tc-subtlety

[mental dirge]

The 67 mil dps sim is under lock

[golden rain]

picture this workflow:
you have local testing with your local simc build for quick iteration to dev on apl changes
you think you found an apl increase, you push it to github for PR review. At this stage you could also trigger a snapshot simc in a docker container if you want everyone to have access to the same changes locally
your PR is reviewed by fuu, he agrees its a increase, merge it in
nightly raidbots build comes in and picks up apl changes every day at midnight, now everyone can run it

[fickle shore]

be careful or I start leaking your guys 300k sim

[mental dirge]

Thats it?

[golden rain]

at each stage latest apl is available, any ongoing changes are neatly threaded into PRs, you dont have to wait for people to be online to dev

raidbots auto picks up all changes

[remote sparrow]

not everony here knows git or knows what docker is.

[rugged solar]

So to summerize the discussion, and not loop:

• I completely agree that a pipeline to sync changes faster would be good, i lack the time right now to do so. But feel free to make a pull request for the latest change if you have the time for it.
• Transparency is not a problem, everyone can join and view this channel and the latest changes are well documented
• Communication is easier on Discord because thats where the majority of players are.

[golden rain]

i do think for sub the pin system is fine and also works

only the pipeline stage is lacking

[fickle shore]

ok thats not how it works, people in here usually only have a very focuses vision, they will go "oh I think with this build I could be pressing ss after vanish", implement it, send the sim here, it probably wont work with all builds or in aoe, but the idea is there, fuu can iterate on it, find and solve the issues,

it also is jsut faster

[golden rain]

but for assa or outlaw like i have no idea how to get apl

besides dm whispyr or whoever

[fickle shore]

I mean

open the pins maybe

[remote sparrow]

well he meant u push it, but fuu essentially tests it before merging it

[rugged solar]

if there is a pin for sims

moment

[remote sparrow]

which is essentially what we do here

just more formal

and version controled

[fickle shore]

which is a way more annoying process

cant just talk about it directly

cant just itterate over it on live

[remote sparrow]

yeah, im pretty sure we would all come back here anyways

[golden rain]

you can have automated tests that run in CI and runs your apl through a 5 min sim, 2 minute sim, aoe sim, dslice sim

whenever you push a change

[fickle shore]

its also prolly 2 people (fuu and the guy proposing is idea colaborating)

[remote sparrow]

a big part of the fun is talking to each other and saying hey i notice this thing what do you guys think

[fickle shore]

I mean you are free to setup that lol

have one that test all the relevant builds aswell

[golden rain]

ye exactly

a good CI + version control system adds a lot

takes some initial setup effort but ye just trying to make a point on how it could be good

[remote sparrow]

yeah I dont actually know how to set that up with simc. maybe you can install it locally and just have a script that runs all the configurations

[fickle shore]

you are still ignoring all this iterating process that people do over in here

[remote sparrow]

but idk how to run it any other way

[rugged solar]

you can typically get the apl from linked sims too:

[fickle shore]

and you are hugely increasing the barrier of entry

[rugged solar]

ignore the tabs, i was just ehm testing some apl stuff

[ivory kestrel]

yea that is what i do for some things

[golden rain]

iterating process is handled on PRs via commits from you or fuu or comments on PR review

[ivory kestrel]

there is a part for the botton where you can get it

[golden rain]

each commit auto tests vs relevant build and aoe vs/st settings

[fickle shore]

I mean if you dont understtand how that is vastly different from just tlaking in discord

idk

[rugged solar]

i also added a easteregg to the video ^^

[remote sparrow]

I do actually think that its a good idea, will save a ton of time for fuu too. like any change you make, you can just have the scirpt test all the different configurations for a set list of talent setups you predefine

[golden rain]

ye

you can have a github notifications bot that xposts pr openings or comments to channel here

[rugged solar]

yes, i thought about that too

making a test setup

[fickle shore]

I mean a script that run a bunch of builds and different fight style is good

[rugged solar]

that i can easy run via cli

[fickle shore]

there is litteraly no reason to operate thrugh github tho

[remote sparrow]

githubs like w/e I only like it for version control

I do think that having essentially something there so we can always find previous versions is good

[golden rain]

if you have a testing script everyone needs to have simc installed locally, docker containers, a beefy pc, etc

[remote sparrow]

for example, an issue we had right now, was that for 10.1 we introduced a lot of changes

that became obsolete with tier gone

and fuu had to do clean up

[golden rain]

github allows you to run the testing scripts on beefy test servers so when random joe makes a 1 line change to simc apl in the UI they dont need to know how to run the test scripts

they are just automated jobs on ur pr

[remote sparrow]

but if we had version control we could just get the base APL

[fickle shore]

no, only interested people would have the script (aka fuu)

[fickle shore]

but random joe need to setup github and whatever the fuck

[remote sparrow]

idk how to do this with simc. do you?

[golden rain]

you dont you just go to github and do a ui change

[fickle shore]

when they barely know how to change an apl line

[remote sparrow]

can only write a script to test it locally

[golden rain]

if you can run it locally it can run on CI

usually actually the same script runs both locally and in CI

[remote sparrow]

yes but youd run it locally again, so u cant use their servers

[rugged solar]

but to talk a bit bigger picture

i am also aware that tooling can be improved

i just priotize diffrent things atm

[golden rain]

yeah just making some ideas here

[rugged solar]

about automating testing

it is a nice qol

but the main time is in the waiting time to finish a sim not to set it up

[ivory kestrel]

my face when sim gets to 75%

FATAL ERROR

[rugged solar]

so even if you can compress the sim setup

what is nice

[ivory kestrel]

oh whoops wrote rotten instead of the_rotten once again

[rugged solar]

you still need the same time overall

other specs did take diffrent priorities

and i think thats fine

for what its worth, what i and the old tcs always tried to do was to be as transparent/open as possible

[golden rain]

https://github.com/simulationcraft/simc/blob/dragonflight/profiles/Tier31/T31_Death_Knight_Unholy.simc for example profile could be pinned i just edit like this @fickle shore

same thing as here basically

and then the automated tests run on their change before fuu even looks at it

[fickle shore]

no not same thing you are ignoring the whole having a conversation thing lol

[remote sparrow]

but also as zac mentioned, it does increase barrier to entry

for people who are not comfortable with git

[fickle shore]

also just so you know

you mentionned outlaw

[golden rain]

yeah i havent opened outlaw channel yet

[ivory kestrel]

outlaw you can do the same as sub do a sim and just get the APL

[golden rain]

[fickle shore]

where we litteraly have a link to a github where solo have the latest profile stored

[ivory kestrel]

from the lower part

[fickle shore]

ye I know

[golden rain]

noice

[fickle shore]

but we dont work on github there is no point

its just to have an easy fall back

altough 90% of the time you just open the latest sim sent in the channel

and go from there lol

sin is a bit differnt, whisp likes working how he does so it might not be of your liking

its understandable

[remote sparrow]

yeah,i think the new tools that are beneficial are :

1. version control
2. automatic testing tool

[golden rain]

i guess, i just dont know if any work has been done since latest sim or not as a third party

like you could be mid way through testing a vanish condition and on github i can see your pr and peek at it

maybe take that change and go in a different direction

[fickle shore]

• people that are interested in tcing follow the channel so they know whats happening
• no one does prs

[rugged solar]

reformed

[ivory kestrel]

so you want to see things like this?

