#tc-research

1 messages · Page 1 of 1 (latest)

coarse anchor
#

Hello, I'm starting to look into a project for optimizing APLs, and I'm wondering if simulationcraft has any plugin interface? I didn't see anything in the github docs so I'm starting to read through the code, but figured I'd check here first.

#

Also, is there a dedicated simulationcraft discord that would be better to ask these kinds of questions in?

regal agate
#

There is a (not very active) project to generate the apl with machine learning.
You can find it if you look at the aplgen channel on the Simc discord.
link to the discord: https://discord.gg/tFR2uvK

coarse anchor
#

Excellent, thanks!

plucky mural
#

but it needs to start with a reasonable APL as the training material correct? or does it start from scratch with a list of available actions

regal agate
#

I haven't looked a lot into the project, but i invite you to join the discord above and read/ask in the channel designated to it

knotty oriole
#

Also might be interesting to train it with like all our official APLs over the last 3 xpacs or something.

fluid drum
#

That'd be pretty sick if they could get this to work 😮

#

That's way back though from like Legion... Just looked at some of the code they were trying. 2016?

coarse anchor
#

Yeah, I'm thinking of how to apply reinforcement learning to find a policy (i.e. APL) that will maximize DPS. The project fuu pointed out is already trying to do this, so once I gather my thoughts and do a bit of research I may try to collaborate with them

coarse anchor
#

This is similar to what Deepmind did with AlphaGo: start with training on a database of human games, then switch to self play

#

It does look like most of these efforts have gotten stuck on integration with simcraft. Seems like getting an agent hooked into the simulation in place of an APL will be one of the main challenges with this kind of project

knotty oriole
#

The major challenge here is making APLs where simplicity is also enforced as a positive and not just massive complexity bloat for minimal improvement. Still needs to be “sensible” even if it comes up with different approaches.

#

Also need to run against various sim types or lengths to avoid too many timing specific local optimizations

coarse anchor
#

Yes! these are excellent ideas

regal agate
#

To sim for the seasonal affix, just paste this below your /simc in advanced sim.
Note: This is just a estimate, so not a perfect representation as it gets slightly impacted by stat dr.
(Sims are for 5 and 10 stacks)

📎 season_4_affix.txt
regal agate
#

This should include all S4 items on max item level including WoD items.
Can be just copy/pasted in top gear under your gear profile.
Note: Not 100% sure about the WoD Dungeon loot table, but will update if something changes.

📎 gear_s4.txt
rocky fern
#

Just as a heads up, although not technically a rogue-specific issue- The Mechagon bracers [Wraps of Electrostatic Potential] don't appear to be getting used in simulations I've run so far. I doubt they're actually any good, as I don't recall them being worth losing secondary stats for the damaging effect, but figured I'd let you folks know in case it ends up being better than other options.

willow knoll
#

They were bad back in BFA because it shares a cooldown with trinkets? I'm pretty sure

#

Trinkets, or combat potions I forget which one it blocks

rocky fern
#

Hm, I'd have to test it real quick

#

I don't remember that bit admittedly, I'll try them out to see what they overlap with

#

That might also be why it wasn't showing up in sims if that's the case since I had DPS potions enabled
Edit: So I tried them out- No overlaps with either of the on-use trinkets I had (xy'mox/IQD), no overlap or other cooldown issues when using a DPS potion, and it appears the bracers DoT cannot critically hit. I did look it up on Wowhead, for what it's worth, and the active effect Electrostatic Induction does not appear to share cooldowns with anything.

raw knoll
#

there was no cooldown sharing that I can remember originally. it was basically just macro it to something else.

knotty oriole
#

I will check the SimC implementation, although I don't really see why it wouldn't work

#

I'm not sure it's actually good though since it has no secondary rating on it

#

Might be though

raw knoll
#

it had some pvp use for extra burst, but was never really sought after in pve.

knotty oriole
#

So looks like it just needs APL support for the use_item action for it since it's a non-standard slot

#

Seems pretty side-gradey with the same ilevel bracers

#

So not bad/not particularly good either

#

But they may be upgrades in some cases so will check in some support for it tonight

rocky fern
#

Yeah I figured it'd be an edge case since it does miss out on secondary stats, and also has positional requirements on top of that too. Just got one from a Junkyard spam I've been doing and figured I'd get the word out in case people don't realize how good/bad/whatever it ends up actually being is all, especially since it just flat-out wouldn't show up in any sims I tried. ^^;

mystic jolt
#

The same happens with the kara ring afaik

noble pendant
#

Kara ring works

regal agate
#

the damage part of the ring is just very low

knotty oriole
#

The issue with it back in Legion was that the proc was not set up to scale from ilevel correctly

#

It's possible they have fixed this

#

Also very possible they have not 🙂

#

Or I think the issue was that it had some max scaling level of 45 or something silly

#

yeah

Name             : Collapse (id=234142) 
School           : Shadow
Spell Type       : Magic
Max Scaling Level: 45
#

👌

noble pendant
#

But the mechagon ring is a nature proc

#

Oh Kara ring

#

dentge nvm

spiral light
noble pendant
#

The tooltip alone is already KEKLaugh

spiral light
#

and this is just insane as well

#

not only is the damage portion already absolutely terrible, it has a chance to go on a 5 min cooldown

#

but yeh the damage exactly like it says on the tooltip. so terrible.

knotty oriole
#

So it’s a risk/reward thing

#

I forget the optimal strategy from Legion. Would just sim it if it worked.

#

But it doesn’t 😄

spiral light
#

too bad the reward is so terrible that it's not worth even thinking about

knotty oriole
#

If the proc was scaling with ilevel correctly looks like it'd be around 160 DPS at 304 ilevel

#

I'll go ahead and put in the APL logic tonight just in case the spell data gets fixed 🙂

spiral light
regal agate
noble pendant
#

The ring scaling got hotfixed apparently

#

Might be worth looking into again

regal agate
#

The Season 4 affix is now supported by simc/raidbots you can use the below to sim it:
(advanced sim on raidbots)

📎 s4_affix.txt
lavish zenith
#

u guys ever thought abt some weird conditions for necro in the apl during spike dispatch only periods? i noticed sometimes on packs i would be dealing a fk ton of damage with shark buff compared to traditional "good" buffs, esp in this specific window. my thoughts are the buffs that normally give value in certain ways don't actually have any value here

broadsides, worthless cuz extra cp doesn't matter cuz u just go 0->6 dispatch 0->6 dispatch

bt doesn't matter cuz spike and dispatch already so energy efficient u never oom and usually overcap energy

snc almost no value cuz ur not ssing. only time it good is when there's like a 1-2s break between the spike dispatch spam where u could try to fish for a gsw opp in those globals which it does seem decent for

gm kekw

tb is actually decent, still mid but more notable than the others cuz it gives cdr on the vanish to fish for the shark if u don't already have and the bte so 1-2 extra gsw shooters after like 5 or 6 spikes cdr

and rp is goat

i mostly notice it when mobs are dying and, pack has like ~15s left, and im almost perma spam spike dispatch. if i have shark i should technically reroll the shark, but like almost every roll combo is going to be less dps than shark, except maybe shark + tb. so in these situations i've actually been trying to sit on the shark. happens maybe once per key. maybe more frequently in a key like junkyard or lower.

maybe it's too niche of a scenario and doesn't occur often enough to make a dif in a sim or anything. does the logic at least check out?

knotty oriole
lavish zenith
gusty ridge
lavish zenith
#

oh yeah i am TROLLLLLLLLLLLLLING ok bs not worthless

#

still stand by the others tho

shrewd summit
#

Where can we report items not found in Raidbots? annoying to see that all backs are upgrades xD

unknown.png unknown.png
regal agate
shrewd summit
#

Thanks, someone else reported it apparently 🙂

delicate mountain
#

Is there a way to sim Seasonal Affix?

warm rain
#

scroll up a bit

finite sparrow
knotty oriole
#

Unfortunately

#

Although with only 30% uptime it's probably not too significant

#

Not sure if the data gets recorded and just needs to be added to the report or if it isn't recorded for DoTs at all. Will look into it

finite sparrow
#

was curious cause the thing is 6ppm but only 17 procs

#

is that weird?

#

thought the 17 was not counting refreshes

grave wharf
finite sparrow
#

oh

#

the rppm is overridden in its implementation

grave wharf
knotty oriole
#

I presume it was tested at some point if Emallson made the change

#

I can ask him about it

grave wharf
#

weird implementation, would be pretty easy to tell in logs if the uptime is actually 30% vs something closer to 50-60%

knotty oriole
#

https://github.com/simulationcraft/simc/commit/81d0920b4b46d5e3b3bec27f984c8c0e93c22b9d
https://github.com/simulationcraft/simc/commit/a4643b8f6a4d3d7d8fd9f322182a82a5890526c2
@finite sparrow

GitHub
[Stats] Fix refresh data collection for DoTs · simulationcraft/simc...
  • Due to various refactors over time, stats->add_refresh() was only being called in extremely narrow situations
  • Move stats collection to the main refresh() function to collect data for all...
GitHub
[HTML] Add refresh count to the detailed ability chart · simulation...
knotty oriole
#

Something to keep in mind is that Sadistic Glee does have the "Proc cooldown on a per target basis" flag on it which we have only modeled somewhat recently in SimC (I believe after SoD was current they started doing it more for things like Weaponmaster and such)

#

So this probably needs to be tested more

#

May just work fine though

#

I retrofitted the general proc logic with the target_specific logic a bit ago so it should work, but all needs to be double-checked in-game for AoE probably

knotty oriole
#

Don't really think the target-specific stuff is super relevant though considering Outlaw's lack of non-proc yellow damage

finite sparrow
#

awesome, thanks a bunch

knotty oriole
#

Based on some quick testing it seems like it may use target-specific RPPM in a very strange way where it does have increased trigger attempts in AoE but they only go on your primary target

#

ST uptime seems about right though based on looking at some more recent logs

#

Seems around 26% in most Guardian logs from this week

#

But there could definitely be something funny here in AoE (which would only increase the value, but not linearly like one might expect with it triggering on secondary targets with equal frequency.. ends up triggering a lot of refreshes, but my swarm dummy testing had like 60% uptime or something)

finite sparrow
#

if only I had the weapon to test

knotty oriole
#

So based on looking more at the logs, kinda leaning to thinking it maybe gets to 6 RPPM if you are behind and the target is slowed

#

Despite the tooltip saying "or"

#

Might be 3 RPPM with one or the other but 6 RPPM with both or something

#

Going to test more to isolate

knotty oriole
knotty oriole
#

Well initial tests with both behind seem to show that Crippling+behind is significantly better than boss+behind (probably the full 6 RPPM)
Tomorrow I will test front+crippling for non-boss as well as the impact of cleaving to non-boss with Blade Flurry.

waxen hamlet
#

This is only relevant against targets that can be CCed right?

knotty oriole
#

Yes. Will change the evaluation for dungeon use mostly.

rich temple
#

for outlaw, is there an amount of crit at which it becomes notably better to BtE on CD rather than following the 'minor minmax' to only BtE w/RP or to refresh

half karma
finite sparrow
#

yea looks like at some level of crit it is just neutral instead of a small gain

rich temple
#

nice, do you remember offhand what the base crit % of that profile is?

