On the Effectiveness of Redoubt versus Single Mobs

[moduspwnens]

Effectiveness of Redoubt versus Single Mobs

After seeing countless knocks of Redoubt, I decided to do an actual study of how much it would help given various levels of avoidance. Currently, while main tanking, it does nothing in most cases, as generally Tankadins gear with avoidance and block high enough that with Holy Shield, any extra block % is irrelevant. This is essential when mobs land crushing blows, but current Beta reports say mobs three levels above the player (raid bosses) will no longer crush.

So who cares, right? Well, assuming avoidance levels stay the same, we shouldn't. Holy Shield would stay up all the time anyway, so that extra block % from Redoubt would make no difference. There's no way we can know, though, until we see the level 80 raid gear. There is reason to believe that avoidance and block levels won't reach 70% versus a level 83 mob, and here they are:

Sunwell Radiance
This was put in place by Blizzard in Sunwell to balance high avoidance tanks in Sunwell.

Difficulty of Encounter Tuning
It's very difficult to balance tank damage when the tank has high avoidance. After all, if he just hits harder, the tank is more likely to be gibbed. It's much easier to tune the difficulty of Sunwell encounters because it's expected that the tank will be taking a hit at least 50% of the time.

Death Knights
Due to Death Knights, it's unlikely we'll see any block rating (they can't block) on anything other than tier gear. Defense is also a little inferior to them compared to warriors and paladins, but we can certainly expect that to be on gear.

Warrior Shield Block Changes
If it is to be the case that passive avoidance and block levels are at 70%+, warriors will be at a disadvantage. Paladins will be able to block every hit (much as they do now), but warriors can only block for double every once in a while (plus their passive block, whatever that ends up being). Otherwise, they'll take normal hits.

20% Attack Speed Reduction to all Tanks
Doesn't directly point toward avoidance levels, but because this makes a much bigger difference when avoidance is low (think Sunwell bosses with and without Thunder Clap), it's a possible indicator that avoidance will be comparatively low and tanks will be more even in the new tough 10 mans.

...but here's the meat of the argument.



I learned some basic-to-moderate Java in my classes last year, so I developed a simple program that allows you to put in your stats (miss, dodge, parry, block with Holy Shield), and then simulates a swing from a boss based on those values. The finished product allows you to enter how many swings you want to simulate, and then summarizes and averages them at the end.

It assumes the following:

A single attacker.
Attack speed of 2.0.
Redoubt lasts five swings (10 seconds).
Blocks are partial.
Redoubt procs on both full hits and blocks.

I could have made it work for multiple attackers, but we already know its power with them, so this thread focuses primarily on Redoubt versus single targets.

So, here's some of the test scenarios and results:

Test Case Scenario A
In this scenario, I plugged in my normal avoidance levels from my armory. If everything is working correctly, there should be no normal hits, as I am uncrushable with Holy Shield. I'm also using values versus a level 70 mob.
Miss: 12%
Dodge: 25.81%
Parry: 18.27%
Block: 53.41% (23.41% + 30% Holy Shield) (43.92% fills the combat table)
I simulated 10,000 swings. The results?


Misses: 1168 (11.68%)
Dodges: 2602 (26.02%)
Parries: 1838 (18.38%)
Blocks: 4391 (43.91%)
Hits: 0 (0.0%)
Blocks (Redoubt): 0 (0.0%)
Redoubt Procs: 421
Just as I expected. There were no normal hits, and although Redoubt proc'd many times, no blocks came from the 30% block given by Redoubt. Also, the averages for each are fairly close to those I input at the beginning.

Test Case Scenario B (Warning: 1.98MB)
For this one, I answered a question I've had for a long time. For simplicity's sake, by the stats on my armory, when I'm tanking Kalecgos in Sunwell, how many blocks does Redoubt get me? Even though I'm uncrushable outside of Sunwell, it's generally not worth sacrificing stats to become so in Sunwell (aside from versus Twins). I'll be applying the penalty versus level 73 mobs. So here are the stats:


Miss: 6.4% (12% - 5% - 0.6%)
Dodge: 5.21% (25.81% - 20% - 0.6%)
Parry: 17.67% (18.27% - 0.6%)
Block: 52.81% (23.41% + 30% Holy Shield - 0.6%)
Regular Hit: 17.91%
For this, I simulated 100,000 swings. This is something I wanted to be accurate (although it's not a big deal, it takes less than ten seconds). Here are the results:


Misses: 6434 (6.44%)
Dodges: 5185 (5.185%)
Parries: 17618 (17.618%)
Blocks (from Passive + Holy Shield): 52812 (52.812%)
Hits: 12352 (12.352%)
Blocks (Redoubt): 5595 (5.595%)
Redoubt Procs: 7090
So there's our answer! In my gear, in Sunwell, I can expect Redoubt to account for a little over 5.5% block, while without it I would have been stuck with that 17.91% full hit rate (now reduced to 12.482%).

Test Case Scenario C
Now, let's get a test case that would be a best case scenario for Redoubt (versus a single target). This would mean no avoidance. It will have a chance to proc on every single swing.


Miss: 0%
Dodge: 0%
Parry: 0%
Block: 0%
Regular Hit: 100%
I'm just gonna do 10,000 swings to save myself some hard drive space. Here's the results:


Hits: 8774 (87.74%)
Blocks (Redoubt): 1226 (12.26%)
Redoubt Procs: 1003
So, in a best-case scenario versus single targets, Redoubt will give around 12% chance to block.

Conclusion: Nobody knows where our gear will put us, but if we're at Sunwell levels of avoidance, which I have argued is probable, we can expect a little over 5% block, which is on par with the Warrior talent, Shield Specialization. If you have a set of hypothetical avoidance values you'd like for me to run, post here and I'll run them.

[moduspwnens]

I thought about putting this in the Advanced Training forum, but realistically this doesn't apply (due to crushing blows) except in Sunwell and perhaps in WotLK, so I thought I'd put it here. Let me know what you think.

[QuantumDelta]

2.7975% vs Brutallus then?
0% Twins
Insanity numbers Felmyst (wonder if it's worth simming 20 mobs)
Distinctly average @ M'uru (Doors)
Random @ KJ

[moduspwnens]

2.7975% vs Brutallus then?
0% Twins
Insanity numbers Felmyst (wonder if it's worth simming 20 mobs)
Distinctly average @ M'uru (Doors)
Random @ KJ

I don't know where you're pulling your Brutallus numbers from, and Felmyst and M'uru would involve tanking multiple mobs (where we know it's valuable), and for Twins, you would presumably be uncrushable.

Redoubt would likely be more effective on Brutallus, as he attacks much faster than the mob simulated. I guess as far as Sunwell, this simulation would likely only be accurate for Kalecgos and Felmyst herself.

[QuantumDelta]

doubled the number of attacks and completely forgot that it's more likely to actually be quadrupled are you sure it'd be more?
It was 4am when I posted that and it's 5am now and I'm pulling an all nighter and wont be able to sleep for a while
So the math part of my brain's in total shutdown mode I guess :<
Since you questioned the number I guessed at I've completely lost the plot :<

[moduspwnens]

doubled the number of attacks and completely forgot that it's more likely to actually be quadrupled are you sure it'd be more?
It was 4am when I posted that and it's 5am now and I'm pulling an all nighter and wont be able to sleep for a while
So the math part of my brain's in total shutdown mode I guess :<
Since you questioned the number I guessed at I've completely lost the plot :<

It's pretty late here, too. I'll think about it in the morning, or wait for Jere or Dorvan to tell us the right answer.