[rugged solar]

just one thing

[ivory kestrel]

where you can peek in things people are doing

[rugged solar]

if i do a change

[golden rain]

ive never had anyone call me reformed before

[rugged solar]

you still don't see my browser

nor know what i work on

[fickle shore]

ok but thats on you fuu

[ivory kestrel]

too late we already got the spyware in

[fickle shore]

why arent you streaming it ??

[ivory kestrel]

we know what happens in those 120 windows

[rugged solar]

i could but without microphone

because my enviroment is loud usually

[remote sparrow]

its fine

play the ne worlds music int he background

[rugged solar]

and streaming without mic is fairly uninteressting

[remote sparrow]

or old worlds music

or kda

banger

[rugged solar]

but besides

the point is, if i put in ideas in the pins to test

or the latets apl

there can always be a conflict of 2 people working on the same change

github, discord or whatever don't change that

the 2 things that a github sync improves is visibility and the chance of external input

[fickle shore]

ye idk if external input exists really

[rugged solar]

visibility in that people who typically don't look at the apl see the change

and tools that use simc as a backend have faster access to changes

[fickle shore]

whats the probability someone interested isnt in the discord

[rugged solar]

exernal input is questionable

like stated

one contribution in 7 or 8 years

[ivory kestrel]

tbh some people that don't speak english do dm me about things

[rugged solar]

not rly that promising

[ivory kestrel]

but only in that case i guess

idk if any french/german dm other people that are around here

[rugged solar]

also, blizzcon now

[ivory kestrel]

tru

[fickle shore]

ye

never for tcing tho

[rugged solar]

i have some ppl who dm me in german

and anyone can if he wants

but thanks for the conversation

tooling is def. something i want to improve on

[ivory kestrel]

i don't usually use it that much so idk

[remote sparrow]

I really do think that soviet is correct with 2 things, 1) an automated testing, even if its just locally for fuu. saves a ton of time. 2) version control

so we dont have to re clean the APL after we made a lot of weird changes to the core APL because of 10.1 tier

or cases like that

[fickle shore]

it can save times ye, its easy enough to have a big file with like different builds and sim that in different scenarios

i have done it in the past

can end up annoying very fast tho

usually you can get away with only simming the few main builds

(2)

[ivory kestrel]

i mean someone could do an access file and hook it to excel with some scripts

but not me

[remote sparrow]

yeah, my thought was to jsut make a file with all the builds you want and like splits you want : [dslice, single target, different target count], and different time lenghts

and it just loops through the configs and runs them all hopefully parallel

if I can install simc on a linux I can actually right a script or this

or a simple .sh file that just runs everything

[rugged solar]

wdym

thats not how cleaning up the apl works

you don't just roll back changes

you nee to test again

[fickle shore]

I think thats what he meant

like make changes > check with different builds / fight style

[remote sparrow]

yes, thats what i meant. poor wording on my end i guess.

[fickle shore]

I did that manually in the past

a little script could help but also

eh

[rugged solar]

ah fair

the host is very talk active

[remote sparrow]

its my first blizzocon! im very very excited

[coarse laurel]

I’m not sure I would get too hung up on the workflow for PTR APL work just based on this type of PTR. Rogues haven’t had a rework of this magnitude since 2016. Most work is far more incremental.

I don’t think anyone’s workflow is optimized for a once in a decade event. 😄

[remote sparrow]

nah, surely subs gonna get some actual talents in 11.0. surely

[rugged solar]

@coarse laurel did you read/see the bugfixes i mentioned earlier?

[coarse laurel]

I scanned them but will need to catch up later tonight.

[rugged solar]

tl;dr:

• DB now works on spells it converted
• Shadowed Finishers seems fixed
  -- Works with damage amps now, damage calculation is now using damage pre armor calculation it seems

[past fable]

Also: Replicating shadows doesnt benefit from talent amps if you have dark brew, behaving the same way the tier rupture proc does. (still doesnt benefit from finality or nightstalker unfortunately)

[magic helm]

Do we have new race sims for 10.2?

[rugged solar]

you can use the race template from #subtlety-faq if you want to sim races, but i will update my sheet soon

[coarse laurel]

As far as I can tell, Shadowed Finishers was just changed back to how it used to work. I don't think it has anything to do with the damage dealt or armor or anything. It just has its own coefficient again.

Unless you guys are seeing otherwise

Effects          :
#1 (id=805486)   : School Damage (2): shadow
                   Base Value: 0 | Scaled Value: 0 (coefficient=0, delta coefficient=0.05) | AP Coefficient: 0.08780 | PvP Coefficient: 1.00000 | Radius: 0 - 10 yards | Target: At Caster (22) -> AOE enemy (15)
Hotfixed         : Delta (0 -> 0.05), AP Coefficient (0 -> 0.0878)
Description      : $@spelldesc319175

Effects          :
#1 (id=819933)   : School Damage (2): shadow
                   Base Value: 0 | Scaled Value: 0 (coefficient=0, delta coefficient=0.05) | AP Coefficient: 0.22346 | PvP Coefficient: 0.00000 | Target: Enemy (6)
Description      : $@spelldesc196819

And were added back into the Mastery whitelists, etc.. I assume the CP multiplier was also fixed

Probably just works exactly how it used to

[remote sparrow]

seems like they just reverted it to how it was yep. and the CP multiplier also seems to have been fixed

[remote sparrow]

shadowed rupture went from doing 1.5% to 5% but still seems too weak in the sims. Realistically shuld be close to 9-10%

^ old

^new

sims are a bit higher as a result of this change

u could essentially double the evis bonus from shadowed finishers (probably an over buff), to get an estimate for the spec with bis build around 238k

[past fable]

Yep

As long as that own coefficient means it ends up doing roughly ~48% of the physical effect (with all the amps) thenwere good

[rugged solar]

ah perfect, do numbers match?

does dd/vt work in the sim for spells that get converted?

[past fable]

no

[rugged solar]

alright, so that needs to be changed then

[past fable]

And like armin said, shadowed finishers still doing roughly half of what its supposed to

[rugged solar]

could be explained with the armor calculation we discussed yesterday

[coarse laurel]

I didn’t make any changes yet. Didn’t have time last night. Will revert back to the old setup today.

