,007=0,0021 around 70% block during shieldblock 0,7*0,0021=0,00147 so 0,147% damage reduction extra with 1/3 uptime 0,049% extra reduction. Pretty nice improvement if you take into account that mastery starts out around 0,71% damage reduction per point.
I see what you're saying and updated the damage reduction.
By the way, it should be possible to get 100% critical block chance with SB and HtL at some point in Cataclysm with some hardcore mastery stacking.
So in other words, Shield Block is granting 25% critical block extra on top of what the blues said it was supposed to do? Nice
i whole heartedly agree with the poster Booi, in page 9 arguing less about total incoming damage and more about tank survival.
From the study above, the maximum difference between each extreme, that is pure avoidance or max mastery is never above 1%
however, incoming damage with maxed out mastery is far far smoother, predictable and over all heal-able than a pure avoidance tank.
what i would however want to know is how the balance between mastery versus stamina affect survival rate(for physical incoming damage of course)
what i would love to see, but i am incapable of doing is an actual simulation
tank get hit every X seconds (which he can avoid, block)
tank receive heal from X healers every Y seconds
play with tank stat and healing power
follow tank health and record death time
run simulation hundreds of time get an estimated survival rate depending on stats
can also be used to compare tank survival balance.
Forgive my blindness as I've just scanned the thread to see if this was mentioned, but is there any data that takes the meta (+5% block amount) into account when comparing avoidance vs. mastery?
I'm doing some math on the effectiveness of our Meta-Gems, but I'm stuck calculating my Critical block chance.
Till now i used my World of Logs data, but I'm unsatisfied with these numbers.
Hope someone can help me asap
Critical block chance shows in the character sheet under mastery, its rounded to full percentages there so not sure how accurate that is. Then you can add hold the line, uptime from WoL * 10%.
For shieldblock the extra crit block is parry+dodge+miss+block-50, procs from windwalk or trinket uses can increase this a bit more if gained before shieldblock is used.
So its a quite hard to calculate an average Crit block chance in a fight of 5 Minutes.
Finally i took my WoL data and did my math with it.
When I'm finished translating my stuff, I'll post is here
There it is:
Last edited by Shaizhowfn; 01-08-2011 at 01:44 PM.