For a better idea of how to compare your runs, please post your parameters.

Well, I didn't want to post too many details, as those options are still not finished.

BUT as long as people understand that those are still under test, it can only help if more people test them ;p

Just to be clear, "under test" means:
- can have some bugs
- can be changed before going "public"
- change can be also in how you use them (so different option names, way you use in simcraft etc)
- some options mentioned here exist only in latest build, so you will not get them by downloading packed EXE, you need to download latest source and compile
- but most of those are already in latest published version, only not announced. For example, while current non-published build is 2509, latest published build also include movement and armory parse
- ...


With that said, these are few new options where testing would help most:

1) "smooth random"
- this should reduce variations in numbers between two runs, especially if you have small number of iterations (1000 is default) and you want to calculate stat factors
- to turn it on , use "smooth_rng=1"
- it is by default turned on in latest "opt_raid.simcraft" which is used in "Warlock_T8.simcraft"
- if you want to test and compare it to regular random, set (after opt_raid) option "smooth_rng=0", and then compare numbers. Increase of iterations , like "iterations=10000" , is also suggested if "smooth_rng=0".

2) movement
- For initial test, just replace in "Warlock_T8.simcraft" line "opt_raid.simcraft" with "opt_raid_mov.simcraft"
- for more detailed tests, try to make your own "cocktail" of movements/stuns/invuls in "opt_raid_mov.simcraft"
- each of those movement types ("invul","moving","stun"..) can have options like (period=,duration=,first=,last=,distance=...)
- "moving" means you can use insta-cast actions. "stun" means you can not dps during that time (but dots are still ticking). "invul" means boss is going untargetable/invulnerable, and even existing DoTs will be removed
- it would help if people would try to "mix" those commands to represent several Ulduar boss fights, so we can try and find some "cocktail" mix of those to represent "average" movement raid. Right now "opt_raid_mov.simcraft" is just rough guesstimate

3) armory parse
- this is also new, and allows you to check and simulate DPS and stat values for your own char instead of BiS chars that are given
- way to test that is to make your own "test.simcraft" file with something like:
- of course, instead of "eu" you can use "us", "cn", or even include whole URL
- you can also put several players from same realm in same line
- and you can select for each player to use alternate (secondary / inactive) talents and glyphs, by putting "!" before name
- and as already mentioned before, you can test movement by changing to "opt_raid_mov.simcraft"
- you can turn on debug output to check if your items are parsed correctly. Just put "armory_debug=2" in config file (2 shows you stats of each item, 1 only total parsed stats)
-example of one config file that would test your alternate talents with 2 more warlocks from your guild in "moving" raid, and want to show debug info also:



Once more, while these options are already good enough to be used, they are still under test, so don't get surprised if something does not work as intended. But as I already said - more people willing to test them can only benefit simulator ;p

This scale factor was in first run. I was wondered, but SCALE_FACTORS.BAT shows lag=0 too (after very long time of waiting )
Yes, smooth_rng = 1 by default in opt_raid.simcraft. I have not changed anything.

I just want to point something out.

Factoring in movement/boss imune/stun time makes any of these sims even less valid than the base sims are.

Here's a quick look at whats up.

Original Simcraft issues

• Stand and shoot only
• it's the PERFECT Raid
• It's not a benchmark, it's the bar set so lofty it's nearly impossible to top
• There isn't a single boss in Ulduar that works exactly as simcraft is, ie Stand and Shoot.
• The benefits of movement can't be calculated, though moving is always a dps loss, the benifits of say LT or Corruption go down do to it.

Current SimCraft issues

• Your fight, isn't my fight. IE What you set up your fight as may be FAR worse than how I set mine at, You may move every 30 sec, I may move every 60.
• No fight in ulduar has set movement 100% of the time, You can't look at a single fight and say "Oh I won't be doing anything here, here and here.
• The benchmark is set low. It's nice to know in the worst situation out there what the best spec is, but it's rare it will ever be that bad
• The benefits of movement STILL can't be calculated. Yes you move, But in so many of the fights are you just moving? Sure you can LT sometimes, but maybe sometimes you have to waist a GCD on something else, Or you can stop and cast, and move again (IE safteydance)

You see, adding in the ability to model a real fight DOES have it's place, but it should be reserved for discussing particular fights and the merit of x spec over y.

A much much better thing to model vs movement and stuns and immunes is pushback. Most fights have it, and most fights have some current/measurable amount. Like mimiron p2, You can legitimately model the output of damage from that for pushback.

Now, Don't get me wrong, great stuff, really nice tool to add into the simcraft arsenal, But just take what you see with a grain of salt.

Deleted. I think this question is better asked in Simple Questions section.

Well, if you do not find simulation of movement useful, you can just not use it ;p

But I believe it can be useful. Some comments:

while your (movement) fight may not be same as other fight, in most cases they will all result in changes of stats values in same direction. They may differ in DPS, they may differ in exactly how much stat values changed, but in general (even for two very different movement fights) stat values will tend to change in same direction.