[rugged solar]

👍

[rugged solar]

Highlight:
https://www.raidbots.com/simbot/report/ifK7htTiLggx4vA7fPb1yK

[coarse laurel]

Ok, checked in the Shadowed Finishers changes

What was confirmed as actually happening with the Dark Brew stuff?

[past fable]

It works

Like rupture & sepsis gets increased by deeper etc

#1065728795455266888 message

Only replicating is a bit different, similar to how the 2p rupture works

[rugged solar]

Flagellation improvements:
Based on [Eleems Apl modification](#1065728795455266888 message).

What does this change do?

• Syncs Flagellation with symbols of death if possible.

Lines Changed:

actions.cds+=/flagellation,target_if=max:target.time_to_die,if=variable.snd_condition&combo_points>=5&target.time_to_die>10&(variable.trinket_conditions&cooldown.shadow_blades.remains<=3|fight_remains<=28|cooldown.shadow_blades.remains>=14&talent.invigorating_shadowdust&talent.shadow_dance)&(!talent.invigorating_shadowdust|talent.sepsis|!talent.shadow_dance|talent.invigorating_shadowdust.rank=2&spell_targets.shuriken_storm>=2|cooldown.symbols_of_death.remains<=3|buff.symbols_of_death.remains>3)


New APL:

[coarse laurel]

Not sure what you mean by this

[past fable]

Replicating doesnt double dip

[coarse laurel]

Wasn't Replicating just changed to be non-residual anyway? So it has its own modifiers anyway

I think Replicating was also basically just reverted to its old behavior

[past fable]

Well ye, but it currently double dips in sims

From dark brew

Since it increases both rupture and then also increases replicating

[coarse laurel]

It shouldn't. RS damage isn't based on Rupture damage at all in sims

Anymore

[past fable]

Not at my computer rn but replicating got more like 20% more damage than 10% when i checked

[coarse laurel]

I think it actually does double dip looking at the spell data

When the new labels were added

RS Rupture effect was already in the Dark Brew modifiers

The new effect was this added a bit ago

#5 (id=1083911)  : Apply Aura (6) | Apply Percent Modifier w/ Label (218): Spell Periodic Amount (22)
                   Base Value: 10 | Scaled Value: 10 | PvP Coefficient: 1.00000 | Misc Value: 22 | Misc Value 2: 2696 (Label) | Target: Self (1)
                   Affected Spells (Label): Secret Technique (282449), Black Powder (319190), Eviscerate (328082), Lingering Shadow (386081), Rupture (394031)

But

#3 (id=1083909)  : Apply Aura (6) | Add Percent Modifier (108): Spell Periodic Amount (22)
                   Base Value: 10 | Scaled Value: 10 | PvP Coefficient: 1.00000 | Misc Value: 22 | Target: Self (1)
                   Affected Spells: Rupture (1943), Deadly Poison (2818), Crippling Poison (3409), Numbing Poison (5760), Wound Poison (8680), Crippling Poison (115196), Shadow Blades (121471), Gloomblade (200758), Double Dose (273009), Shadow Blades (279043), Nothing Personal (286581), Instant Poison (315585), Flagellation (323654), Sepsis (328305), Sepsis (328306), Flagellation (345316), Numbing Poison (359078), Rupture (360826), Sepsis (375936), Amplifying Poison (383414), Flagellation (384631), Sepsis (385408), Atrophic Poison (392388), Sepsis (394026), Rupture (394031), Deadly Poison (394324), Instant Poison (394326), Wound Poison (394327), Amplifying Poison (394328), Flagellation (394757), Envenomous Explosion (426581)
                   Family Flags: 15, 20, 51, 78, 97, 109, 118, 124, 126

This is the original modifier list in Dark Brew

It also contains 394031

So I think this is a bug, but probably one that happens in-game

Since these are just passive

[past fable]

Youre correct, i just forgot replicating doesnt work with stratagems

So they roughly even out

Mb

[coarse laurel]

This 2696 label is what is used in Deeper Daggers as well, and it's correct there because the main DD direct list is far more limited

But the custom Dark Brew list they have is a lot more comprehensive

So regarding Dark Brew then, I'm guessing this must be scripted since I don't really see any aura references in either Veiltouched or Deeper Daggers regarding, say, the basic Rupture spell ID

Kinda a pain in the ass... ><

[rugged solar]

i assume so

[coarse laurel]

I'm just gonna have to go on faith here that everything works that is supposed to... lol

[rugged solar]

it was only few spells he needed to add DD to and only rupture for VT

[coarse laurel]

https://github.com/simulationcraft/simc/commit/290e64dbb812981080096120094705614e6b72d2

Alright, that should do the trick. Kinda taking it on faith there aren't any double dips or everything captured, but should be functional 😛

You guys will probably want to keep an eye on it

[ivory kestrel]

will look at it tomorrow more but only dipper that was kinda hard to see was replicating i saw you talking to eleem about it

[rugged solar]

ah great

thanks

guess need to re-sim dark brew builds ^^

[past fable]

Kinda inconsequential and irrelevant but heres some very slight improvements to some of the races, to have em line up their cds with flag instead of symbols (flag was better than blades)

actions.cds+=/berserking,if=buff.flagellation_buff.up|fight_remains<20
actions.cds+=/fireblood,if=buff.flagellation_buff.up|fight_remains<20
actions.cds+=/ancestral_call,if=buff.flagellation_buff.up|fight_remains<20```
<https://mimiron.raidbots.com/simbot/report/w7dx2xkiCF9zzuiciVfaw8>

[turbid pulsar]

only tested the normal 2min build for this and all i did was remove the !buff.premeditation.up on backstab

[turbid pulsar]

ran some more tests on it, seem like its only a gain for non shadowdust since it looks like it ends up delaying cds for some reason

[golden rain]

We take the non dust gains

[rugged solar]

thats interessting

because the !buff premediation was a gain for both gb and premediation when i introduced it

but will double check

[rugged solar]

probably better to use shadow blades

[mystic sequoia]

Erhm, from checking quickly what you linked on mobile, you removed the usage fully if you don't talent flag, no?