#

just out of curiosity too if you happen to have it, what's the gain on not BtE-ing on CD?

finite sparrow
#

that profile is 40.37% and the gain for venthyr bb is <0.2%

knotty oriole
#

@stray bolt moving here since I think this channel is a bit better to discuss this

#

I did some testing after raid as well and it does seem like it's possible it's applying the damage bonus to all weapon damage abilities (but not AP-only abilities)

#

Not sure what they did to do this, since this wasn't the case before (quite a few people tested this)

#

But definitely appears Sinister Strike, Pistol Shot, etc. are affected. But I gotta test some more things

stray bolt
#

@knotty oriole It seems a fix has been implemented, but a wrong one.

#

take off weapon and equip it -> normal
normal + signet -> normal
normal with signet - signet -> weapon damage buffed but not AA
buffed +signet -> normal again

knotty oriole
#

Well there is definitely a bug that I saw while testing where the AA bonus gets stuck or inverted after swapping shortly after combat, at least on dummies

#

Relogging appears to return things back to normal

#

That is kinda a different issue

stray bolt
#

I equip the signet and only aa buffed, but after taking it off my SS increased. I take off my weapons and re-equip it, damage gets back to normal.

knotty oriole
#

Relogging fixes the odd issues with swapping the ring but still see an increase to abilities after doing that

#

But swapping the ring in-game definitely has some very strange effects

#

I don't know, the whole thing is quite strange and buggy

#

I think the way they are injecting it is into the weapon AP calculations but swapping the ring itself doesn't actually change those cached calculations

#

So it doesn't update until your Agility changes

#

Celestial Guidance proccing seems to force it to update

#

e.g. relog I will do 80 damage with Poison Knife until Celestial Guidance procs then I will be doing 125 damage even after the buff goes away

knotty oriole
#

The further I dig into this, the stranger it gets. The ring is definitely having some interesting effects on a wide variety of skills and may also be normalized differently based on MH weapon type.

rustic sphinx
#

very likely MH weapon based

knotty oriole
#

Not strictly MH weapon based for all abilities, some appear to use OH or Hybrid AP coefficients but many attacks are

rustic sphinx
#

what a headache this ring turned out to be

knotty oriole
#

Whole thing is quite strange, I've been testing for a very long time trying to figure this out and still experimenting with a few different ways this may work

#

Have some ideas but nothing has quite matched perfectly yet

fluid drum
#

Where was the original comment that started this investigation and finding?

noble pendant
#

was a conversation in assassination where he posted details screenshots last night

#

nothing crazy, can just lookup from:lament has:image probably

viscid oracle
#

Good thing I farmed both sets just in case. When are sims going to be updated for this?

finite sparrow
#

should be around 1am cst unless seriallos decides to rebuild raidbots sooner

tough hornet
#

Is this a rogue unique bug or other melee classes (havoc, monk etc) also have this interaction with jeweled signet?

noble pendant
#

melee

spiral light
#

Fuck

junior jetty
knotty oriole
#

It doesn't specifically, it's just that Spirit Wolves themselves are set to inherit AP from their owner so it's just the same as any other ability we'd have

#

This is different from the pet bugs with the ring in S3

#

Since this buffs typed AP, it basically will work on any ability or spec that uses WDPS->AP conversion, which is all the melee specs at the very least. And possibly Hunters.

#

Most beneficial for Feral due to their attack speed

#

For them it is about 10%

hearty olive
#

feral is going nuclear holyyy

fluid drum
#

Ahh, for some reason the original convo wasn't popping up for me. But yeah, this is a crazy finding. Good stuff

regal agate
#

Just to keep this in mind, the ring could get bug fixed given the impact and 100% will for DF if it stays in the loot pool.
So the benefits of getting it might be only temporary.

sly meteor
#

Back to CoS prison feelscryman

sage estuary
#

Better than vault prison

dusky oracle
#

Yeah i really hope this is a bug and im not perma 1k behind the entire tier lol

terse prawn
#

glad i gave up farming it in NA due to it not simming that crazy last week Gladge perma behind all tier now

regal agate
#

please keep in mind that regular discussions should happen in the spec channels ( #subtlety, #assassination, #outlaw )
This is to reserve this channel for theory crafting questions and discussions.

mystic jolt
knotty oriole
#

copy="Jeweled Signet of Melandrus 298 (Slot 1 / 16 Haste) (2),Vyrîe"
finger1=",id=134542,ilevel=298,enchant=1600haste"

#

Yeah I wonder 🤔

#

Real sim

#

So no 😛

finite willow
# lavish zenith still stand by the others tho

I have been doing the same-ish to a select degree because I have noticed this too. I wouldn't say I'm fish'n for a shark boi but while it's up and have a flow of SBS and mobs are falling off on the backend of a pack I ride it out over directly rerolling. This feels more of a "feel good moment" than actual gain if you factor in the CDR you get from the SBS-Dispatch synergy that would prob let you roll into more Buffs in the long term.

finite sparrow
#

if triple threat is 21% chance then shouldn't the average number of hits be 81.9 * .21 in this case

unknown.png
#

what could be causing tt to have 21.8 procs instead of ~17.2?

knotty oriole
#

I believe it's assuming the second TT hit has a chance to proc TT again since it calls the SS OH hit again. But looking at logs that likely isn't accurate.

finite sparrow
#

i had a suspicion that thing was oversimming a bit

knotty oriole
#

This would be pretty small though

#

Like 0.2% or something?

#

I'll test

finite sparrow
#

comparatively yeah, but ~25% worse? 😄

knotty oriole
#

Probably not totally linear due to energy overcapping but I'll look

knotty oriole
#

Although guess I need to specifically test this just to double confirm

#

Still seems ahead of SoH even for WM builds so probably nothing changes anyway

finite sparrow
#

ill afk dummy a bit

viscid oracle
weary scroll
#

Hopefully I'm on the right track with this one, but in sims it looks like TT might proc off itself infinitely, or at least after a large amount of chances. I put the 21% in a calculator and got that result, which is close to the linked sim and to other sims I tested. I also tested lower ranks of TT and that seemed to follow as well.

IMG_1668.png
knotty oriole
#

Was just a matter of confirming if that was possible or not. Usually these types of things with cyclic calls on Blizzard's side actually can

#

(Like some Havoc DH mechanics and such)

#

But it appears not in this case. Likely TT proc is attached to the primary Opportunity proc script and not the impact of the OH ability itself

opal whale
#

Hey, I was just playing around with things relating to the scars trinket because I got it in my vault. I noticed that it would never blow up the final rune on my sims.

https://www.raidbots.com/simbot/report/rwszW6WufQ9DSGY6dBKheh

I looked at the line for scars and it didn't seem quite right
actions.cds+="/use_item,name=scars_of_fraternal_strife,if=!buff.scars_of_fraternal_strife_4.up|fight_remains<30

If its sub 30 then it'll be 29.99999 so it'll never land the explosion because it lasts 30s so I altered the line to be

actions.cds+="/use_item,name=scars_of_fraternal_strife,if=!buff.scars_of_fraternal_strife_4.up|fight_remains<31

for a small gain
https://www.raidbots.com/simbot/report/jUin8eZNmms1QeSx44zAQz

might be more potential with scars idk. Not an expert by any means 🙂

noble pendant
#

makes sense to me, thanks for finding that oversight

knotty oriole
#

I suspect some timing thing happened with this trinket setup in SimC at some point since this condition did something at some point in the past. But hadn't re-checked it recently. Will make the change for sure as it's simple enough. May just make it <35s or something to avoid health estimator issues.

#

Given the recent buffs to the trinket does seem relevant here for simming purposes, at least for Assassination and Sub. Probably will continue to be awful for Outlaw by the looks of it.

opal whale
#

I played a round with it a bit during lunch to see if there's potential with it. Unrealistic sim, but was to mainly to compare explosion to just pure stats in an AoE Patchwerk situation.

https://www.raidbots.com/simbot/report/kdjaRsCC4cqAVMN2FQ8WyL
With explosion (I removed the if condition on the line, so it should just spam it)

Vs

https://www.raidbots.com/simbot/report/dxW3SM6LBCYmv6QnMH4iBh
Default APL that will just keep the stats and the explosion wont proc at the end (<30s)

I played with raid_event.adds.in<29-raid_event.adds.duration instead of the <31s condition on dslice to try and get it to activate on adds and got this
https://www.raidbots.com/simbot/report/gUYoWBQ1npSxMoN4QAggJH

Idk, will look at it a bit more when I get some time, might not bear fruit though and I'm very new at changing APL things.

regal agate
#

This was a old optimization i did, but i am not sure it still holds up (assassination):

- actions.cds+="/use_item,name=scars_of_fraternal_strife,if=!buff.scars_of_fraternal_strife_4.upfight_remains<30"
+ actions.cds+="/use_item,name=scars_of_fraternal_strife,if=!buff.scars_of_fraternal_strife_4.up|buff.scars_of_fraternal_strife_4.up&cooldown.vendetta.remains<=0&fight_remains<=60|fight_remains<=30"```
half karma
#

Instead of doing multiples sims

regal agate
#

Adding this to the top also helps seeing the relative loss:chart_show_relative_difference=1 relative_difference_from_max=1

opal whale
#

Thanks for the info 🙂

knotty oriole
#

But still could be beneficial potentially

regal agate
#

There is potential to consider flag too, but i haven't tested that

#

i think whispyr is atm looking into improvements, maybe he found something

knotty oriole
#

Whispyr was also working on evaluating some raid_events changes probably similar to what you are trying @opal whale but this probably needs further iteration. All this seems worth looking into now that the trinket is buffed a fair bit.

#

There wasn't too much gain playing around it before since it was quite bad before the hotfixes. But now does seem strong enough that it could be worth.

#

Just need to figure out the new target thresholds

finite sparrow
#

how do you sim 2 target cleave but the mobs are not bosses so they can be slowed?

#

i know you can use a raid event like raid_events+=/adds,name=CleaveBoss,count=2,cooldown=300,duration=300 in advanced sim, but then how do you force raidbots to ignore the patchwerk 1 target that is selected at the bottom

#

afaik adding that would make it 3 targets, 1 boss (from the bottom settings) and 2 non bosses (from raid events)

half karma
#

from dslice

#

I think that should do waht u want

finite sparrow
#

that's right

knotty oriole
#

This is how it works on the initial boss for DS sims

#

/adds,name=Boss,count=1,cooldown=500,duration=135,type=add_boss,duration_stddev=1

#

This will count as a boss for the purposes of CC effects

#

Or wait I'm dumb

#

I thought you were asking the opposite, my bad 😄

finite sparrow
#

i didn't know that either, was that always a thing?

knotty oriole
#

I added it some time ago. It's not super new but semi-recent this year.

fluid drum
#

@finite sparrow do you have the sims for using BTE on CD vs BTE conditionally for the various bounty stacks?

finite sparrow
#

yes, with no crit stacks it's just a small minmax +0.2% or so. with 10 stacks it became neutral.

