Theck's MATLAB thread - Cataclysm/4.x

[Iminmmnni]

the values from the Ability Damage post:

How does SD and Inq factor into the ability damage table? The ability damage post indicates that the rotation is N/A but the rotation has a significant effect on parsed damage due to SD and Inq uptime.

Is there a calc* script in the codebase that will display (Inq) and (SD) damage?

[tlitp]

[CAD] is a low-level script, it doesn't compute the aggregate output; EG/GrCr/SD/Inq are all ignored. Think of it as being a tool to check scaling and damage formulas.

[Caeles]

If Inquisition is up, hit HotR every 2nd GCD
If Inq is down, hit CS every 2nd GCD.
If at 3 HoPo and Inquisition is up, hit ShoR.
If at 3 HoPo and SD is up, hit ShoR
If at 3 HoPo and you haven't got a SD Proc to use and Inq is down, hit Inquisition
Hit AS between CS/HotR if it lit up. Don't push back Inq, ShoR, HotR or CS to hit it.
HoW if you're in execute phase.
Judge for Mana & SD procs.
Whatever's off CD. Skip Consecrate unless you have mana coming out your ears.

With regards to the above priority, is this saying that HotR is stronger than CS for single target if Inquisition is up? I didn't see this reflected in sim 56, which is why i'm curious where this priority came from.

[knaughty]

With regards to the above priority, is this saying that HotR is stronger than CS for single target if Inquisition is up? I didn't see this reflected in sim 56, which is why i'm curious where this priority came from.

Probably mis-reading HoW for HotW? Mistake on my part by the sound of it.

[Awyndel]

Do the sims with concecrate take mana into account? It could affect J and conc priority.

[theckhd]

Do the sims with concecrate take mana into account? It could affect J and conc priority.

No, they don't. That's why I've included both 'X' and 'X>Cons>HW' versions of the important queues. We've seen that Cons/HW can be treated as an optional DPS increase via filling gaps, mana permitting. Given the long cooldowns (and thus low DPS contributions) from both, it's highly unlikely that it's worth prioritizing Judgement higher just to get more Cons/HW off.

[Awyndel]

Well that settles the J question.

I'm just wondering about X-conc-hw vs X-hw-conc. One would run oom faster then the other. Perhaps putting conc last would allow us to keep casting. I'm going to test that next raid.

[Cassyboy]

Well that settles the J question.

I'm just wondering about X-conc-hw vs X-hw-conc. One would run oom faster then the other. Perhaps putting conc last would allow us to keep casting. I'm going to test that next raid.

Very simple, do what your mana permits. You may fail at the first while but after a few days you know exactly when consecration will make you oom.

[knaughty]

Easily answered by looking at damage-per-mana. Cons will run you OOM but if it's higher DPM you still use it as filler.

[theckhd]

I have another round of updates ready that incorporates GrCr HP on cast and AS not proccing seals. In addition, I'm updating the default gear set to a T11 heroic set. I've already updated the gear set post, feel free to peruse it and point out errors/suggestions/etc. However, I'm on crappy hotel wifi right now, and have a busy week planned, so it may take some time to get around to updating the rest of the posts.

[tapdov]

Is there any reason a similar code setup wouldn't work for ret?

[Iminmmnni]

Is there any reason a similar code setup wouldn't work for ret?

The approach is feasible for any rotation where the state space is of a manageable size. Prot's maximal rotation has ~150k states which takes about half a second to generate the graph and 3-4s to converge the state probabilities. The approach is not feasible for specs where arbitrary haste must be modeled accurately. Prot is ideal as we essentially ignore haste so we use a 0.5s step size (we could use 1.5 but 0.5 allows for possibility of 8s judge-on-CD rotations). Total prot state space size using 0.5s step size is ~10^16 which increases to ~10^25 with 0.1s step size.

The state generation and convergence code was designed with reuse for other specs/classes in mind so yes, in theory it can be applied to ret. I can't say whether ret is practical as I haven't been ret since levelling to 80 when wotlk was released so I'm really out of touch with ret mechanics. I can see it working if we only took certain haste data point and modeled with 0.1s step size. I'll brush up on ret mechanics, code up some basic ret rotations and see how many states they use.

[RedAces]

hey,

Is there any reason a similar code setup wouldn't work for ret?

I think it would, but you'd have to incoporate all ret mechanics into it. Use SimulationCraft until then.

Bye, RedAces.

[Cavalorn]

Due to the WoG-CD I had a discussion on some ways to improve the remaining selfheal.
I'll be interrested in some sims which compares the SHPS of different talentspeccs with SoI activ and including JotP(Meelehaste -> some more SoI-Proccs) in fights where TPS/DPS doesn't matter.

In short: JotP(1/2/3) vs. RoL(1/2/3) vs. GbtL(1/2) vs. Reck(1/2)
With a #28(WoG>CS>AS>J), #29(WoG>SotR>CS>AS>J) and/or #30(WoG>CS>AS>J>Cons>HW) single-target-rotation don't take overheal into account.

Napkin-Math with the numbers from the "Ability Damage"-Post give me for 2/3 JotP in a 7min patchwork-fight. ~15 extra-hits from JotP * 3,1K(SoI) / 420sec = ~110 SHPS
If this is too complicated or too much work, feel free to skip my request

[tlitp]

The approach is not feasible for specs where arbitrary haste must be modeled accurately. Prot is ideal as we essentially ignore haste (...)

If we are to be rigorous, haste isn't a smooth manifold for Prot. Autoattacks (via parryhaste), SoT (via Censure), ability usage (via mana income, via JotW) - they all exhibit "ugly" scaling with haste. It's fair to say, however, that their effects are somewhat lost in the noise generated by other abilities/mechanics.