That means you could get better information about "real" relative value of stats, since almost all fights in ulduar have some movement. So if almost any movement simulation shows reduced value of SPI, I would prefer to value it that way when I decide among items.

True, just as you can look at single fight and say "I will cast nuke here, curse here and here, and ..oh..I will crit here and here"
But you can still fairly well model boss fight. For example, if Ignis is casting jest every 30sec, and you know you need to stop casting for 3sec, you can model it. Or you can model that you will once end up in pot for 10sec.

Right now opt_raid_mov is set to have 10% movement. It may be to much or too little compared to real average uldual raid, but it is not "worst situation"

It is already set that simulaiton use other things too while moving, not only LT. And if you really want to "safetydance", then just set more shorter movement periods.



But to reiterate, use movement simulation only if you think it is useful to you. If you can not find way to make it useful, just use default "ideal" simulation.

Zal: I'm not sure if this is normal, but there is no 'proc uptime' for Illustration of the Dragon Soul when I run SimCraft with illustration_of_the_dragon_soul=1. Is it being modeled properly? Or are we assuming that people pre-tap and that it is basically a flat +200 SP.

Very sure that the trinket is modeled starting from 0 SP, otherwise there would be no need for the illustration_of_the_dragon_soul=1 clause in the gear file. The instructions would simply ask you to add 200 to the spell power attribute.

I have a question regarding the optimal raid parameter. How do I see this list in order to change it to something sub optimal, eg remove orc racial, etc.

Im having some questions about Backdraft, and the relation between haste/DPCT. Ive been wondering if the Backdraft proc, could change the priority regarding what spell to cast under it, in between rotations sometime I come with some quick decisions to make, and for some time now, Ive been debating myself on those issues.

Do we follow the normal rules regarding DPCT and execute our rotation always like that? Is it worth to delay a cast to "fit-in" some cast (higher in DPCT) into backdraft? If we (eg, due to movement), still have a backdraft proc, and Conflag is off CD, is it worth to delay conflag or use those backdraft charges, or ignore it and conflag?

Some enlightment from someone more comfortable with this issue would be great

Tks in advance

Regarding Illustration of the Dragon Soul: You are correct.

Regarding the optimal_raid parameter: Most options are simply "last setting wins". The optmal_raid paramter is a little special in that it toggles a ton of options, so....

Config 1:
optimal_raid=1
improved_shadow_bolt=0

Config 2:
improved_shadow_bolt=0
optimal_riad=1

These configs are different! In the second one, the optimal_raid=1 will stomp all over the previous improved_shadow_bolt=0 setting.

If you don't know the name of any option, the format is (almost!) always:

(1) all lower case
(2) spaces become underscores
(3) remove apostrophes and the like

IotDS should really be calculated as a constant 200 spellpower due to the fact that you can stack it way before the fight starts. I'm rather surprised to hear this is not happening.

Two questions:

1) The higher the Lag value indicates more sensitivity or less sensitivity to lag?
2) For the purpose of gem optimization, how would one use this data? (Edit: would scale factors with movement be more relevant when optimizing gem/gear?)

It's up to you. If you want to stack before the fight, set illustration_of_the_dragon_soul=0 and add +200 sp to the gear.

Why don't all of the destruction sims include using the Doomguard when it's off cool down? It would appear that using it always yields higher results.

I would still keep points in empowered imp for the times when the DG is not available.

They can occur while you are casting, and if stun/invul they will cancel cast.
You do not specify when exactly they occur, but how often ( you set period/cooldown ), and how long they last (duration).
You can also define "randomness" by using standard deviatiation: "cooldown_stddev=" and "duration_stddev="
But even without using "stddev" options, between two runs they will happen at different spells, because there is also randomness for lag, spells miss, procs etc between runs.

I imagine it's cause Doomguard doesn't have Avoidance. You'd have to factor in how much time you'd spend microing your DG to avoid it being gibbed to accurately gauge it's feasibility in most Ulduar fights. Using a DG as Affliction provides such a huge boost to DPS because your normal pet's DPS is low in comparison, which is why it's included.

Because the 4-piece bonus is worth like 150 DPS, much more than you'd gain from any two individual pieces.

If you're affliction, the 4-piece bonus is only worth 87.5 dps. I'm not sure if you could gain more than that by replacing 2 pieces with ilvl 239 gear, but the 4-piece would probably still be better on Hodir and Vezax where you get a huge haste boost and shadow bolts make up a larger % of your total damage.

Dont forget to look at the 4pc Bonus as crit rating and not as simply 5% crit. It is equivalent to 230 crit rating. Sure crit may not be worth as much as haste, but that much crit from 2 items is significant.

Remember that the 4-piece bonus only applies to SB/Incin and as such isn't comparable to 230 crit rating. The translation into dps seems to me to be more constuctive.

Incinerate is typically 40-45% of my damage done. So I'd say 92 crit rating would be a fair comparison.

That would only be reasonably accurate if all the spells in your rotation could crit and if all your crittable spells had the same crit rate. The simcraft results on the dps-delta are the better choice for comparison than trying to determine a static combat rating equivalent.