#

i didn't sim 5 stacks or 15 stacks or anything else. if someone wants to find the exact point it is neutral instead of a minmax, it'll depend on their initial crit

#

and i'd still do the minmax anyway, since it helps with target swapping

fluid drum
#

Awesome thanks, so I don't have to harass Koji for the sims 😂

finite sparrow
#

I think using invulnerable raid events may be conflicting with blade flurry targeting

example https://www.raidbots.com/simbot/report/mKsLB9GpW1q7GM9VzJzzoY:

raid_events+=/adds,name=SlowedBoss,count=4,cooldown=300,duration=300,duration_stddev=1```
this is invulnerable + 4 targets, 5 minutes

changing it to invulnerable + 5 targets (<https://www.raidbots.com/simbot/report/927Uum37vUSRKctHQuUJGM>) is only a ~0.5% gain, which makes me think the invulnerable thing is getting hit by blade flurry instead of the 5th target
knotty oriole
#

DS and DR fight styles do this automatically, but won't apply with pure custom events

unique zephyr
#

Hello, has there been any headway into how much value the WFT changes will bring?

willow knoll
#

thats something wordup does

regal agate
#

probably a noticable amount but we will have simulations sooner or later

unique zephyr
#

ok cheers

knotty oriole
#

But we don't have exact values yet

thorn gale
#

Is there a Raidbots or SimC for beta? If so, where can I find it.

noble pendant
thorn gale
#

Thanks

#

Whats the best way to import data from the beta server?

noble pendant
#

there's a beta version of simulationcraft on curseforge

#

you can get your character import like that

#

but given how shaky beta implementations are for simc currently, idk how much info you can really get from that

#

I end up just using the template char or making a profile

thorn gale
#

I just what to sim some talent builds with the template character

noble pendant
#

yeah that's fine

#

just remember that a lot of stuff might not be implemented properly and the APL's don't support it fully yet

thorn gale
#

Thanks

solar swift
#

Hello, in df build sim Duskwalker's Patch affects on deathmark (instead of vendetta)

unknown.png unknown.png
noble pendant
solar swift
#

But in the game it does not work on deathmark

spiral light
#

:P

#

No worries, we've been told it should work.

solar swift
#

well then ok, I hope they'll figure it out 🙂

knotty oriole
#

Yeah I put it in advance of the patch due to the comments.

cursive valley
#

But if they make it work on live it’s more than welcome 👍🏻

minor quartz
#

good luck out there this week gents

dusty delta
#

For Outlaw it seems like the dungeon slice sim doesnt use Vanish so no gain from Shadow dust

finite sparrow
dusty delta
#

maybe its just in quick sim?

finite sparrow
#

your /simc is missing your covenant

#

it's because of this

unknown.png
#

generating simc input from raidbots top gear results will remove your covenant atm

dusty delta
#

okay sorry my bad then didnt know

#

so y if i create the simc direct from the addon and quicksim it uses vanish

finite sparrow
#

you had no way of knowing if i didnt tell you

#

since vanish rules with dust are tied to which covenant you are, in the event you have no covenant then vanish doesnt get used

tender mauve
finite sparrow
#

support for them may have been discontinued

tender mauve
#

Gotcha, I'll look into it thanks!

valid relic
#

raidbots still showing NYI for rogue APL's

#

anywhere we can find the APL's?

noble pendant
#

can you link the error?

#

plenty of people have been able to sim

valid relic
#

i mean its here in the SS

unknown.png
#

and also their google docs

noble pendant
#

yeah don't bother with those google docs they're rarely updated

#

that NYI is a general disclaimer I'm pretty sure

#

do you have any issues when you try to sim?

valid relic
#

im more trying to get my rotation helper working

noble pendant
#

the APL's should all be functioning fine

regal agate
# valid relic anywhere we can find the APL's?

just run a sim, you can then just look at it in the html report in the profile section.
For rotation helper, hekili and Hero damage will be updated over time, so you need to have some patience.

valid relic
#

yeah cheers

potent ferry
#

didn't spend a lot of time on it though, might be some holes in the logic

regal agate
#

great work 👍

potent ferry
#

the ambush thing probably needs some conditionals for talents like FW/hidden opp/CTO

knotty oriole
#

Will investigate a bit. I have seen some build setups where it is a loss to remove the low energy lines and I suspect there is probably a more narrow condition for when this is a loss vs. a gain that I'm still trying to identify.

#

the !cooldown.dreadblades.ready condition to me feels a little "accidental" inasmuch as we are currently holding Dreadblades for Flag right now so that is a somewhat indeterminate/unspecific time window and I think there's possibly a more specific check that may be more targeted

potent ferry
#

I intended to dump stacks prior to db

#

But if we sync it with flag we need a much more complicated condition for that

#

I also only tested ds and my char/build

knotty oriole
#

I guess the way I'm looking at it is that I think it could be better to be more specific about when we shouldn't be dumping stacks rather than conditionally enabling when we should based on a variable length of time? If that makes sense. e.g. if it's fine when Dreadblades CD is ready then it may also be fine when Dreadblades is not ready but Flag is finished, etc.

potent ferry
#

The other part of it is to not use opp at 3 stacks of you don't have gsw buff

#

Might be interested to isolate that part to see effect of only that

#

The db interaction might not matter

#

I'll try some more when the kids are sleeping

half karma
# potent ferry The other part of it is to not use opp at 3 stacks of you don't have gsw buff

The rule we were discussing with solo last time was to use OPP procs on cd but only at under 3? Cps, so the idea is to not use procs while building but instead keep them after a finisher (obviously if you get to 6 stacks while building you would finish early akin to bb), I tested a rough version of that against your OPP rule and it seemed to perform slightly better, might be interesting to explore OPP rules a bit more

#

In db tho I would agree with your OPP rule (never use unless 6 or GSW)

potent ferry
#

Unless you have gsw proc or opp proc is running out yeah

#

Making sure you use them on low cp might be more important, or not. Haven't tried

knotty oriole
#

Just to give you an idea of the breakdowns

#

I split out the lines for reporting purposes

half karma
#

What is the purpose of the sinister strike line

#

Oh just for reference I see

knotty oriole
#

Yes

#

This is just for debugging

#

For even more detail

#

(2nd line is the stacking portion not persist)

#

So I guess my curiosity here would be

#

a) The majority of fallthrough SSs are outside of cooldowns

#

b) Dumping stacks prior to Dreadblades does seem to be a small gain

#

But that leaves to me some lack of identification of what conditions are actually beneficial here during the "normal" rotation

potent ferry
#

Then the mail gain should be more related to stacking for gsw procs?

knotty oriole
#

Yeah I do wonder if that is most likely

potent ferry
#

With vanish you can get bte with very little time in between

#

This allows you to not have to hold bte

half karma
#

I don't think so, the rule I explained earlier don't even care about gsw and it's still slightly better

#

There might be more to it tho

knotty oriole
potent ferry
#

Holding for bte will also make you use fth directly after bte

knotty oriole
#

0.2-0.4% can really pop out of nowhere

potent ferry
#

So you may be right zac

half karma
#

Will have to see yea, happy we can already find some dumb rules for less than 1% dps thoCheesin

potent ferry
#

This is just for fun

#

And understanding the spec better

knotty oriole
#

I would not normally care about this for prepatch but I want to determine if this is something relevant for the "real" DF APL

#

In terms of some generalized conditional rule

#

Since FtH+GSW+Dreadblades will all still be there

half karma
#

Oh it should be yeah, don't think dust as anything to do with just a normal FtH rule

potent ferry
#

Should try without dust then

knotty oriole
#

Ok I think I isolated a much better rule

#

Actually

#

One second 🙂

half karma
#

Would only impact GSW really

knotty oriole
#

This is in ST

potent ferry
#

Sweet

knotty oriole
#

(Baseline is with your change)

#

The difference here is instead of the dreadblades cooldown check

#

I made it a CP deficit check

half karma
#

Idk if you did but you should probably look at the finisher rule as well koji

knotty oriole
#

The gain here makes sense as it's more about generator efficiency

potent ferry
#

This doesn't dump stacks before db at all right?

knotty oriole
#

Correct. Or, well, it will be just using stacks whenever it will get 4+ CP

knotty oriole
#

Although was more planning on doing it for DF branch

half karma
#

Well I think this is relevant for both tho?

potent ferry
#

We can probably squeeze out a bit more hehe

half karma
#

Well for me this rule only works if you have a proper finisher rule

#

That's the thing I tried earlier

#

I had to change the finisher rule

knotty oriole
#

I mean it's somewhat relevant for both

half karma
#

So it doesn't send 6 stacks but instead finish early

knotty oriole
#

If we weren't generating 7 CP with like 80% of our generators in prepatch

#

🙂

half karma
knotty oriole
#

But you can see here with that change the breakdown shifts quite a bit

potent ferry
#

Finish early if you have 6 stacks and 4+ cp maybe

knotty oriole
#

Most of the fallthrough SSs with Opp up shift into Dreablades only

#

Which makes sense

potent ferry
#

We should not finish early with just 3 stacks at least

knotty oriole
#

Removes 20 of the fallthrough non-CD SSs

#

7 vs 27

half karma
#

Hm

half karma
# knotty oriole

I don't think it's useful to have a special line for GSW, since we are only delaying OPP usage by one finisher at most with this new rule, it would be a better cp gain for maybe or even not less gsw

potent ferry
#

Then you have to hold bte a little?

knotty oriole
#

Will check shortly

#

But for full comparison in ST

#

Putting baseline back

#

compared to baseline

#

Basically getting 9-10 more CP out of this

#

Is the main change

#

Some waste is going to be inevitable (especially with Dreadblades)

#

But 10 CP is 20k+ damage so

potent ferry
#

That's why I wanted to dump before db

knotty oriole
#

Yeah I think it was the right idea, just it was being limited to the window between Dreadblades and Flag rather than the entire time between Dreadblades CDs

potent ferry
#

But dumping might hurt the gsw gain (if there is any)

knotty oriole
#

This change basically opens it up all the time outside of DB

#

Funny enough it's kinda backwards how pre-DF APL used Opp procs to fill in low CP gains

#

Now we want to use SS to fill in low CP gains instead 😛

#

I will check BtE

potent ferry
#

Since we have a rule to not send db unless low on cp this rule might actually force dump before db anyway

#

Hmm. No

#

Maybe add a rule to not send db of you have opp proc

waxen hamlet
potent ferry
#

Might mess with flag alignment though

knotty oriole
#

SS's only purpose in this build is to generate stacks and if we're already at max I'm curious what the value would be if you're already capped

half karma
half karma
#

So more cp agajn

waxen hamlet
#

Yes, the idea is to increase your odds of ending DB with 6 stacks

potent ferry
#

Yeah the other side of the coin of dumping stacks before db

#

Note that my build is using imp bte, so holding bte is even worse than normal build

half karma
#

and db is a bit too good to waste cdr on it

knotty oriole
#

I think there can be a narrow gain for this but it's far smaller with the adjustments I've made now

potent ferry
#

But what is the gain if db except cp?

half karma
knotty oriole
#

Seems only like 0.2% difference with the changes to PS usage in my version

potent ferry
#

Main issue with dumping before db should be the alignment with flag

#

But it takes at most 4 seconds to dump so maybe we can do something smart

knotty oriole
#

I think it's more the point that there isn't any reason to specifically dump before Dreadblades. You should just be using for max CP all the time outside of Dreadblades.

