Mindbender & Shadowfiend

[aica]

How exactly does Mindbender differ from Shadowfiend?

When I look at Simcraft results, which has 25.55% haste at level 90, I see Mindbender and Shadowfiend both attacking 13 times per cast, and MB doing 60% of SF's damage per attack.

Here's how I'm getting the numbers from the Simcraft results page:
All of the sims with Shadowfiend have 40.0 attacks over 553 seconds. Shadowfiend is cast with an interval of 181.7 seconds. This means Shadowfiend is cast 553/181.7 = 3.04 times, and has 40.1 / 3.04 = 13.18 attacks per cooldowns.
The three sims with Mindbender have 116 attacks over 553 seconds. Mindbender is cast with an interval of 62.1 seconds. This means Mindbender is cast 553/62.1 = 8.91 times, and has 13.01 attacks per cooldown.

Average damage done per Mindbender attack is 37055; average damage done per Shadowfiend attack is 60960. Relative damage is 37055/60960 = 0.607

However, when I check out how Mindbender and Shadowfiend differ during live, at 85, using 25.52% haste, I find something quite different:

Code: Select all

                MB    SF
Attacks         12/13 9/10
Relative Damage
   per attack   ~0.66 1

Shadowfiend is attacking much less frequently on live at 85 vs. on simcraft at 90. In addition Mindbender is doing ~10% more damage per attack on live at 85 than it is on simcraft at 90.

Am I doing something incorrectly?

[aica]

I figured some of it out. Twintop's sim uses Berserking and Bloodlust to increase the number of Shadowfiend attacks: Bloodlust w/ Berserking on the first Shadowfiend, and then Berserking with every other Shadowfiend (as the cooldowns line up).

Also, our pets don't benefit from our spell haste, only our melee haste, and not our melee attack speed, so we'll need to remove the 5% haste buff (and not give it the 20% attack speed buff).

Testing has shown that a 0% haste SFiend/Mindbender has very close to a 1.4875 second attack speed (not 1.5! 9.10% melee haste is enough to increase the number of attacks from 11 to 12 in 15 seconds; if its base attack speed was 1.5, it would need exactly 10% haste to get an extra attack, which it doesn't).

I guess that since Shadowfiend has more damage concentrated in a shorter period of time, any spellpower proc that line up with Shadowfiend will benefit its attacks more than that of Mindbender (assuming the spellpower proc is <=12 seconds, of which Jade Spirit and Synapse Springs apply). Is it enough to boost Shadowfiend's damage per attack over that of Mindbender's by 10% over a static case? I guess it's possible.

[Leida]

first off id like to say i am in no way a theorycrafter lol, i can barly read some of it sometimes and i actually like math.

now about sf and mb, yes on a pure one time cast of each sf would be on top. but sf has a 3 min cd while mb has a 1 min cd, so by the time you cast a second sf someone else is casting there 3rd mb, plus mb lasts for 15 sec which is 3 sec longer then sf. so for every 1 cast you get 12 sec of sf and 2 cast of mb you get 30 sec. i hope this help and make sense cause it does in my mind but maybe not in someone else lol.

[aica]

So, as I said in this thread, I've worked out the haste breakpoints for Shadowfiend and Mindbender. The values might be slightly off, because I've worked out the shadowpet's base attack speed purely by testing, and latency can throw off the numbers a bit. If someone has a more accurate value for the shadowpet's base attack speed, I welcome it.

Edit: Note that the 5% column (which signifies 5% spell haste raid buff for DP/VT/SW:P) doesn't actually include 5% spell haste, since pets don't benefit from it.

[Venaliter]

Pretty convenient, actually.

8055 is VERY close to 8085.

The Shadowfiend breakpoint is 10161; very close to the 10124 breakpoint on VT. We'll get that on the third raid tier in MoP; maybe even the 2nd.

Solid work.

[Blackmorgrim]

Work on the mechanics of Shadowfiend was something I felt was really missing from our Cata theorycrafting. I'm glad to see we have people stepping up to look into this time.

Also helps that Shadow hasn't had huge mechanical changes every two weeks like the first few months of cata.

[Shaz]

Hi Aica.

Thank you very much for your efforts on the SF.

I was looking for such information a long time (since mid of WOTLK) and never found it and my efforts of testing it weren't as successful.

One more question I have. Do you have any ideas of base damage and spp coeff for shadow fiend. Because of procs avg damage is hard to evaluate for getting this numbers.

Regards

shaz

PS:
When I look on simcraft http://www.simulationcraft.org/505/Raid ... hadow_T14H , SF damage is 0 in the damage breakdown. Where do I see SFs damage?

[aica]

I've been trying to figure that out myself.

Look at the "pet - shadowfiend" set of data, specifically melee. The shadowfiend in the priest list is just the actual cast, and shows you his uptime and cooldown.