Minus just at the end of fight

[rugged solar]

it is probably better to use shadow blades as a driver

because it is soething you always use

else just

|buff.symbols might be enough

[mystic sequoia]

In practice, I think the current implementation is basically OR symbols

[rugged solar]

can add a talent check

buff.flagellation_buff.up|!talent.flagellation&buff.symbols_of_death.up|fight_remains<20

also given how much we re-use it

might be worth to add a variable

[past fable]

Ye mb i guess thats technically true

[rugged solar]

you realistically won't skip on flag

still good to have a fallback

I quickly checked, shb and flag sync is equal so i did put in this:

actions.cds+=/variable,name=racial_sync,value=buff.shadow_blades.up|!talent.shadow_blades&buff.symbols_of_death.up|fight_remains<20
actions.cds+=/blood_fury,if=variable.racial_sync
actions.cds+=/berserking,if=variable.racial_sync
actions.cds+=/fireblood,if=variable.racial_sync
actions.cds+=/ancestral_call,if=variable.racial_sync```

the diffrence is very minimal tho, 0.1% mostly

[rugged solar]

^ just the latest apl with the above lines:

[rugged solar]

Just a little bit of fun fact

only ~1/3 of the apl remains of the 10.1.x version

if we include comments, it goes down even lower

how that looks in chanegs:

[ivory kestrel]

and more things we need to change ayyy

[rugged solar]

oh possible

was more of a fun comparison

there are def. some changes to be made still, tho i would say the apl is quite good/stable

it is vital to keep in mind the amount of change that the patch had, you can also just scroll the channel and will see how long we iterated over chanegs already

(3-4 weeks)

[low breach]

I guess this is more changes then some exp to exp changes

[rugged solar]

i could do the same for outlaw/assassination

i don't think it is as extreme for them, but will sitll be a lot of change

xpac to xpac changes have a lot of reoved changes

10.2 adds a lot more complexity to all 3 specs

[past fable]

Tornado was ending up as a negative dps talent on singletarget because the use of it in sims screwed up some other spell windows so i made a quick fix, im sure it could use more love but atleast i made it not be a negative dps talent
https://mimiron.raidbots.com/simbot/report/oMagPqrQTmz8mRtznrb3yP
Dslice and singletarget

PS: Its still not a good talent for singletarget, no need to worry

[rugged solar]

👍

[fading mica]

goddamn lmfao thats insane

[rugged solar]

Shuriken Tornado cleanup/improvements:
based on [Eleem's changes](#1065728795455266888 message)

What does this change do?

• Only use Shuriken tornado outside of cooldowns on low target counts.

Lines changed:```ansi
actions.cds+=/shuriken_tornado,if=variable.snd_condition&buff.symbols_of_death.up&combo_points<=2&!buff.premeditation.up&(!talent.flagellation|cooldown.flagellation.remains>20)&spell_targets.shuriken_storm>=3
actions.cds+=/shuriken_tornado,if=variable.snd_condition&!buff.shadow_dance.up&!buff.flagellation_buff.up&!buff.flagellation_persist.up&!buff.shadow_blades.up&spell_targets.shuriken_storm<=2&!raid_event.adds.up


__Lines removed:__```diff
- actions.cds+=/shadow_dance,use_off_gcd=1,if=!buff.shadow_dance.up&buff.shuriken_tornado.up&buff.shuriken_tornado.remains<=3.5
-     actions.cds+=/symbols_of_death,use_off_gcd=1,if=buff.shuriken_tornado.up&buff.shuriken_tornado.remains<=3.5&!set_bonus.tier30_2pc```

__New APL:__

Highlights:
Dungeon Slice

Single Target

Two Targets

[ivory kestrel]

so tornado is back pog

[rugged solar]

Seven Targets

The talent setup i used might get a lot higher numbers than others, so keep that in mind

so the typical win is likely lower

[rugged solar]

Rupture improvement
Idea: change the opener for smarter rupture use.

What does this change do?
Remove double rupture casts from Shadow Dance.

Apl Change:

actions.finish+=/rupture,if=buff.finality_rupture.up&buff.shadow_dance.up&spell_targets.shuriken_storm<=4&!action.rupture.used_for_danse

New Apl:

[rugged solar]

Highlights:
Single Target

single target

Single Target

Dungeon Slice

[azure drift]

how do you format this?

[fickle shore]

omega

[rugged solar]

moment

ill show you how its done simpler

You can use This Website to help you with formatting.

Example:```ansi
actions.stealth_cds+=/shadowmeld,if=energy>=40&energy.deficit>=10&!variable.shd_threshold&combo_points.deficit>4
actions.stealth_cds+=/variable,name=shd_combo_points,value=combo_points.deficit>=3

(you can use diffrent colors too, i just liked these ones for the contrast and readability)

[golden rain]

how is this opener change 1% damage?

[rugged solar]

it is in one instance, it is for most setups 0.2-0.6%

but what it does is, like i mentioned in the pin remove double rupture cast

i did also test alternatives, so this was e.g. contrasted to force rupture before dance

e.g. here is the change compared/combined with a 2nd one

the change does simply reduce the amount of rupture casts

towards

how do you format this
Just to mention it, in case anyone didn't know this.
If you hover over the text box you can copy/paste it with formatting by clicking this button.

[twilit hearth]

is it just me or does the dance is fired simply on CD there and sometimes it just overlaps with symbols or other cds?

like there for example

[low breach]

In playing yes this is normal. Especially with current tier.

[twilit hearth]

but we still want to hold one dance charge for symbols or other cd's, right?

[low breach]

Except lining up flag and blades with your 2nd vanish (roughly 1m into a fight) you just send anything on cooldown currently

[twilit hearth]

thanks, i see that changes are still very dynamic, wowhead gets outdated quite fast

[low breach]

I think this might change with 2p but I am not sure. Only looking at what we currently have.

[twilit hearth]

the screen i posted above is with 10.2 4p

[low breach]

Yeah I know.

[twilit hearth]

thanks though, will follow the channel more closely to avoid stupid questions

cheers

[rugged solar]

it is not as easy

dance is in general aligned with symbols, but it will use spare charges

so when it does depends

[twilit hearth]

in 30th second there its the third dance in the sim, it would mean that we have generated 2 dance over 30 sec (1 dance from dust plus baseline 30 seconds of time spent) and another charge purely from CDR, makes sense when you break it up on pieces

[rugged solar]

it is possible, i am haven't looked at the exact sequence

keep in mind that the sequence is also only one out of 100.000 iterations, so it can be that it is not always 100% ideal

[rugged solar]

just for the fun of it

this just adds a check to never use flag in dance/stealth

[golden rain]

So the sim uses flag mid dance?