#

And only banking charges when CP deficit is low

potent ferry
#

The idea is to get an extended period with 1 builder per finisher

knotty oriole
#

Right but there's no reason to "dump" specifically before Dreadblades if you're not sitting stacks to begin with

potent ferry
#

Consider this sequence: SS (procs opp), dispatch, db
Vs
Ss (procs opp), dispatch, ps, dispatch, db

#

I think we want to force the latter somehow

knotty oriole
#

Adding a conditional to DB to delay seems to be negative at a glance

#

Anyway I think the altered PS trigger conditions make sense to put in since the net gain is rather large at this point.
However, I do need to also revisit if the commented out low energy PS/non-QD lines serve any purpose anymore.

#

I assume they do for the Ambush builds

knotty oriole
#

At least with my more aggressive timing adjustments

#

There doesn't really seem to be any waste here

half karma
#

oh totally forgot you could not take fth xd

knotty oriole
#

yeah lol this is only a prepatch consideration

#

I do enjoy some of the talent tree stuff actually makes life easier

#

Some things are harder, but not all thankfully 😄

#

Same working on the Momentum stuff for Havoc, because it's required to have Unbound Chaos to get Momentum it makes things much simplier lol

#

Ok, on a roll now 😄

half karma
#

what is the gsw merge

knotty oriole
#

actions.build+=/pistol_shot,if=buff.opportunity.up&(buff.greenskins_wickers.up&!talent.fan_the_hammer|buff.concealed_blunderbuss.up)
actions.build+=/pistol_shot,if=talent.fan_the_hammer&buff.opportunity.up&(buff.opportunity.stack>=buff.opportunity.max_stack|buff.opportunity.remains<2)
actions.build+=/pistol_shot,if=talent.fan_the_hammer&buff.opportunity.up&combo_points.deficit>4&!buff.dreadblades.up

#

Disabling the GSW condition with FtH

#

So legendary only

half karma
#

oh wow didnt expect it to be that much better

knotty oriole
#

Probably also a fair amount of CP waste

half karma
#

I must have messed something up when I did my tests haha, tried the same thing but wasnt near as much of a gain

knotty oriole
#

A lot of this stuff is kinda a house of cards, it could have been getting munched by the other lines

half karma
#

you are still not doing any special finisher rules right? maybe I messed that up

knotty oriole
#

I think I will leave these lines split because it's much more clear to debug

#

And we'll be removing all the SBS lines so it'll be fine 🤣

knotty oriole
#

Shurtugal's work has some updated logic that I was going to profile soon

potent ferry
#

So cp efficiency is more important than getting off as many gsw procs as possible

knotty oriole
#

I don't think it really ends up overflowing any GSW

#

Given the conditions in which the charge usage is happening now it would be pretty unlikely not to cast PS between BtEs

#

If not impossible outside of some Shadowdust sequencing

#

After all this build is casting Pistol Shot every 6s on average anyway 😛

vital anchor
#

Oh god just forget what I said

knotty oriole
#

I mean it is... exceptionally unlikely 🤣

#

But I suppose it could happen

vital anchor
#

I read the last condition as buff.opportunity.stacks<2

knotty oriole
#

yeah np!

#

Yeah

buff.opportunity.stack>=buff.opportunity.max_stack
is the generic way of checking this that should work with all ranks

#

If they change anything

#

(Even without FtH this would work)

#

I need to determine if that line is appropriate for non-FtH builds but haven't checked yet

#

Seems good

#

🤣

vital anchor
#

I mean, without FtH, GSW or BB, pistol shot is never used by the 3 lines you proposed.

knotty oriole
#

Yeah the previous versions commented out the low energy PS lines

#

That line still seems to be ok at a glance for the other builds, at least until the Ambush builds get further optimized

#

To just have a !fan_the_hammer check on it probably

vital anchor
#

I tried it on lvl 70 builds, got the low energy PS line to be a gain with a talent.hidden_opportunity check (not having that line is a loss for HO+FtH builds).

knotty oriole
#

A task for another day

potent ferry
#

did a few more versions, some stuff is really really volatile

#

the only thing that seems to be a clear gain is a small adjustment for GSW, add something like this
actions.build+=/pistol_shot,if=talent.fan_the_hammer&buff.opportunity.up&buff.greenskins_wickers.up&(cooldown.between_the_eyes.ready)

#

can't see clear gain with changed finish condition or dump before db, results for those seem to be extremely volatile

#

but even the GSW thing is 0.1-0.2% at most

knotty oriole
#

Yeah I tested a BtE condition there but it seemed not really to do much of anything. Maybe that changes at some point, but it will be even less important without Shadowdust so that doesn't worry me too much

potent ferry
#

yeah, probably not worth bothering with

knotty oriole
#

Ok, all this stuff should be cheked in and picked up in the next raidbots build. Thanks everyone for the iteration and feedback!

wet mountain
potent ferry
#

apl after kojis change only does ps with 6 stacks during dreadblades (or if opp proc is about to run out, but that's unlikely)

#

these two lines:
actions.build+=/pistol_shot,if=talent.fan_the_hammer&buff.opportunity.up&(buff.opportunity.stack>=buff.opportunity.max_stack|buff.opportunity.remains<2)
actions.build+=/pistol_shot,if=talent.fan_the_hammer&buff.opportunity.up&combo_points.deficit>4&!buff.dreadblades.up

#

first one can trigger in and out of dreadblades

#

second only outside of dreadblades

stray bolt
#

For exsang in DF, should it be a dps gain if cast a exsang's remaining cd+12s rupt and cast a 7cp rupt with ER when exsang's cd also up for a rotation of 21s of exsang'd rupt every 45s?

rich silo
#

Hi folks. On line 25 of the outlaw.simc file, there is some arithmetic with talents and buffs: cp_max_spend-buff.broadside.up-(buff.opportunity.up*(talent.quick_draw|talent.fan_the_hammer). How does that work? Do the talents return 1 or 0 if they are selected or not? What about talents with multiple ranks? Just trying to work out the calculation here.

noble pendant
#

that line basically says to finish earlier when you have broadside, or an opportunity proc with a pistol shot talent to avoid overcapping

rich silo
#

Ok, thanks! 🙂

knotty oriole
#

Yeah just binary arithmetic in this way. Basically it is "lower the CP threshold by 1 if you will get extra CP from casting Pistol Shot"

vital anchor
#

And if you want to check the rank you do talent.fan_the_hammer.rank.

regal agate
earnest raptor
#

Im having a weird bug but its about m+ in general, not about outlaw

#

Where should I post it and what kind of info should I share to help ppl work on it?

#

Like, logs, recording of my screen etc

half karma
#

there is also a bug tracker

#

!bug

plush slateBOT
stray bolt
dapper nacelle
#

I'm pretty new at the APL stuff, does this value=combo_points.deficit>=2+talent.improved_ambush+buff.broadside.up&energy>=50&(!talent.count_the_odds|buff.roll_the_bones.remains>=10) mean that the variable is when you have 5 or fewer CP AND are talented into improved ambush AND have broadsides up AND have more than 50 energy (and dont have over 10 seconds on roll the bones buffs)?

#

or is it that you have 5 or fewer CPs AND (either improved ambush OR broadsides ) AND more than 50 energy... that makes more sense but im still not entirely sure how to read it as that :3

noble pendant
#

its easier to break into chunks imo, and from there it's just boolean logic
combo_points.deficit>=2 = do you have 2 or more empty cp
+talent.improved_ambush = add 1 if you have improved ambush

now we get a big conditional
+buff.broadside.up = add 1 if you have broadside right now, but only if we also have &energy>=50, AND we have either:

  • !talent.count_the_odds no count the odds
    OR - buff.roll_the_bones.remains>=10 roll the bones has 10 or more seconds left

so in context, lets say you have both deeper strategems, giving you a max of 7cp. A combo point deficit of >=2 would be 5 or less cp. If you have improved ambush, we lower it to 4, if we also have broadside, more than 50 energy, and roll the bones has 10 more seconds on it, we lower it to 3.

Alternatively, if we only have 6cp max, the deficit would be 4 or less. We don't have improved ambush, so skip that, but we do have broadsides, and we have roll the bones >=10 seconds left, but we have less than 50 energy, so we skip that too, leaving us at 4

dapper nacelle
# noble pendant its easier to break into chunks imo, and from there it's just boolean logic ``co...

thank you blushies i got confused that the + was adding one >.< i've never done any kind of coding really, is there like a key or something anywhere morigiggle1 i can kinda figure it out but its more using what i know about the class to figure out what it says and id really like to be able to do it the other way around - aka I'm hitting a target dummy and thinking about how to optimize vanish better, so i thought, well someone finally linked a sim i can look at ill just read that and look at the conditions theyre using, um... turns out, reading it is kinda hard

noble pendant
#

also just for max clarity, the operators in simc are:

+ #addition
- #subtraction
< #less than
> #greater than
= #equal to
& #and
| #or
! #not (invert)
* #multiply
% #divide
%% #mod
#

forgot one monkaWW

dapper nacelle
#

oh my gosh youre a saint :3 in case theres anyone else like me out here, purrhaps pinning that in here could be useful too?

noble pendant
#

Let me see if there's a wiki page that would be better to pin

#

yeah this is way more involved

#

the wiki is already pinned under

#

the APL wiki here

unknown.png
dapper nacelle
#

ahh okie, i see, thanks ❤️

regal agate
# stray bolt WLOG, we need builds with cheat death for comparisons. maybe?

it makes sense to compare the highest dps builds, skipping utility for dps is also worth diffrent amounts depending on the spec.
e.g. subtlety gains the most of it (with ~3-4% dps), while outlaw and assassination both have lower impact to the point this option is not appealing (assassination esp. in a ER build does gain no dps from it given how weak nightstalker is without shadow dance)

stray bolt
# regal agate it makes sense to compare the highest dps builds, skipping utility for dps is al...

Maybe it would be better adding some tag to notify that sub's sim has a significant gain from not spending two points for the major utility of cheat death. I don't know if we're going to shift to a no-cheat_death normal since any top build for a spec utilising SF and dance will get into this dps gain for no cheat death, which is however a bit of misleading considering cheat death isn't some cheap utility we can just throw away.

stray bolt
half karma
#

he only did these sims to give a rough idea as to how much of a gain the buffs were for the differents specs tho you shouldn't look too much into it for now

stray bolt
west prism
#

Apart from small gains being found for each specs APL (Outlaw probably finding the most), it's actually quite indicative of where the specs stand relative to eachother atm.

half karma
#

yea I more so meant regarding utility dps value like cheat cheat and such since he was asking about that, the differents specs will probably have detailed informations about dps gain/loss regarding utility nodes in the future

magic wolf
#

Hello, i know this is not a help channel.. but i want to research if this APL String is a DPS LOSS without the Opportunity Condition. I think i can manage to change the string, how i want it.. but where do it put it in raidbots? 😄 pistol_shot,if=buff.opportunity.up&(buff.greenskins_wickers.up&!talent.fan_the_hammer|buff.concealed_blunderbuss.up)|buff.greenskins_wickers.up&buff.greenskins_wickers.remains<1.5

noble pendant
half karma
#

careful about not taking the otl-df if you want to sim for prepatch

magic wolf
#

ok 7 dps loss 😄 thanks all

half karma
#

by consuming gsw procs without opp?

