Simulationcraft for Sage/Sorcerer

[CaseyTheRetard]

Boss armor is one source of discrepancy between the spreadsheet and simulationcraft models. The spreadsheet is using 50%=10800 base armor for the target, simulationcraft is using 11400, which would provide 50% damage reduction against a level 53 attacker.

Edit: I can't find a reliable source for Operation boss armor. The 50% number seems to be more the result of conjecture than testing (post #165 in the formula thread). It would be nice if the two tools were at least consistent, even if the actual number is wrong.

[Caltiom]

(01-28-2012 05:33 AM)CaseyTheRetard Wrote:  Boss armor is one source of discrepancy between the spreadsheet and simulationcraft models. The spreadsheet is using 50%=10800 base armor for the target, simulationcraft is using 11400, which would provide 50% damage reduction against a level 53 attacker.

Edit: I can't find a reliable source for Operation boss armor. The 50% number seems to be more the result of conjecture than testing (post #165 in the formula thread). It would be nice if the two tools were at least consistent, even if the actual number is wrong.

53*200+800 = 11400
and
50*200+800 = 10800

As long as the spreadsheet uses a lvl 50 target both numbers are equal. I'm not even sure what level bosses do have, feel free to enlighten me

Epic Power Potion is included, I might add an option to specify the buff amount, maybe later add separate names. Relics are not yet implemented, but shouldn't be a problem to do so.

About LB getting consumed by a unbuffed FL: Good find, I know exactly why this is happening in the simulation. But are you sure it doesn't behave that way ingame, even if it wouldn't be the intuitive way?

lostdummy: If you really want to compare sim to spreadsheet I'd advise you to first run the sim with vary_combat_length=0 and fixed_time=0 to ensure a strict 300s combat. And then I'd compare without Force in Balance / Death Field for now, as you can't assume that you have the same percentage of affliction ticks benefiting from it than all dots ticks are. But I could add specific tracking for that, later.

Then are you accounting for half-finished afflictions in the spreadsheet? By that I mean afflictions cast 15-0s before the end of the fight, which add another execute, but not full damage.

About Wrath: The value in interval means the interval betwen starting the buff ( from 0 to 1 stacks ), which in this case will corelate with the interval crushing darkness. Trigger-interval gives you the average intervall between triggers ( wheter from 0 to 1 stacks, or just refreshing it ), in this case 3.5s. So that would result in around 85 triggers, which is around 30% of 277 FL ticks.

[DPSFTW]

(01-28-2012 07:56 AM)Caltiom Wrote:  About LB getting consumed by a unbuffed FL: Good find, I know exactly why this is happening in the simulation. But are you sure it doesn't behave that way ingame, even if it wouldn't be the intuitive way?

Yes I'm sure, LB only starts on next FL cast if it procs while you already casting a FL.

What About Recklessness Skill? it doesn't show up on combat log at all, are you using a different name?

[lostdummy]

(01-28-2012 07:56 AM)Caltiom Wrote:  Then are you accounting for half-finished afflictions in the spreadsheet? By that I mean afflictions cast 15-0s before the end of the fight, which add another execute, but not full damage.

That could explain main difference in DPE, if you calculate DPE to represent partial damage also. I intentionally calculate DPE to represent damage from full spell, since I presume that is more meaningful/useful number. For example, if DPCT would be based on DPE per CT, priority order decisions for rotation should be based on DPCT from 'full' spell DPE, not partial. Partial damage is counted in total spell damage, if last spell is partially used.

[Caltiom]

Version 101-5 is now available, and sample reports on http://www.simulationcraft.org have been updated.

The release notes contain a complete list of changes and can be found here: http://code.google.com/p/simulationcraft...leaseNotes. If you encounter any issues or have a suggestion/idea please create an issue at: http://code.google.com/p/simulationcraft...ssues/list

As always, if you're interested in getting your hands dirty writing code, helping with documentation, or providing class knowledge, please hit us up on IRC. irc.stratics.com (#simulationcraft) or leave us some Feedback.

For the future the main coding plans are to add set bonuses and relics, and I'd really invite everyone out there to player around with movement-scenarios ( Movement fight styles are available in the GUI ), and improve them ( action list, talents ). That's one strength of a simulation, and I plan to have separate movement profiles / reports if people contribute them.

Edit: Fixed the ReleaseNotes URL, it was trying to add the period to it ~Kor

[Kaedis]

Your spec URL is not longer value, looks like TORhead converted over to a hash system instead of the numerical notation.

I'm also curious as to what the "snapshot_stats" line in the priority list does.

Also looks like you're not including a "force_lightning,if=buff.lightning_barrage.react" in the Hybrid build (which has Lightning Barrage).

Lastly, are you simulating double-dips of Force Barrage for the Hybrid build? It's surely a bug, but it's also a fairly significant portion of their dps. My sim simulated it by giving the buff 2 charges, but setting the remaining duration to no more than 2.0 seconds when the first was used (but it prioritizes using a second FL immediately).

[Caltiom]

Thanks for fixing the broken link.

Regarding torhead links I'm going to quote myself:

(01-21-2012 11:49 PM)Caltiom Wrote:  I know torhead links with the talents in integer form don't work. They did on wowhead, not sure why torhead doesn't have that feature. SimC currently can only decode torhead links, not encode them, so I won't be able to offer you correct links right now.
But I do use encoded torhead links for the profiles, they can be found in create_Jedi_Sage.simc and create_Sith_Sorcerer.simc ( which are the "source" profiles, from which the others get created, which then have decrypted talent links ).
I already wrote to the torhead team if they depreciated the feature completly, or if it just hidden behind some strange syntax different from the wowhead one. No reply so far.

I'm guessing in WoW the wowhead team had a great incentive to offer parsing of integer talent strings not because of SimulationCraft, but because the armory used the integer form too for a long time.
I could just pass the un-decrypted talent string put in by the user to the output, but that wouldn't completly garantuee its correctness.
Edit: I sent torhead another e-mail, maybe they will reply this time.

snapshot_stats is a virtual action, usually called after pre-combat buffs/stims to snapshot the characters buffed stats with which you will start into combat, at the end they get reported in the html/text output.

Casey created the hybrid action lists, so I don't know exactly what is best there and what not. But could you specify at which point a FL,if=buff.LB.react line would have to be inserted? Otherwise the FL action without any expressions of course also takes care of FL's with LB up.

I added the double dip part, pushed Hybrid dps from 1200 to 1237.

[Tyrus]

Thanks for this tool, I find it quite useful.

Is there any plan to add support for other classes as well ? I'm looking for a good assassin spreadsheet/simulationcraft but can't find any. Is there anyone working on it already? what would be needed to start developing one ?

[Kaedis]

Based on DPCT, Lightning Barrage FL should be top. However, if you include theoretical dps gain from Affliction via further LB, it would go just below Affliction and just above Deathfield.

[Caltiom]

(02-07-2012 08:00 PM)Tyrus Wrote:  Is there anyone working on it already? what would be needed to start developing one ?

No, and the second question is kind of answered by the first one

I Added the extra FL line to the hybrid profiles, another small dps boost from 1237 to 1250.