[rugged solar]

it can happen, it does not force it

[golden rain]

I thought we always flag before dance to not waste a global, and to have the dance be amped

Is this really 4.5% loss to force this

[rugged solar]

yes, thats reasonable

well no

the 4.5% would use dance and use flag after

so instead you would need to introduce a delay

to use flag-dance

testing so was around 0.2% worse

putting a beautification for 0.2% damage loss is not that good

[remote sparrow]

Btw fuu, idk if I've seen eleems dust timer sheet. We have a pretty good timer setup now for 2 min trinkets maybe we can get going in sims too

[rugged solar]

hachadino mentioned some dust things to me, i haven't rly looked into it

if you have ideas, sure

we can try

[remote sparrow]

0 all + branch
1.5 all
2.5 all+branch (this go flag dances then vanishes to reset blades)
3.5 empty flag
4.5 blades then vanish to reset flag + branch

I was wondering if we could setup variables for the variations of flag blades vanish

Like their ordering based on the CDs of the trinket

Like if there's 30 seconds left on blades and u have vanish and flag and trinket up : variable variation_1 is true

So u flag vanish blades

If u have all up, it's condition 0 true etc etc

[rugged solar]

Some small apl updates:

• Backstab/Gloomblade during dance should sequence better with this change
• Split up the Thistle tea condition in 3 separate conditions and slightly modified the use
• Simplified trinket condition for flag.
  Both only lead to marginal damage improvements, ill skip the Highlights for this change.

[ivory kestrel]

One thing that i also noticed is that using strikes to consume rotten buffs in Aoe while blades are up might be better on the preseason build

or only using 1 charge on storm to keep FW

as strike can crit way higher than storm can hit even on many targets

+the rotten bug

[rugged solar]

rotten bug isn't in the sim

[coarse laurel]

Which one are you talking about specifically?

[rugged solar]

https://github.com/SimCMinMax/WoW-BugTracker/issues/1141

but that one is irrelevant for strike/storm

[ivory kestrel]

normal shuriken storm does it too

so it is actually quite krangled

you coudl get 2 critical strike shadowstrikes

but instead get 2 critical strike storms and whatever on other mobs

[rugged solar]

oh interessting

[ivory kestrel]

fukq

i have mkv

but yea it can be "little" impact as we just ignore rotten in T31

which ones do embed mp4?

[spiral forge]

What does the mentioned premed/sht fix mean ?

[past fable]

#subtlety message

Aka it doesnt fuck up something

[spiral forge]

Ah thx

[rugged solar]

yes mp4 embed, you can just convert it to mp4

hope this worked:

[past fable]

Even after the buff ashes actually looks kinda ass compared to branch, ive heard its good for assa so decided to investigate the APL use of it and found some gains (for dust specifically) but its still looking very mid
https://mimiron.raidbots.com/simbot/report/9AxMkS3h8SY5GbbkRYHWgo

Haste obviously isnt that strong for us, but still thought that it wouldve been an alright trinket

[rugged solar]

ah nice, what was the apl change

[past fable]

actions.cds+=/use_item,name=ashes_of_the_embersoul,if=buff.flagellation_buff.up&buff.danse_macabre.stack>=3&!raid_event.adds.up&!equipped.witherbarks_branch
Havent checked if it clashes with any aoe stuff but i dont think it should

With that it effectively uses it just before sectech (or before a strike/backstab right before sectech)

[rugged solar]

ah, nice

[rugged solar]

actions.cds+=/use_item,name=ashes_of_the_embersoul,if=buff.flagellation_buff.up&(buff.danse_macabre.stack>=3|!talent.danse_macabre|raid_event.adds.up|equipped.witherbarks_branch)```

^ fixes conflicts i saw, need to run it through other talent combos too but seems good

seems fine with my profiles

[past fable]

I wanted it to check with the sectech condition since that seemed logical to me but it seemed worse than just dm stacks

[spiral forge]

Does the sim assume to pick up the orbs perfectly every time ?

[ivory kestrel]

1 1 7

[spiral forge]

So it’s likely to come out way worse in a realistic fight isn’t it ?

[ivory kestrel]

well unless you are blind or lost many figers in a war or have frecuent seizures that impede you to move in a circle every 2 mins

not likely

to mess up

but i think even many blind people or missing fingers can do it quite reliably

[rugged solar]

the timing discribes the time between pickups, so it means pick up 2 orbs and try to pick up the 3rd one slightly later

it allows for wiggle room, it is not somehting thats super time sensitive

[golden bay]

would love to see you use this trinket optimally on tindral and smolderon

unironically, i want to see it, i don't mean this as flame

it feels like simming heavy movement should also come at a penalty to branch timings

[mystic sequoia]

You can add
dragonflight.witherbarks_branch_timing=1/1/1
In custom apl and it'll sim as if you just pick them all up asap. Or like perhaps 1 1 2 might be a little more real

[rugged solar]

^ can click the copy icon on the textbox:

dragonflight.witherbarks_branch_timing=1/1/1```

[past fable]

https://www.raidbots.com/simbot/report/6zRg3yQCZmWephGqE4iDG9
Checked some conditions for dancing and also found some gains where holding of with full cp just before symbols comes up was worthwhile even though it sometimes leads to some energy wasted (presumably because getting more finishers of in dance is worth it). Theres probably some better way to achieve this than what i was doing which was just adding this to the non-stealthed finish lines:

actions+=/call_action_list,name=finish,if=(combo_points.deficit<=1|fight_remains<=1&effective_combo_points>=3)&(cooldown.symbols_of_death.remains>3|buff.symbols_of_death.up)```

Havent checked any other builds besides the standard singletarget one as of now, so maybe its incredibly bad for other scenarios setups ![peeposhrug](https://cdn.discordapp.com/emojis/657382266921812009.webp?size=128 "peeposhrug")

Oh also the shd combo point thing is just actions.stealth_cds+=/variable,name=shd_combo_points,value=combo_points.deficit=0 to send it at 7 cp

[remote sparrow]

^ this is really interesting im curious if its worth it to sod maybe 2-3 seconds before dance and not finish (if not a flag/baldes) set, to get a bunch of stacks before each dance

might be able to get 1/2 extra finishes during dance

[rugged solar]

probably does not work well with the first dance builds

still a interessting idea

i haven't done a lot of sht specific optimizations, maybe there is room for more

e.g. ER would be a primary option for it

[ivory kestrel]

also something i found but idk how to put kinda is not using tornado and SoD together with shadowcraft as the refill can happen and you get to lose the cps kinda

or the other way around might be better to hold SoD for after tornado then get the refills easier there

[rugged solar]

Did anyone test the tier set over a longer period of time

[ivory kestrel]

for the % on finishers?

[rugged solar]

yes, curios if someone found a updated model that seems to work

[karmic sky]

I couldn't find a satisfactory answer in the Sub chat - does "1/1/7" branch timings indicate picking up:

1. Orbs at 1s, 2s, then 9s
2. Two orbs simultaneously 1s after activation then the final one 6s later AKA 7s after activation?

How much testing has gone into different timings? I did some sims and a bunch were non-trivial improvements, but they might be impossible depending on the answer to my question.