#

or were you trying to achieve something else

#

because I highly doubt consuming gsw without op is only 7dps loss lol

magic wolf
#

i tried to sim if its worth to use the last minute procc for gsw without opp ( the last seconds)

half karma
#

ah, but thats already in sims

magic wolf
#

yeah but with opp procc i think

#

i tried it without

half karma
#

no the line you sent has a OR at the end

#

"|"

#

this thing

magic wolf
#

ohh 😄 my bad .. i think my hackers carrier ends here

#

😄

half karma
#

dw its cool that you are interested in testings things yourself tho

magic wolf
#

ty, It's always fascinating how you guys play with the numbers, so I wanted to try it myself 🙂

half karma
#

PirateHeart keep trying things

regal agate
regal agate
#

if you want to do a before/after sim, you can just paste this below your /simc in advanced sim.

📎 Spell_data.txt
regal agate
regal agate
#

this list should include all DF s1 Raid/m+ items, can just paste it below your /simc in top gear.

📎 gear.txt
regal agate
#

just in case someone needs them, mythi 0 trinkets. (can just copy/paste in top gear like the above)

📎 mythic0.txt
regal agate
#

If you want to sim Dragonflight consumables, you can paste this below your /simc in advanced sim.

📎 consumables.txt
stray bolt
#

I wonder if anything is missing about thistle tea. We have 3 charges at beginning and 1 more per minute, but there are only 4.7 uses on APL
It seems we never drink tea in dm according to sample APL and it only uses 5 charges where some casts are missing.
More tea would lead to a bit more dmg, which will be diminished with longer fight.
At least some extra room for optimisation here.

unknown.png unknown.png
golden briar
#

forcing tea in opener and then just not letting it overcap should be easy enough to implement in simc

noble pendant
#

Not drinking tea during cooldowns is by design since the mastery isn’t worth playing around

#

But there’s probably wiggle room to use energy more aggressively

willow knoll
#

Seems like a no brainer fix lol

#

I assumed that's how tea was functioning since we had pretty minor tea optimizations already pushed internally

spiral light
#

This is for Assa, but probably similar results for other specs if it's not working for Sub/Outlaw either:

actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&energy.deficit>=100&(charges=3|debuff.deathmark.up|fight_remains<cooldown.deathmark.remains)|!buff.thistle_tea.up&debuff.deathmark.up&fight_remains<60|!buff.thistle_tea.up&fight_remains<12

Sync's the tea with the last Deathmark, sends the last tea on its own if ttd < 12s. Can probably optimize slightly bit more.

image.png
noble pendant
#

all within margin peepoHands

spiral light
#

found one 0.4% by slapping some energy-shaishe on top

#

but it's iffy

noble pendant
#

look what they did to my boy

knotty oriole
#

Outlaw already has modified lines

#

But Outlaw only really uses it for energy because Mastery is just whatever random BS

finite sparrow
#

You mean we don't link it with Killing Spree? Sadge

knotty oriole
#

both the best place for the Mastery and the worst place for the Energy 🤣

#

Actually probably technically the best place to use it for Mastery is using it prior to Blade Flurry initial cast in AoE since the hits can trigger MG now

spiral light
#

tea is just a weird talent, there's rarely if ever a good place to get full benefit from it

#

you have to sacrifice either the energy or the mastery

spiral light
#

Noticed that the sim is actually holding onto 2 charges at all times and only using if energy deficit is >100 and DM is up. Simply allowing the sim to use tea during Kingsbane windows when energy deficit is >100 shows a noticeable gain.

#

actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&energy.deficit>=100&(charges=3|debuff.deathmark.up|fight_remains<cooldown.deathmark.remains|dot.kingsbane.ticking)|!buff.thistle_tea.up&fight_remains<12

*cleaned up the line slightly

hearty knot
#

100k assa PauseChamp

noble pendant
#

DTB in the dirt?

#

it's happening

willow knoll
#

Needs another 15% buff

#

Nerf outlaw

#

It's unfair that it does any aoe damage, meanwhile assa is doing 20k more st than most specs in the game KEKL

proud wadi
#

Weird, I could've sworn this is tc-research not wow-general or a spec channel

regal agate
#

^ please stay on topic in this channel, it will be the heaviest moderated one

raw knoll
spiral light
# raw knoll I asked Fuu to try to sync it with CDs for sub since the apl wasn’t using tea un...

Yeah like Whispyr said earlier, syncing tea with cds hasn't shown gains previously so we just default to using it for energy. However the sin apl wasn't using all the charges during the fight so we would end the fight with 2 charges instead of 0 which is what changed. Also forcing the sim to use tea charges during Kingsbane for energy and the mastery proc. But again it's only a gain outside of Deathmark windows where you gain the benefit of the energy and the mastery proc, using tea during all cds the beginning is still a loss. Sin apl also uses the first charge at around 45 seconds currently even with all of these changes.

raw knoll
#

assa I could maybe understand if it pops them during execute.

knotty oriole
spiral light
knotty oriole
#

I'd imagine DTB still has something it wants that is better though

#

fwiw the Outlaw line is

#

actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&(energy.base_deficit>=100|fight_remains<charges*6)

spiral light
#

@knotty oriole Came up with a really good Tea line for Kingsbane and non-Kingsbane builds:
https://mimiron.raidbots.com/simbot/report/ptsnNtstQoKJrRuUmQ63nK

actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&(energy.deficit>=100|charges=3&(fight_remains<cooldown.deathmark.remains|dot.kingsbane.ticking|debuff.deathmark.up)|fight_remains<12)

I had 2 separate lines for KB builds and non-KB builds at first but managed to merge them into one single line. Had to rearrange and pretty much rewrite the entire line from scratch.

image.png
Raidbots
Advanced Sim - Zzz - 98,461 DPS - Raidbots
knotty oriole
#

Does it actually need the fight_remains<cooldown.deathmark.remains check anymore?

spiral light
#

I tried without it several times and it was a 0.2% loss everytime

noble pendant
#

Holy

#

What the fuck

knotty oriole
#

Did you try something like the Outlaw line I posted?

#

Instead

#

fight_remains<charges*6

spiral light
#

~~I did yeah, was quite a bit worse than having all the checks. ~~nvm I'm dumb.

knotty oriole
#

No I don't mean in place of the checks

spiral light
#

Yeh

knotty oriole
#

actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&(energy.deficit>=100|charges=3&(dot.kingsbane.ticking|debuff.deathmark.up)|fight_remains<charges*6)

#

or something like that

#

idk I'd check but I've got like 3 plates spinning atm lol

spiral light
knotty oriole
#

Nice I'll drop that in tonight

#

I have some other minor changes

lavish zenith
#

can one of u guys post a tldr for my dumbass? its literally just tea rules from before + use it during deathmark + use it during kingsbane? icant

unique zephyr
#

Isn't there also ER now?

noble pendant
unique zephyr
noble pendant
#

that's bis gear

unique zephyr
#

424 or smth right

#

cheers

noble pendant
#

422

#

but it's like a mix of 421 and 424

quasi mulch
#

How can I reduce the uptime of Corrupted Rage in sims?

noble pendant
#

think it's a little jank atm

knotty oriole
#

No controls for it yet

#

Assume it will be added a bit later, but lots going on

stray bolt
stray bolt
knotty oriole
#

Should be dot.deathmark.ticking probably but I just checked and it made no real difference to the performance of either spec

stray bolt
#

Thx. a bit of sad feeling dtb does not enjoy drinking tea like kb.

regal agate
#

Subtlety also had a small room for improvement with Tea. (sadly no 2% improvements, we are already fairly optimized)
https://www.raidbots.com/simbot/report/uaaMn5Es3ojQDpEq1BZRcw```diff

  • actions.cds+=/thistle_tea,if=cooldown.symbols_of_death.remains>=3&!buff.thistle_tea.up&(energy.deficit>=100|cooldown.thistle_tea.charges_fractional>=2.75&buff.shadow_dance.up)|buff.shadow_dance.up&!buff.thistle_tea.up&spell_targets.shuriken_storm>=3|!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)```
image.png
spiral light
stray bolt
spiral light
stray bolt
#

I think Koji's comment would work. tea does not require gcd, which leads to plenty opportunities to use right after you cast something. we would be capping energy but not like at 100% all milliseconds.

spiral light
#

The apl is using tea at the same time as other gcds so it checks the energy and the only time that it would be able to use tea is in between gcds, but that"s Koji stuff.

#

Yeh exactly.

potent ferry
#

I don't see how that could matter? You can't use energy between gcds anyway? Isn't it better to see how much energy you have once it is time to use the next ability? If you think tea is not used enough then tweak the energy threshold instead?

spiral light
prime narwhal
spiral light
#

Tea is probably implemented wrong in simulationcraft then if you can use it at 100% energy in-game now.

proud wadi
#

but before you 100% couldn't

spiral light
#

Yeah, I was still under the impression that you couldn't. But good to know.

prime narwhal
#

it got changed after one of the HC raid tests.

stray bolt
regal agate
#

Yes, the 2nd part of the condition is the one that will trigger regular when only playing with one charge.
condition: shadow_dance,if=variable.shd_combo_points&fight_remains<cooldown.symbols_of_death.remains|!talent.shadow_dance&dot.rupture.ticking&spell_targets.shuriken_storm<=4
The only condition it has is that rupture is up and the target count is below 5 targets

#

the condition in essence skips the usual constrains that Shadow Dance is tied to on lower target counts

plucky mural
#

@knotty oriole I feel for you guys this time around. So many build alternatives.

  1. Does haste still increase Energy rate (added in BFA)
  2. Are we expecting to see high crit build variants with seal fate? Or is it not really that impactful ?
knotty oriole
#

afaik Haste mechanics are still the same. As for Crit, seems solid for all the specs right now.

chrome hound
#

Does anyone have a resource showing which rogue spec is likely to be preferred for each boss in the upcoming DF raid?

regal agate
maiden cargo
#

@knotty oriole how robust are the crafted item implementations? Not just JC neck (which doesn't seem to include secondaries), but the proc effects? https://www.wowhead.com/item=193451/slimy-expulsion-boots

On a simple first glance these sound like they would drastically outperform any other boots, but don't know how simmable they are yet

Wowhead
Slimy Expulsion Boots

This epic leather armor of item level 350 goes in the "Feet" slot. It is crafted. Added in World of Warcraft: Dragonflight. Always up to date.

knotty oriole
#

My guess is that the passive Haste reduction doesn't work but the damage proc does. But I'd have to look

maiden cargo
#

Unless the proc rate is atrocious I have a hard time believing those aren't pretty insane, or any of the crafted items (vault 2 set frost damage wrist/waist)

tiny narwhal
#

its 4rppm

#

hover over the text in wowhead

#

There are quite a few items that seem to be very good

maiden cargo
#

700 damage every 3 seconds on the helm at 350 ilvl too

#

The items seem pretty undervalued in sim profiles atm.

