Theck's Simcraft Thread - WoD/6.x

Schroom wrote:hasn't it yet happened? I thought it would have been implemented along with the other changes in that hotfix.

The tooltip still says 10% so I wasn't sure

theckhd wrote:

golfinguy wrote:Here it is:

http://pastebin.com/18u3EF1b


(shields up - ready for my embarrassment)

Sorry, still haven't gotten to this. It's on my to-do list, but I may not get time until this weekend.

No worries, it was one week tanking duties and our regular is back. Still want to be better prepared next time, but if the weather continues to improve I'll be golfing this weekend! ;)

Klaudandus wrote:

Schroom wrote:hasn't it yet happened? I thought it would have been implemented along with the other changes in that hotfix.

The tooltip still says 10% so I wasn't sure

Most tooltips cannot be trusted when things are hotfixed, a client patch is usually required to fix the tooltips. Case in point, look at the tooltip for Shield of the Righteous, then look at the value stated in the buff while it's active (the value in the buff is the accurate one).

Hi, I was messing around with SimulationCraft and noticed some things in my results that I couldn't really explain.

First, it looks like Shining Protector only gets one multistrike% roll instead of the two it seems to get ingame (despite what the tooltip says), at least judging by the percentage of my heals that procc Shining Protector.

Secondly, it seems like Shining Protector proccs off of Light's Hammer's Arcing Light in SimulationCraft, which it doesn't do in game (even though Arcing Light can multistrike. Shining Protector is weirdly selective, for some reason)

I'll look into both of those (though if you have logs illustrating the behavior in-game, that would help speed it up).

golfinguy wrote:Here it is:

http://pastebin.com/18u3EF1b


(shields up - ready for my embarrassment)

I can't replicate this.
All 6 actors
HA_Sera
HS_DP
HS_SW

The "All 6" sim is what I get if I put your pastebin in the Simulate window. For the three individual tests, I just selected that actor and copy/pasted that section into its own new Simulate window. The results seem consistent with one another as far as I can tell.

This was with SimC 603-22 (just what I had handy on this computer). Can you give me a little more info? In particular what version you're using, though if you have the individual reports that would be helpful too.

WolfMeat wrote:Hi, I was messing around with SimulationCraft and noticed some things in my results that I couldn't really explain.

First, it looks like Shining Protector only gets one multistrike% roll instead of the two it seems to get ingame (despite what the tooltip says), at least judging by the percentage of my heals that procc Shining Protector.

Secondly, it seems like Shining Protector proccs off of Light's Hammer's Arcing Light in SimulationCraft, which it doesn't do in game (even though Arcing Light can multistrike. Shining Protector is weirdly selective, for some reason)

The Shining Protector implementation is based off of how it worked in beta, so yes, it's only making one MS roll per heal. Apparently that changed somewhere along the way, because it's definitely able to proc twice now:


Will do some inquiring to figure out whether this is intended, and if so, modify the model.

Testing the LH thing now, but I think you're correct (this is harder to tell from logs). However, I'm pretty sure that's a bug. Again, will inquire and post back here once I have a resolution.

LH confirmed bugged, should be fixed on PTR in a few builds.

Confirmed that SP is working as intended, so I need to update the implementation in SimC. That should help Multistrike out a bit.

theckhd wrote:LH confirmed bugged, should be fixed on PTR in a few builds.

Confirmed that SP is working as intended, so I need to update the implementation in SimC. That should help Multistrike out a bit.

Cool! Sorry I was away and unable to provide my logs. (I should probably have thought of that in the first place)

On a side note, regarding the usefulness of Multistrike, I noticed Shining Protector doesn't work with quite a few spells for no discernible reason.
It works with the entirety of Riptide, but not with the HoT portion of Eternal Flame.
It doesn't work with Healing Stream and Healing Tide Totems. I'm not sure if those can even multistrike, but then again Shining Protector proccs off of Lay on Hands, which also doesn't seem to multistrike.
It doesn't work with Divine Star, Penance, or Power Word: Shield - I don't know if it's supposed to work with absorbs though. Sacred Shield also seems a little weird in that regard, as getting the 100% multistrike buff doesn't seem to give me more multistrikes on the shields at all.

... Sorry if this got a little off-topic. I hope some of it is useful to you though. I can probably generate some logs of these if you need them. :)

Sacred Shield's multistrikes won't show up as separate events in the combat log. The way it's coded, it makes crit and multistrike rolls and just wraps that into the base absorb of the shield you get. Can test this by just casting it on yourself while out of combat and checking the log, you'll see different absorb values from tick to tick.