You can pick up 2 simultaneously pickup if you:

• walk forward to the midpoint of the front left and right ones and let them drift into you.
• angle yourself to put one orb in the middle of the boss hitbox, grab one of the other two then press shadowstep

[turbid pulsar]

1/1/7 is just press trinket wait 1 sec pick first then pick 2nd 1 sec later, then kite last one for 7 sec

[karmic sky]

Gotcha ty

[rugged solar]

tthe numbers is the time between pick ups

you can modify it if you want to want to have diffrent behaviour

e.g. if you want to pick up all 3 immidiatly:

dragonflight.witherbarks_branch_timing=1/1/1```

[karmic sky]

Is there any particular reason why the T31 profile for Sub has Haste+Vers on its helm? Does Sub actually want at least a bit of haste? Or some sort of "soft" breakpoint?

[rugged solar]

it did sim highest when i created the profile

[karmic sky]

What's involved with building those out? Not doubting your methodology, I was just surprised the profile uses crafted helm and crit+vers cape, instead of crafted cape and crit+vers BOE helm.

It'd result in less haste but it's gotta be possible to offset that somewhere else

I think it's a net +20 agi +22 secondaries making that swap

[rugged solar]

basically, simming a lot of configurations

it is a multi stage process, when going over talents, gear, etc...

[rugged solar]

Slightly better Ashes use, need to look into it more with two on use and on other fight styles
https://www.raidbots.com/simbot/report/a2nED6GgAja8LbvXEn2rKU

[ivory kestrel]

maybe need to look into 1 min on ashes+branch etc

think assa did something for it similar

[low breach]

This for general pw or smolderon?

[lilac crescent]

I assume this is using ashes before sectech instead of with dance?

(so 3+ DM stacks)

[rugged solar]

yes

it is just syncing it with secret

[lilac crescent]

that's how I've been using it either way

[rugged solar]

👍

[lilac crescent]

nice when the feelycraft is confirmed

[ivory kestrel]

top feelycraft was always using it after but before 1 sec

now with 2 secs it is less important tho

cant get sec tech+`2 finishers anymore

[past fable]

Did we accidently change it away from syncing it with sectech? Im sure thats the first thing we changed about its apl, curious as to why its not doing it rn

[rugged solar]

i thought i had it in too

but seems like i didn't

[past fable]

@rugged solar 👀 https://www.raidbots.com/simbot/report/2WANX3toFEFLrrwfJaPj1j

[rugged solar]

👀 what did you do?

[past fable]

actions.cds+=/use_item,name=ashes_of_the_embersoul,use_off_gcd=1,if=buff.cold_blood.up&(buff.flagellation_buff.up&talent.invigorating_shadowdust|buff.shadow_dance.up&!raid_event.adds.up&!equipped.witherbarks_branch)
Not sure the offgcd is needed but i was tinkering with other stuff before

[rugged solar]

my change was only a small test

ahh cold blood

ye makes sense

i thought you could maybe simplify it

[past fable]

Could probably add something to make sure you use it before the encounter ends

[rugged solar]

but will need to test

test 2:yaml actions.cds+=/use_item,name=ashes_of_the_embersoul,if=buff.danse_macabre.stack>=3&(buff.cold_blood.up|!talent.cold_blood)
https://www.raidbots.com/simbot/report/h5qMVkrifkd92d3AoFjuDj

but need to test other talent combos and fight styles

[past fable]

https://www.raidbots.com/simbot/report/u6QoCausB5XfonTL7xVsoz
Swapping more stuff to mastery makes it about the same as branch actually with this apl change

[rugged solar]

oh interestsing

[rugged solar]

Ashes of Embersoul improvements:
In collaboration with Eleeem & Raz.

What does this change do?
Syncs Ashes use with Cold Blood / Secret Technique

Changed lines:

actions.cds+=/use_item,name=ashes_of_the_embersoul,if=(buff.cold_blood.up|(!talent.danse_macabre&buff.shadow_dance.up|buff.danse_macabre.stack>=3)&!talent.cold_blood)|fight_remains<10
actions.cds+=/use_items,if=!stealthed.all&(!trinket.mirror_of_fractured_tomorrows.cooldown.ready|!equipped.mirror_of_fractured_tomorrows)&(!trinket.ashes_of_the_embersoul.cooldown.ready|!equipped.ashes_of_the_embersoul)|fight_remains<10
```
__New APL:__

[rugged solar]

Highlights:
St simulation

[balmy linden]

So ashes is better now?

[rugged solar]

better than bark if played ideally? no

closer to bark, yes

Dungeon Slice - Non Dust

but to give a quick tl;dr about bark

if you can use it ideally
so pick up 2 and then delay the 3rd orb as long as possible.
It is the best trinket by quite a bit

then more you go down from that, then coser it gets

so overall, bark will always edge out other trinkets if you can do 1-1-9

and will be slightly better with 1-1-7

[golden rain]

if im reading this line correct it doesnt care about syncing with a cold blood sectech in blades, but any cold blood sectech in the fight (even outside cds)?

[rugged solar]

yes

it does not consider blades

[past fable]

Sectech clones doesnt work with blades anyway

[remote sparrow]

Looking at my own logs (and some of ther other top logs) for branchmark avg uptime per cast (as in, the full 20 seconds etc) on an easy fight (volcross) avg uptime seems to be roughyl 15ish seconds (just looking at logs havent pulled in any actual data)

realistically in harder fights, this might drop, so embersoul might actually be the better play

[rugged solar]

eleems sim also shows something to explore

changing stats can work well for ashes

narrowing the gap even more

[remote sparrow]

as in dropping some haste?

[rugged solar]

going more mastery/vers heavy

[remote sparrow]

yeah thought so as well.

also there are some other considerations for ashes vs branch as JPC said. if youre getting aug'd ashes seems to be more overall damage for the team vs branch (if played perfectly) is higher damage for you

[rugged solar]

its a interessting dynamic

i like it

having options is good

[ivory kestrel]

ashes also unlocks the secret build sepsis pog

[rough dove]

so we can basically macro ashes with sectech now for raid?

[ivory kestrel]

with cold blood in this case

but uh can make it do weird things

[remote sparrow]

I dont think u can sync them

imagine a 3:30 min fight

u can send cds at 0,1,2:30 or 0, 1, 2:00, 3:00.

in that case, u probably wanna do the third one, which in this case its not a good idea to macro it. since it would use it on the 2 min go, which is not what u want sine thats a solo flag go vs the 3 go which is flag and blades

[rough dove]

yeah ok true

i just keep it on a seperate bind then

[polar belfry]

why ashes would be better with aug?

[past fable]

Follow-up to these changes and with an extra change of removing the shd threshold too, aka wasting some shd uptime to have it not do dances right before other cds (i checked changing the condition to other fractials, but 0.90-1 all simmed the same)
https://www.raidbots.com/simbot/report/5KN4XtafQC1cPxeWBtDrR7

[rugged solar]

thats quite good

[past fable]

Baseline is the ashes change + some swapped around gear with full vers+mastery

[rough dove]

what is meant by dont dance 2?

[rugged solar]

what the line does is to not use dance if your cooldowns are ready soon

[ivory kestrel]

also lets you pool for some extra sht i guess

[lilac crescent]

me already doing both these in raid, feelycraft wins again

[ivory kestrel]

ehehe

[lilac crescent]

just push buttons bro it works

[ivory kestrel]

crit all sec techs with sophic best advice i can give

[rough dove]

@past fable quick q bout the optimization w/ holding dance

whats the max amount if time we would hold dance for cd's?

like whats the condition you used that it comes out as a gain

[past fable]

Its no different to what guides are saying, its literally just a slight sim dif

Dont dance unless you have symbols/flag/blades up

[rough dove]

hmm ok

cause on fyrakk prog e.g

i send a naked dance with sectech before first move and then send dance+sod w cd's after move

so according to the min/max i should just hold the dance?

cause if i send a naked dance with pooling before i still get my 2 dances + sectech into second flag + sb window

[split crystal]

are there plans to implement the ashes changes?

[rugged solar]

i did ask koji already to pusht the apl update to simc, not sure when he has time to do so

[coarse laurel]

Probably tonight

[modern fern]

Sorry,but what is don’t dance at 2 stacks mean?

[past fable]

Dont dance at 2 stacks of dance

[modern fern]

Cuz sync with symbol?

[past fable]

yes

[modern fern]

Cool , I've been troubled by this problem a lot recently, especially without swift death

[uncut niche]

Blades cant be up for every dance tho, right?

[rugged solar]

no, you can't have blades for every dance

[past fable]

Not all 3, either one of them

[kind rover]

!finality

[rugged solar]

#subtlety <- that's a command for the main channel

[golden rain]

https://github.com/simulationcraft/simc/pull/8573 made a pr for ashes change if you wanna +1 @coarse laurel

[past fable]

#tc-research message
Continuation of this ^:
Swift death still not a good talent for non-dust but atleast its not negative dps. Hes simming without a bunch of raid buffs similar to his m+ group i guess but should be better either way
https://www.raidbots.com/simbot/report/auiWjQvTwVRNCCG73zckVt

appending this onto the flag line (which is starting to seem too long and convoluted for its own good)
&(talent.invigorating_shadowdust|cooldown.symbols_of_death.remains<3)

[rugged solar]

oh, thats interssting

[golden bay]

Swift death is not terrible in aoe, which is what I was simming in the first place.

The reason im exploring non dust builds in m+ is that it just scales better with aug than dust.

It simmed better in aug simms (which basicly sim party dps) than dust, but my aug friend did those so its a trust me bro source.

[rugged solar]

it also works better with aug in raid

but there is a bit of a irony because the loss to play DB is higher than the gain from aug

unless we manage to find more improvements to DB builds to bridge the gap

[golden bay]

well, in the simms we did, DB was ahead if you considered the aug dps aswell, on all target counts

but that was in dungeon slice build

[rugged solar]

unexpected

[golden bay]

i don't think you can sim this on raidbots, you need the simc client

[rugged solar]

you probably can using advanced sim

but not entirely sure, would need to talk to the evoker tcs to know how their simulations work

[ivory kestrel]

you can play with 2 on uses to be a bit better with Aug in general play

that might make it go higher than DB again

as aug does not benefit much trinkets like crab/pips

that can proc whenever and lose value as they proc outside dance/when you get the buffs

[golden bay]

i dont see how 2 on use trinkets would matter, your ashes is on their breath anyways

[coarse laurel]

Thanks. PRs definitely easiest for me atm. I've been very busy over the holidays with family stuff!

[remote sparrow]

Someone mentioned that Lariet might be undersimming in sims right now, apparently has a lower proc chance implemented or something. does anyone have any further info on this?

[ivory kestrel]

lariat should sim fine it has around 3x% uptime

[rugged solar]

not aware of a issue, but maybe i missed a discussion aorund it

[ivory kestrel]

make sure people are not getting confused with having split stats

as that would make it 1x% of both

[ivory kestrel]

Just in case people do not know but still to have somewhere to link from #subtlety the change that makes us do more dps got pushed today https://github.com/simulationcraft/simc/commit/a8703e468302088df6c1e65b92eae861d5b1d185 Basically it reintroduces the bug that is still ongoing from BP finality making it proc always and also how shadowblades works with our new 2P

[sweet arrow]

Attached is a profile I wrote from (mostly) scratch to model the Smolderon fight. It includes boss vulnerability raid events to represent the boss phases.

The profile is broken down into three action lists to model each phase of the boss: (1) the opener, (2) the vulnerability phases, and (3) the downtime between vuln phases (referred to as 'chill' in the profile).

Right now the profile sims at 322k, whereas the top parse is ~351k. Few reasons that I think this is happening:

1. Per this message #subtlety message it looks like Shadow Rupture can be min-maxed. Some of the top parses get 12-15k more damage out of shadow rupture than this profile.
2. I suspect some of my rupture handling is sub-par. It reapplies rupture with the finality buff up while in sdance, at least.
3. The vulnerability is modeled as a damage vuln on the boss, whereas the actual fight places a damage buff on the player. I'm not sure if this is strictly equivalent in the wow world.
4. There isn't any pooling done when going into a sdance. Not sure if this is important. On a visual inspection, the profile rarely goes out of enery while in sdance anyways.