#

It looks like the snowball set is best for all three specs by a sizeable margin, but I can't get the neck to sim properly and I'm not fantastic at Sims. Above my pay grade at this point

tiny narwhal
#

I don’t think all the effects are properly implemented on the backend

maiden cargo
#

Maybe not, but this is probably for fuu/Koji/whispyr/seli, I'm out of my depth

maiden cargo
#

Someone may want to confirm/re-run things, but I'm seeing +1k for sin to run the crafted 2set, +950 for sub to run it, and +2k for outlaw to run flaring cowl+slimy expulsion boots

#

Compared to the bis profiles from fuus sheet.

regal agate
#

will look into adding crafted gear when i update the sheet the next time

stray bolt
#

@knotty oriole I think if a SecT is cast in the middle of a tornado, the CB gets consumed by the auto SS. Put it to the end should be a gain or something wrong there (fuu said it was not a gain). Or we delay that SecT and make a space for CB after the auto SS.

unknown.png
regal agate
#

lower cdr can also have side effects when you get secret up too late to lineup

#

e.g. if i make it more strict, it gets worse and the sample sequence becomes weird

#

^ for example

#

ofc either of this simulations are just surface level looks at it and not modifications i looked into in detail

stray bolt
#

Then I have to push my 0.3 delay strategy here.
tornado - late dance (or dance after 1st SS)- BP - 0.3s delay SecT - CB after auto SS - BP after akkali - BP - SS - BP - SS - BP -(dance finish)

#

we can exploit CB on the SecT shadow dmg and the following BP.

#

This may not show good gain in sim, but in keys we have more opportunities to apply this.

stray bolt
#

For sub ST, would it be a gain if dance 0.5s after SecT and CB following by an evis? So we get crit SecT shadow dmg and a 4/5cp evis.
dance akaari(crit) akaari(crit) 1evis(crit) - 2strike - 3evis -4 strike - 5strike - 6evis - 7strike - 8strike
or
dance akaari(crit) akaari(crit) 1gb(crit+FW) - 2evis - 3strike - 4evis - 5strike - 6strike - 7evis - 8strike
compared to
dance - 1gb - 2evis - 3strike - 4SecT CB - akaari(crit) akaari(crit) 5strike - 6strike - 7evis - 8strike

waxen hamlet
#

one thing to keep in mind is CB is bugged with evis and BP as well

#

a CBed evis only gets the crit buff for the phys portion

#

BP is similar, the extra shadow damage has a slight delay so CB doesn't work correctly

stray bolt
#

It seems working fine on live for BP at least

prime narwhal
stray bolt
regal agate
#

if you want to sim with the exsang nerf, you can use this:

override.spell_data=spell.200806.cooldown=180000
override.spell_data=effect.295690.base_value=80```Example:<https://www.raidbots.com/simbot/report/mAtEQqNsRGFc1KN6Sjkgex>
image.png
stray bolt
dense scaffold
#

hello :3 i was trying to test an apl thing using the outlaw profile from the dungeon slice talent sims pinned in #outlaw but my results were about 3400 dps lower than what I thought should be the same sim, so i'm curious what I did wrong, or what to do differently. Relatively new to this and not a coding expert but would like to learn more sim stuff, any help would be appreciated hearthuggies
original profile: https://www.raidbots.com/simbot/report/6W9N5nuLXgzGoxF5ueKrh4
my sim: https://www.raidbots.com/simbot/report/73s4E7ccs1FLBUubrDdhSo

Raidbots
Top Gear - Fuu - 116,669 DPS - Raidbots
Raidbots
Advanced Sim - Fuu - 113,316 DPS - Raidbots
dense scaffold
#

also i guess a valid question is does it matter, if it gives me the same answer about my apl

dense scaffold
#

also also, i changed the line

actions.cds+=/blade_rush,if=variable.blade_flurry_sync&!buff.dreadblades.up&!buff.shadow_dance.up&energy.base_time_to_max>4&target.time_to_die>4
to
actions.cds+=/blade_rush,if=energy.base_time_to_max>4&target.time_to_die>4

and it appears to be a 0.5% gain in dungeon slice and 0.2% in patchwerk with the HO dance build, presuming i'm doing this correctly
https://www.raidbots.com/simbot/report/7GuATBvUdbsQXJppPvAVSH

Raidbots
Advanced Sim - Fuu - 113,863 DPS - Raidbots
half karma
half karma
#

I do believe that's where the gain comes from rather than pressing in dance

#

Also, still on phone so very hard to see, but you don't have to copy the whole apl everytime you make a profile with a change

#

You can just copy the "action group"

#

So here you are just changing the action cd you would just need to copy only that on every profile

dense scaffold
#

i took the bf sync out of the original, still seems ok to use during dance. i would think it'd be optimal to BR right before dance, but not sure what rule/condition would be best to test that, or if it'd even be practical if you're not specifically holding BR and SD condition is pretty random/abrupt

oop have to edit, I guess i messed up copy pasting something and put the bf sync not during dance condition at the very bottom of the CDs list on both "no bf sync" copies. seemed to make 0 difference in the results tho when i fixed it and ran it a couple more times

loud temple
#

Generally speaking have we identified the damage reduction of mind numbing vs atrophic

noble pendant
# loud temple Generally speaking have we identified the damage reduction of mind numbing vs at...

that's really not something that can be averaged. Numbing is mostly used as an attack speed slow for tank mitigation, while atrophic is a raid-wide damage reduction. The fights damage profile varies way too wildly to come to any sort of concrete numeric conclusion. The consensus right now is raids -> atrophic, and mythic+ -> numbing. Mythic+ usually involves more tank damage than group damage, numbing affects less players than it would in a raid, etc etc

loud temple
#

Thank you very much for the information!

calm spindle
#

do we have any current trinket sims yet?

noble pendant
calm spindle
#

but only the top few, right? I'd be interested in others aswell - is there a bigger list somewhere?

noble pendant
#

the simulation is a comparison of all

image.png
calm spindle
#

gotcha, thanks

orchid rune
#

Any sims for?

image.png
hearty knot
#

I think solo said that non of the effects are implemented in simc

finite sparrow
#

The mastery is now

maiden cargo
noble pendant
tough hornet
#

note trinket tuning will likely occur, since some are way out of line and some are useless. also not every trinket is implemented in simc and can be compared yet

knotty oriole
#

Manic does a lot of damage

vivid wraith
#

Good afternoon - has there been any Sims on the slimy expulsion boots as of yet, they are looking much better than all other options for feet

stray bolt
#

Seems we might want to reconsider CB in sub aoe. Or revisit CB use with a tornado.
When we use CB after secT in a tornado, it crits the following auto SS before the akaari dmg. Thus, the following skill or the akaari cannot benefit from CB.
Even with my suggestion, delay secT for a small window to cast CB after the cast of auto SS, it does not do good in m+ due to SS has trajectory and can be self-delayed. Unless the pack is somehow extremely squeezed together, like dummies in town.

west prism
#

You just delay sectech to the last nado spin

#

Instead of trying to pull it off mid tornado

stray bolt
golden briar
#

there should be no diff in CDR

#

only on first one

stray bolt
#

4x7x0.7+t+tx7x0.7/2=t+tx7/2 t=18.7
4x7x0.7+t+tx7x0.7/2=60 t=11.7
correct, there should be no cd difference at 2nd finisher in the following dance

median kite
#

Asking here, wowhead is recommending double sophic enchants...are buff enchants stacking or is the rppm low enough that two chances to proc it are worth it?

noble pendant
#

at least that's my understanding, I don't have them to actually test

#

they seem to all have a 1rppm

west prism
#

The agility stacks, 1 is 800, 2 is 1600 from the proc

noble pendant
#

holy

west prism
#

(At least it did until last day of beta)

#

Havent had it on live

noble pendant
#

it didn't used to do that with the bfa and sl enchants right

west prism
#

Nah

noble pendant
#

super weird

west prism
#

Just some df thing

unkempt sail
#

I’m pretty sure even burning crusade enchants stacked

noble pendant
unkempt sail
#

Yea but when was the last time we had a simple stat proc

prime narwhal
#

bfa

tiny narwhal
#

Was it?

#

SL literally had celestial guidance

regal agate
#

Should include all Crafted Gear, you can just paste it below your /simc in top gear.
Raidbots also has Dropoptimizer, what does support crafted gear.

📎 crafted_gear.txt
fluid drum
#

You probably meant to post this in #assassination. This channel is for sim research and optimization.

viscid oracle
regal agate
#

what do you mean with rep?

viscid oracle
#

items list for all the grindable rep items @regal agate

#

So wrath neck/back, cobalt ring, etc

dapper carbon
viscid oracle
#

okay and?

regal agate
#

i don't have a list of all possible worldquest/rep/etc items no

#

you can filter it on wowhead if you rly want a complete list

regal agate
#

The balance changes, if you want to sim with them. you can paste them in advanced sim and your /simc below.
Note: Subtlety Cold blood change is not included as that needs change in the simc implementation.```yaml
#subtlety: bp nerf - ~2% nerf on 5 target aoe
override.spell_data=effect.805458.ap_coefficient=0.08778
override.spell_data=effect.805486.ap_coefficient=0.08778

#outlaw: dispatch buff & pistol shot nerf - No noticable change
override.spell_data=effect.622.ap_coefficient=0.29925
override.spell_data=effect.269972.ap_coefficient=0.376

#assassiantion: poison bomp nerfs - ~1.5% nerf for builds using Poison Bomb
override.spell_data=effect.477503.base_value=80

#

If you want to compare current to tuned value, paste this below your /simc in advanced sim:

#Balance Tuning
profileset."balance changes"+=override.spell_data=effect.805458.ap_coefficient=0.08778
profileset."balance changes"+=override.spell_data=effect.805486.ap_coefficient=0.08778
profileset."balance changes"+=override.spell_data=effect.622.ap_coefficient=0.29925
profileset."balance changes"+=override.spell_data=effect.269972.ap_coefficient=0.376
profileset."balance changes"+=override.spell_data=effect.477503.base_value=80```
unique zephyr
#

@regal agate do you have a consumables string i could borrow?

regal agate
unique zephyr
stray bolt
#

is the signet still bugged? or only AA damage in S1? how is it simmed now?

finite sparrow
#

it was silently fixed back to AA damage recently, and this is in simc

maiden cargo
#

Hey, was looking at last update of rogue outlaw on simc and was wondering what 'rtb_buffs.normal=0' and 'rtb_buffs.longer>=1' meant

finite sparrow
maiden cargo
#

thx gonna look at it

willow knoll
#

in what way

stray bolt
#

Two items need some help/attention if possible:

  1. Outlaw HO+CTO build needs some attention on dance condition, cast on cd seems better and a gain, currently it runs at 74s interval where there are 1 or 2 missing casts
  2. HO builds should have a different reroll logic considering SnC buff way too strong and TB/crit quite weak. Current reroll logic makes the build sitting with TB in single buff sounds like a dps loss (and possibly some weak 2buff combos).
    We have conducted some tests and compared results from raidbot. Outlaw's sim seems weak possibly due to these two points. NS bug influence has already been considered.
    Q. There are conditions checking if subt buff for ghostly and other skills, does this also check dance?
half karma
#

1.https://www.raidbots.com/simbot/report/eirY1PKpFfYER24D9dCY36 did this quick might be some weirdness, but yea we tested and it was better to hold dance (the nightstalker bug isnt in simc as It should get fixed for raid, if it isnt then unluck)
2. rtb matter way less than you think, we mainly want to press it as less as possible with ho this was an old sim so I dont have a link anymore but you can see how it barely have any impact

Q. tried, didnt matter

image.png
Raidbots
Advanced Sim - Fuu - 110,835 DPS - Raidbots
#

you guys seems to always have ideas and suggestions which is great, but doing apl modifications is honestly not that hard, if you think something could be a gain you could try testing it yourselves, I could help if you don't know were to start, more people knowing how to modify apl is always good

stray bolt
# half karma you guys seems to always have ideas and suggestions which is great, but doing ap...