The mechanics are a little wonky too, it only makes one crit roll and two MS rolls (i.e. the MS rolls don't crit independently), so you'll always get multipliers of 1.3 (one MS), 1.6 (two MS), 2.0 (crit), 2.6 (crit+MS), or 3.2 (crit+MS+MS).

theckhd wrote:

golfinguy wrote:Here it is:

http://pastebin.com/18u3EF1b


(shields up - ready for my embarrassment)

I can't replicate this.
All 6 actors
HA_Sera
HS_DP
HS_SW

The "All 6" sim is what I get if I put your pastebin in the Simulate window. For the three individual tests, I just selected that actor and copy/pasted that section into its own new Simulate window. The results seem consistent with one another as far as I can tell.

This was with SimC 603-22 (just what I had handy on this computer). Can you give me a little more info? In particular what version you're using, though if you have the individual reports that would be helpful too.

I'm dumbstruck. I was using the latest (603-24) on my desktop, but decided maybe something was wrong so I downloaded it on my laptop and ran new simulations there - to get a fresh start. The only two things I did was turn on Scaling for secondary stats, and selected TMI as the 'scale over' factor (with survivability being my first priority, I'll worry about meters later). Please let me know if that is not the best way to sim for prot.

Using the Protection T17H template (may be a bit before I see Mythic gear in significant volume) I ran it again, and get the results below. Versatility showing up as second best - which surprises me.

PROT T17H: DP/Seraphim
Simulate Text


But it gets even stranger when I copy the gear from that template and sim all of the talent combos (save EmpS). The results are here:

PROT T17H: Automation Sim
Simulate Text

Look at what happens with Haste and Versatility from Holy Shield to Seraphim! For all Seraphim sims, Versatility is the top stat. And for Seraphim/HA - haste falls to the floor. I simply do not trust what I've done at this time.




I've included the text from the Simulate tab for each, in hopes that it points out something foolish I've done. Save that, I'm guessing its how I set the options in SimCraft - but the two changes above are the only things I've touched (on the laptop anyway, which is where these were generated).

golfinguy wrote:Save that, I'm guessing its how I set the options in SimCraft - but the two changes above are the only things I've touched.

I didn't get a chance to thoroughly go through your newest results and setup yet, but one thing that stands out to me is you're still using the "Fluffy Pillow" boss, while Theck is using the "TMI_Standard_T17M" mythic boss. Last I looked, you do need to set the boss type on a new install. The other bosses have somewhat different damage profiles than the TMI bosses, so perhaps that's contributing to the differences that you're seeing?

That said, if what's mostly bothering you is the scale factors for haste, those are going to be fairly unreliable, especially for Seraphim. I've certainly gotten scale factors that look like yours when running Seraphim sims, depending on how close to one of the haste breakpoints your gear happens to be. In addition, versatility often does very well, at least on Seraphim sims from what I've seen. Even looking at the graphs and scales on Theck's blog post, eg:


Versatility doing well with Seraphim makes some sense. It's one of (if not the) most reliable stat(s) besides Armor/BonusArmor/Stamina to reduce your chance of a spike during the 5 second window that you're pooling HP before you can cast Seraphim. In fact, now that I look at those scale factors above, all of the EH stats seem to scale "better" for the Seraphim actors than for the equivalent HS actors, probably for the same reason? Interesting.

Anyhow, TLDR: Your most recent results don't look that odd to me unless I'm missing something, but the different boss used might be why you're getting slightly different output than Theck.

If you're still having other issues with the automation tool, I can't really help much there, as my experience with it is much more limited. Sorry.

There are some weird differences there. Looking at the report, the single-actor Sera_DP, the DPS is higher because it's casting almost twice as many SotRs. This also means that actor is getting more SoI procs. The combination probably explains the significant increase in survivability.

Looking at this between pulls during raid, so I'm still hunting for the cause of that anomaly.

Found the problem: your automation actors weren't getting holy power from Judgment. They were all taking Long Arm of the Law, while your single-actor sims were using Pursuit of Justice. There's a bug in SimC that's causing Judgment not to generate holy power if you take LAotL, so that's the problem. It was fixed earlier today, so until release 603-25 avoid using LAotL in sims.

WHEW!

OK, seems I'm all better now. The seemingly innocuous talent difference (bug) accounted for the single vs automation issue I was seeing. And the fluffy_pillow accounted for the crazy Versatility values I was seeing.

Thanks all!