Error: Actor 'fuu': Action 'rtb_reroll_kir_cto': Cannot parse expression from '(rtb_buffs.normal=0&rtb_buffs.longer>=1)&!(buff.broadside.up&buff.true_bearing.up&buff.skull_and_crossbones.up)&!(buff.broadside.remains>39|buff.true_bearing.remains>39|buff.ruthless_precision.remains>39|buff.skull_and_crossbones.remains>39)': No expression found.
Hi, trying to do it from the profile from your link. can you run it from your end? it seems a corrupted profile? at least something is not right

finite sparrow
#

use nightly here

image.png
stray bolt
#

got it working finally, thanks

stray bolt
finite sparrow
#

the reroll rule in the APL is kind of in limbo while we test KIR stuff, you may want to isolate out the latest KIR reroll changes first

#

it’s possible ho and kir should use different logics but it’s looking like a lot of text and complexity for 0.3% gains, was hoping a simplistic catch all rule for all builds and sim styles would be what is in the APL by the end

willow knoll
#

i mean we can work on the complexity later

#

u can just sim both builds with different rules individually

#

and then in the future we can work on adding code to tell the difference automatically

#

since if what hes saying is true about the values being large, we can just keep the sims the way they are and then suggest to people otherwise

stray bolt
#

There is an increasing amount of voices from cn community talking about low sim numbers on HO ST build recently.
They claim about 2k-3k difference in dps, where sims number is much weaker.
One thing is the NS bug and dance abuses.
But I took a look into logs and sim reports.
It shows a big difference on rtb buff uptime considering CTO builds.
Current sim report shares a general rule on rtb reroll while SnC has a huge impact on HO builds compared to TB and BS, and this is where I start to dig.
However, there is another issue that sim report's overall rtb buff uptime is generally below logs.
For example, with my aggressive reroll rules, SnC's uptime is still only 46%, while logs show much better uptime around 55% to 65%.
Ofc my log samples does not cover as many iterations as sims, but this could be another direction.
Maybe it would be better to test more next week if NS bugs are fixed.

willow knoll
# 
For example, with my aggressive reroll rules, SnC's uptime is still only 46%, while logs show much better uptime around 55% to 65%.```
#

what are you exactly saying here?

stray bolt
#

potentially we're having better cto proc on live?

willow knoll
#

Are you saying that its much better to keep SNC (makes sense) and when actually rerolling for it you have much larger uptime?

willow knoll
stray bolt
#

yea, it is like from sims report, we have SnC sitting at 46% with more reroll rules but other buffs sitting at 35%ish. From logs I received there is a considerable more uptime for other buffs sitting at arount 43%.

#

I mean for me the concept is clear that HO builds should not have big gain anywhere in APL considering cast builders or finishers. That's why I was surprised at the first place when several players talking about big difference than sim'd numbers.

knotty oriole
#

That issue is worth about 3.5%

#

Regarding buff uptime, I don't really think there's any proof of anything like this without more logs. The differential in uptime will be wildly different just to pure RNG. You would need 100s of pulls of logs to get any sense of one buff being higher than another.

#

We also can't control for players themselves biasing uptime by their own execution of rerolls.

knotty oriole
#

Does look like NS bug has been killed in tonight's hotfixes

hearty olive
#

F

half karma
#

W

pastel flare
#

My hidden op assa build...

regal agate
#

@stray bolt mentioned that the bugfix to NS is not applying correctly on his server

stray bolt
knotty oriole
#

NS seems to be working correctly for me.

#

No 50% bonus any more from Stealth

stray bolt
#

yes, switch to sub and try some attack.

#

and also NS does not provide 8% for other skills on outlaw

#

im not sure if there was another hotfix in between

full stratus
#

Also just talking to a guildmate this bug was able to be replicated for a DK sim so it is potentially an issue with the potion and not rogue so to speak.

stray bolt
#

It also comes with an aggressive reroll conditions.

image.png
half karma
#

Could you detail a bit what the changes are ?

#

Or paste here the lines you changed, its a bit hard to find

stray bolt
#
  1. reroll condition
  2. PS at 5 or 4 cps
#

one more potential change is vanish condition, not for ambush but for surprise

half karma
#

Take em by surprise you mean?

stray bolt
#

actions+=/variable,name=rtb_reroll,value=rtb_buffs<2&((!talent.fan_the_hammer|!buff.skull_and_crossbones.up))|rtb_buffs=2&(!buff.skull_and_crossbones.up&buff.grand_melee.up|buff.ruthless_precision.up&buff.true_bearing.up&!buff.adrenaline_rush.up|(!talent.fan_the_hammer|!buff.skull_and_crossbones.up)&buff.loaded_dice.up)

half karma
#

I think we looked at this one and it didn't seem worth to play around the buff

stray bolt
#

actions.build+=/pistol_shot,if=talent.fan_the_hammer&buff.opportunity.up&combo_points.deficit>=2&!buff.dreadblades.up&(!talent.hidden_opportunity|!buff.subterfuge.up&!buff.shadow_dance.up)

half karma
#

Ok that does sound interesting

#

This is probably only worth for ho? Or have you tested for kir builds aswell?

#

(the pistol shot rule)

stray bolt
#

well, we believe current APL is way too conservative for HO.

#

pistol_shot,if=((talent.killing_spree.enabled&cooldown.killing_spree.up)|buff.adrenaline_rush.up|buff.buried_treasure.up|buff.skull_and_crossbones.up)&talent.fan_the_hammer&buff.opportunity.up&combo_points.deficit>=3&!buff.dreadblades.up&(!talent.hidden_opportunity|!buff.subterfuge.up&!buff.shadow_dance.up)

half karma
#

Yeah your apl changes clearly shows that

#

That's cool

stray bolt
half karma
#

Ok thanks will take a look, great work

stray bolt
#

So far, the reroll and PS changes seem solid gain; vanish part maybe not so accurate

#

I think the PS change is more like loosening the PS condition. I try to merge it with current "careful" conditions together.

#

I dont want to just set it to PS at 5, which seems too aggressive. However, it seems not achieving any better results yet.
I tried to set it from PS at 4 with buffs to PS at 5 with buffs, and it gets better.
maybe we just PS at 5?

#

The top3 PS at 5 or at 4 or at4 with buffs are basically the same dps, sidegraded

half karma
#

I mean the ps at 5 rules is simplier and seems to perform the same as the Complexe one

#

We probably can take a deeper look at it maybe

#

But we also try to keep the apl simple

stray bolt
#

yes, could be a good direction. i mean if it is just the same or PS at 5 is better, it suits me anyway.

#

although, it would be difficult to explain?

knotty oriole
last sundial
#

Bt is the energy buff right?

half karma
#

energy roll yea

knotty oriole
#

It’s 4 energy regen unaffected by modifiers. It’s a minuscule part of our regen.

last sundial
#

Only 4 energy?

knotty oriole
#

Sorry 5 now but the fact that it’s post modifier makes it very weak.

stray bolt
#

For single rtb buff SIM, BT is actually not that bad, on par with TB and RP.

#

So I think it could be a potential indicator for early PS.

#

If we PS early, basically we waste cps for full dispatch and more frequent ambush that is also a resource generator. Instead of storing opportunities, we just try greedy strategy to speed up ambush-PS cycles. If we fail to get opportunities from ambush, energy would provide good help.

stray bolt
half karma
#

yeah we have been discussing that aswell, idk if you tried on a normal ho build yet?

stray bolt
half karma
#

we also have been thinking that it might be related to gs since gs makes it more likely to have ps procs and higher cp, so Im asking if you tried this new rule with more classic HO builds ? (without gs )

stray bolt
#

Not yet.

#

But can be done soon

#

Let me try

finite sparrow
# 
override.spell_data=effect.269972.ap_coefficient=.376```
use this for the dispatch/pistol shot tuning, im not sure if they are in simc yet
knotty oriole
#

They aren’t I don’t think

half karma
#

this means they arent right ?

stray bolt
#

choosing sepsis

half karma
#

probably a gs only rule then yea

#

make sense

#

sepsis is not casted as much as gs so might be the same

stray bolt
#

indeed, gs is really a frequent castee

potent ferry
#

Has anyone tried delaying reroll a bit if no loaded dice and adrenaline rush is coming off cd soon?

half karma
#

yea, Idk if I can find it again tho

#

IIrc it didnt do anything

potent ferry
#

No need to look it up if it was already tested

half karma
#

yeah did that

image.png
#

didnt check if loaded was already ready tho

#

Im not sure if that would even happen

#

but could if you get lucky

#

but didnt seem promising

potent ferry
#

Yeah doesn't look promising

half karma
#

(this also would still press rtb if you have no buffs)

#

I assumed this case would have been bad to hold

potent ferry
#

Yeah I was thinking of those two things as extra conditions. No rtb buff up at all and loaded dice already up

potent ferry
#

here are some apl changes I've been working on
https://www.raidbots.com/simbot/report/gjgA8KMDPv3FdEAVUZZHb4
The tea build is a build without improved ambush (so full util)
The dance build is a build with cd and acro (but not all the other stuff)
The tea change (don't tea with opp proc) is probably too small to bother with)
The other two changes are:
don't reroll during subterfuge, seems good for both builds
don't vanish with any opp stacks, seems neutral for dance but good for tea

All in all, pretty insignificant stuff for HO dance, but 0.5% is decent for HO tea at least

image.png image.png
Raidbots
Advanced Sim - Fuu - 108,599 DPS - Raidbots
regal agate
#

also thanks for your contribution

potent ferry
# 
+ actions.cds+=/roll_the_bones,if=buff.dreadblades.down&buff.subterfuge.down&(rtb_buffs.total=0|variable.rtb_reroll|variable.rtb_reroll_kir_cto)
- actions.cds+=/roll_the_bones,if=buff.dreadblades.down&(rtb_buffs.total=0|variable.rtb_reroll|variable.rtb_reroll_kir_cto)
...
+ actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&(energy.base_deficit>=100&buff.opportunity.down|fight_remains<charges*6)
- actions.cds+=/thistle_tea,if=!buff.thistle_tea.up&(energy.base_deficit>=100|fight_remains<charges*6)
...
+ actions.stealth_cds+=/vanish,if=talent.hidden_opportunity&!buff.audacity.up&(variable.vanish_opportunity_condition|buff.opportunity.down)&variable.ambush_condition&variable.vanish_condition
- actions.stealth_cds+=/vanish,if=talent.hidden_opportunity&!buff.audacity.up&(variable.vanish_opportunity_condition|buff.opportunity.stack<buff.opportunity.max_stack)&variable.ambush_condition&variable.vanish_condition
#

something like that

#

that's easier to read maybe

regal agate
#

perfect

potent ferry
#

it seems that the subterfuge window is more important for HO tea than for HO dance, might be more improvements to be found there

knotty oriole
#

So one thing I will mention is that it is typically not a good idea to put many dramatically different profiles in the same sim as it can widen the error margin beyond what may be intutive

#

It's typically best to iterate on 1-4 profiles at a time with small differences as once there is a very large delta between top and bottom profiles in a sim you can lose fidelity in the other changes

#

As an example

#

If I extract the profiles out to their own sim

#

It's a bit larger in DS but can be different than the bulk result

#

Still worth investigating this but just a bit of a warning since I notice you have a lot of profiles in that sim

potent ferry
#

Yeah I did run this only for the tea build on the side to confirm that it is around 0.5% but wanted some context when sharing here

knotty oriole
#

Yeah that's all good, just wanted to make a note

potent ferry
#

Easy to make mistakes with a sim with that many profiles too, fully agree

knotty oriole
#

Always hard to tell when there are slightly larger DS gains vs. ST gains if it's not an AoE-specific change because that can just be the result of some timing shift but will look into it

potent ferry
#

It was only 0.2% for the dance build in DS too though

#

Only the tea build had 0.5%

#

the change to the vanish line is probably not good for the dance build, otherwise it would have been found already I think. I assume it is slightly negative if you run enough iterations

#

clearly positive for tea though

knotty oriole
#

One observation is that running it through ST the only real difference over the 5m fight is that the changed APL gets 1 more Ambush in

#

However the strange part there is that it has roughly 1 less Vanish

#

Due to the stricter conditions

#

It's unclear to me what alignment is changing to allow more Ambushes to fit into the existing Subterfuge windows specifically

#

Unless that is just the RtB change

potent ferry
#

if it is ST it is not due to BF I guess

#

otherwise that would be my first bet

knotty oriole
#

Probably the Subterfuge condition on RtB makes sense but would need to restrict it to HO builds most likely

potent ferry
#

yeah, that one is the clearest gain for both HO builds

#

I suspect we can find more subterfuge optimization, especially for HO tea

#

maybe throw up some extra lines for statistics for what we do with subterfuge up that isn't ambush

#

if we could make the DS sim BF, then stealth, then ambush it would probably be a gain. Not sure how to do that though, also pretty hard to execute in practice

knotty oriole
#

Yeah I have some ideas, but will have to investigate more tomorrow

last sundial
#

reroll broadside/ss when going into db

#

if they're at less than 15 seconds left

#

since having something like crit/true bearings would be better no?

noble pendant
grim inlet
#

does dungeon slice not apply WFT?

regal agate
regal agate
#

If you want to sim with all nerfs you can put this in the header section in EXPERT MODE.

📎 nerfs.txt
hasty parrot
#

Hi all, apologies if this is not the correct space to ask this, but does anybody know how "good" is the default APL in SimC? I am using the addon called Hekili to help me with my rotation and was wondering if there are more advanced APLs out there or the default one from SimC is good enough?

regal agate
hasty parrot
#

Ok thanks a lot fuu

regal agate
winged iris
#

Do we actually know how Resounding Clarity is nerfed ? Does it randomly choose 3 out of the 4 cp previously charged or is the 5th cp just not charged anymore ?

noble pendant
#

random

winged iris
#

So every single Sin opener is random now.....

#

KB/EXG openers will feel really bad depending on crits and cps charged

#

I don't know if it's possible to changed them in a way where we can always have that 7cp rupture before exg outside the opener

#

With ER + SBS we should always be able to hit the charged rupture on pull, worst case scenario is 0 cp -> ER crit -> 3 cp with 2-4-5 charged -> SBS bringing you to 4/5 is fine but after that SBS can randomly crit and bring you to 6..... I don't know if it's big deal, it's to early to tell, but i feel like nerfing by adding rng in this way is not the best choice.... Even if the spec needs the nerf

regal agate
#

raidbots had a update earlier, ptr=1 should work now too to sim the changes (nightly build)

knotty oriole
#

ER change will require a code change.

#

Not checked in yet.

regal agate
ruby agate
#

I'm seeing bronzed grip wrappings underperform compared to the toxic boots proc I also wear - is it possibly because procs are getting "wasted" by getting proc'ed by the healing effects I self-apply (slice and dice healing, crimson vial)? I see 66 healing procs over a dungeon and only 74 damage procs. Compared with 137 damage procs from the toxic boots. I'm wondering if the sims maybe aren't modelling this proc rate correctly and thus giving misleading embellishment comparison results. The comments in the code indicate the sim is undersimming by only proc'ing on auto attacks, but it may in fact be over-simming

regal agate
#

Bronze grip performs on average worse than boots, this is just a consequence of power level of the effects.
But the reason people go for weapons is that they are the highest damage upgrade item you can get early on, and if you already craft weapon you can just spend a bit more and put bronze grip/fang on it for a small extra gain.

#

additional, bronze is slightly more proc dependent, so how well it performs is also depending on your luck to a degree

knotty oriole
#

All yellow damage hits are presumed to proc the heal until we whitelist them

#

I did the testing for this the other day but isn't implemented yet

#

Since AAs always proc damage and are so frequent, you can more or less look at this as a 50/50 roll, or maybe a 60/40 damage biased roll due to some yellows triggering damage in-game

#

But sims should be fine for this

ruby agate
ruby agate
#

(and the heal proc is very not good, it's crazy undertuned)

knotty oriole
#

Every ability procs a specific one

#

Ambush procs heal

#

Finishers don’t. AAs don’t.

hasty parrot
#

Hey all, does anybody know if there is a way to compare WoW Logs and the current APL to see where I am going wrong in my rotation?

regal agate
#

there is no tool/script to do that. But overall a better way is to ask for a log analysis in the spec channels if you reach the point of not finding mistakes on your own.

hasty parrot
#

Ok thank you fuu!

midnight moat
#

Or do a log cmp, in wcl and compare your self to other better rogues

stray bolt
#

@knotty oriole @noble pendant For sin's dance, I'm checking a sim -- https://www.raidbots.com/simbot/report/bbRJ5sxEdtbS1T1LTGViwG

  1. dance has a condition that prevent its use to refresh imp'd gars. (1st cast is used at 35s, can be used at 11s ish to refresh?)
  2. vanish is also not used for anything like refreshing imp'd gars or use MA during dm. (1st cast of vanish is at 58s)
  3. significant cast missing for dance, but this is due to pandemic loss issue I think? (1st cast at 35s and 2nd cast after 2min with IC)
Raidbots
Quick Sim - 水忍者 - 95,742 DPS - Raidbots
knotty oriole
#

It's been on my list to revisit this a bit now that the expressions are updated to have the correct odd Shadow Dance duration returned for MA and Imp Garrote

#

(Since they don't overhang Shadow Dance duration by 3s like normal stealth)

#

Although I think most attempts at syncing things with Deathmark have not really been positive so far it may be worth revisiting that given the various tuning changes that have happened with Exsang, etc.

#

I guess probably Carnage could definitely see some Dance improvements since currently the Carnage syncing up is mostly oriented towards Vanish

#

(At a quick glance, not really seeing any gains here though messing with some of the Dance conditions for Carnage unfortunately.)

noble pendant
#

I also kinda haven't cared to optimize it much because dance is just so so far behind

#

maybe if you can find like 1.5% you could justify it on council specifically

#

but any hope for general use would require like 5+% worth of gains

#

and that's a tall order

knotty oriole
#

Changing Shadow Dance to have a Garrote refreshable condition seems to be neutral at a glance and my attempts to sync it with Carnage seem negative so hard to say so far. It's definitely a bit awkward to play around.

stray bolt
stray bolt
#

and only 1.3 casts of vanish

knotty oriole
#

I checked in some Shadow Dance optimizations after looking at some various things tonight. Some weren't specific to Shadow Dance though (funny enough, the Garrote change was mostly relevant for Vanish due to the timing edge cases and not really impacting Dance much)

#

Not sure where you are seeing that 5 minute count though

#

Even with the unchanged profile I'm seeing 3.5 Shadow Dances

#

With my modified profile it is 4.5 though

#

Still appears to be a rather far way off other builds though

stray bolt
#

in the link

knotty oriole
#

I was able to gain about 1.7% in DungeonSlice and 0.6% in ST.
Your particular build I ended up getting 1.8% in ST due to fixing an odd timing issue I noticed with Crimson Tempest and Kingsbane (in ST only, it was still about the same in DungeonSlice for that build.)

knotty oriole
#

Likely this is still quite a bit behind Tea and ER setups unfortunately though, but always good to improve things

naive glacier
#

Dance with KB sync, ambush and envenom som for duration?

stray bolt
# knotty oriole I was able to gain about 1.7% in DungeonSlice and 0.6% in ST. Your particular bu...

Could you please link me the sim so I can use the same profile and do some tweak?
Currently I'm thinking about:

  1. maybe we dont open with IC in some scenarios (like 5 targets), and we put a normal opener dotting 4 targets and IC imp'd gar from dance a bit later. This gives us 5 gcd with crit to do some other stuff instead of 5 gars
  2. 8 normal gars without pandemic=4 imp' gar+2 normal gar, if this does not last long enough for next dance, it becomes 8 normal gars +6 imp'd gars with 2 pandemic. if we delay IC, it could be 8 normal gars + 6 imp'd gars with better pandemic. Or we delay IC further for a sync, and we get 8 imp'd gar with maximum 7 full pandemic. This will cause miss cast of IC, so maybe it should be limited by combat time.
deft widget
#

I'm a total TC/numbers noob, but is there any chance I could estimate how much my Jeweled Signet of Melandrus dealt (contributed to dmg) during my whole dung run? Can I somehow estimate it based on amount of melee casts and dmg they dealt? like I had 2278 melee casts that dealt 6,10m dmg, is there any chance I can count the dmg that the ring provided? EDIT: + does it proc on Main Gauche as well?

finite sparrow
#

i'm not sure how it interacts with the damage formula, but i think this will work in raidbots advanced sim, so you can see the difference:


profileset."disable ring effect"+=override.spell_data=effect.343994.coefficient=0```
#

does it proc on Main Gauche as well

no, it just boosts melee damage, does not boost main gauches that proc from melees

hasty parrot
#

Does this APL for roll the bones looks alright to you guys?

#

rtb_buffs < 2 & ( ! buff.broadside.up & ( ! talent.fan_the_hammer.enabled | ! buff.skull_and_crossbones.up ) & ! buff.true_bearing.up | buff.loaded_dice.up ) | rtb_buffs = 2 & ( buff.buried_treasure.up & buff.grand_melee.up | ! buff.broadside.up & ! buff.true_bearing.up & buff.loaded_dice.up )

spring trench
#

Can anyone advise on how to edit my current gear to ilvl 418?

#

I tried copying the line in my simc but I don't know how to change the ilvl

regal agate
#

scale_to_itemlevel=418

spring trench
#

Oooh where do I put that in?

#

Under my Simc?

#

e.g # Primal Molten Shortblade (405)

off_hand=,id=190505,enchant_id=6643,bonus_id=8836/8840/8902/8802/8845/8795/8937/8960,crafted_stats=40/49

#

Primal Molten Shortblade (405)

off_hand=,id=190505,enchant_id=6643,bonus_id=8836/8840/8902/8802/8845/8795/8937/8960,crafted_stats=40/49, scale_to_itemlevel=418

#

Like this?

regal agate
#

just put it in advanced and your /simc below

spring trench
#

ok

#

oh but I just want 1 particular item

#

not my whole body

spiral light
spring trench
#

roger, will give that a shot

spiral light
#

So for the weapon it would be # off_hand=,id=190505,enchant_id=6643,bonus_id=8836/8840/8902/8802/8845/8795/8937/8960,crafted_stats=40/49,ilevel=418

spring trench
#

thank you!

#

last question, how do I swap the embellishment for sim comparison?

spiral light
#

You can do that in Top Gear.