5. I think SnD is being refreshed sometimes during burst phases. Not if I should foricbly refresh it just before a burst phase.
6. The symbols of death timing in the profile is usually 4-7s later than what most parses do. Right now the profile tries to symbols when dance is up, when there's <3s left on symbols buff, and when there's >4s left on dance. I'm not sure what rules top players are using around symbols

Yeah, the profile is not super useful. I did this 33% to better understand the spec, 33% to better understand simc, and 33% for fun.

[rugged solar]

interessting project

The reason top logs can be higher is typically because they are one iteration

so they are affected more by random inpact factors

you don't rly play around the shadow rupture proc, but your output depends on good ones

it is a big impact factor

The Simulation plays a lot better around the vulnerability, but is very hacky

its very visible in the damage breakdown

compared to the default (that does not consider the vulnerability at all)

[sweet arrow]

that is by design. here's a chart from a log

I think the default profile only happens to get those two peaks because they naturally line up with the first and second vulnerability

[rugged solar]

yes

there is no vulnerability specific handling in the apl

what made you decide to write a new apl

instead of modifying the standard apl for vulnerability events

[sweet arrow]

I did start by modifying the default. Let me see how far I got with that.

306k overall (322kish on the custom one linked above).

[rugged solar]

oh interessting, not too bad

[sweet arrow]

I think there's an easy improvement in my changes to the default profile - its not using symbols anywhere near enough

[ivory kestrel]

you DO play around rupture for the highroll(gambling) and like fuu said across a lot of tries you will se the most dps being the ones that hit

all the shitters that procced it on a non finality one outside dance

are buried

[sweet arrow]

symbols of death usage. 'Expected' is where a top parse used symbols, 'actual' is when the profile did.

[ivory kestrel]

so in sims you cant really go and try to emulate "gambling" as across 100000 iterarions gambling is not really gambling

[rugged solar]

it actually would be

it would show you how successful gambling as a strategy is

[ivory kestrel]

you will just see a higher spread

as when the gambling fails you would do even less dps

but highest dps would go up i guess

[sweet arrow]

I do use the dps range as a metric to control for that

[ivory kestrel]

i actually wonder how it would pan out

[rugged solar]

i assume you could check how common gambling is by looking at rupture casts

i don't expect it to be super common, but maybe i am wrong

[rugged solar]

the max is 374k, so above the top log even

it is a nice project tbo

very cool

[sweet arrow]

I made a tindral profile too. its pretty accurate as well

uses a raid_event.stun to model the flight phases

[rugged solar]

haha great, you did rly sink some time into it

[sweet arrow]

oh yeah.

I was seeing these giga parses from other rogues and just couldnt understand how they were doing them. Building the apls seemed like a reasonable way to figure it out

[rugged solar]

it seems more effort than breaking down cooldowns and fight timers

i am also not sure your vulnerability timings on smolderon are correct

https://lorrgs.io/spec_ranking/rogue-subtlety/smolderon

would be the exact timings

rank 1 log has very good timings for parsing

the boss dies just as the last intermission ends

it is also interessting to see that the fight timer got down this low already

rank 13 killed it in ~5 minutes 45 secs

[rugged solar]

^ this might be another factor why the sim is lower

[sweet arrow]

the first intermission starts at 1:08, but you don't get the actual damage buff until a few seconds later

[fickle shore]

well you could chose to delay the damage buffs a lot

[past fable]

Althoug hyperfocusing on a single top log is likely not ideal since even those contain missplays. Like we see with symbols here, you can clearly very easily fit another symbols in before the first intermission, yet the log didnt

[fickle shore]

if that was your desire

[rugged solar]

oh, you are correct

[sweet arrow]

you're not wrong, but I needed something to base the timings on. Otherwise I'd just be shooting in the dark, adjusting numbers until I got good dps

[coarse laurel]

Yeah I think you’re right that some high roll optimizations aren’t going to show up in average DPS. You’d have to optimize for max DPS being the metric rather than mean. (Not sure I think this is great other than parse fishing though. But you’re right that highroll optimizations aren’t always gonna yield a higher mean. Maybe even lower.)

[ivory kestrel]

Yea that is why sometimes you dont really want to put them in or well it does not make much sense the situation is rare enough where you have X happening while Y and then hit a % on Z for it to just be a gamba thing you say it is possible but just leave it there

we just have 3 bosses where it can lead to big highroll gnarl-smold-tindral

[rugged solar]

If something is rare enough and clearly better, it is easy to have it in the apl. A lot of the optimizations we have are exactly this

It becomes complicated if you want to do something that is overall a damage loss with the promise of higher variance

we never tried to remove gambling if it is overall damage positive

if we want to explore tier set gambling, it would mean having the 0-33-66-100 proc rate in the sim and a option to check the state is required. And unless i missed the commit, the sim is still using a proc chance.

[split crystal]

Implementing the 0-33-66-100 proc rate in the sim would answer how to proceed when you have 100% chance for shadow rupture but no finality before dance. It makes sense for the sim to only rupture in dance with finality given the current implementation.

[past fable]

Its worth noting that the rupture in dance thing is worthwhile even without tier, although not by as much

[split crystal]

I also wonder, for the opener, if it's worth dancing before the first rupture - sim might have a different result if it knew there was a 0% chance for 2p proc, and your 2nd rupture cast would be inside the 2nd dance with finality and 33% 2p chance like normal

Essentially delaying dance in standard opener by 1 global

[golden rain]

3 mins sim: https://www.raidbots.com/simbot/report/teCCQ2Pr3pDENkMDdpChGA
4 mins sim https://www.raidbots.com/simbot/report/q9dmViRktdZc43HP3Hdu8X
5 mins sim https://www.raidbots.com/simbot/report/h5mM7YiGwKXEyTBW6fqSoa
did some sims at lower fight times trying to force a triple dance opener, seems doing a triple dance opener but saving dance for CB in cooldowns otherwise is a consistent gain on faster fights. Gains:

+2.5% at fight lengths <=3 mins when doing fixed sims
+1.3% at 3 mins (variable)
+0.6% at 4 mins (variable)
-0.3% at 5 mins (variable)
much worse on long fights

apl changes:
adding this stealthed list to vanish strike near the end of dance for premeditation if at 2 vanish charges (copy pasta from 10.1 apl)

actions.stealthed+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2&(cooldown.secret_technique.remains>=30|!talent.secret_technique)&cooldown.vanish.charges>1

shadow dance condition changed to dance if >1.75 stacks of shadow dance and the cold blood condition is not met, and also to dance regardless of sectech CD. This last part is necessary because 3x dancing on pull can cause a future dance to not have a sectech if unlucky with procs

actions.stealth_cds+=/shadow_dance,if=(dot.rupture.ticking|talent.invigorating_shadowdust)&(variable.rotten_cb|variable.sdh_threshold)&(!talent.the_first_dance|combo_points.deficit>=4|buff.shadow_blades.up)&(variable.shd_combo_points&variable.shd_threshold|(buff.shadow_blades.up|cooldown.symbols_of_death.up&!talent.sepsis|buff.symbols_of_death.remains>=4&!set_bonus.tier30_2pc|!buff.symbols_of_death.remains&set_bonus.tier30_2pc))

[rugged solar]

nice

[golden rain]

I think with some more investigation/cleanup it's possible to find a gain/be neutral for 5 mins as well. Note that the variance went up from the default sim slightly and I noticed some edge cases:

• there are some bad rng iteraitons where dance is used a little weirdly (without SoD)
• it might be possible to avoid the unlucky dances without sectechs by sending SoD first and then using dance a few seconds into your SoD, rather than using them together
• probably some gains to force 3x dances at future cooldown usages (e.g. 2 mins or 2:30), in particular on the last set of cooldowns in the fight is almost certainly a gain
• on bad iterations cold blood is used pretty late; its possible delaying dance/flag/blades a few seconds if cold blood is close to coming off cooldown to guarantee 2x cold bloods in the CD window is gain as well, as long as you dont lose a cd usage throughout the fight

[rough dove]

so how would we play out an optimal opener with triple dance then?

cause the lines kinda confuse me

i assume vanish after sectech and for the 4th dance we would send sod and a late dance to secure a sectech dance window no?

and for the 1min cd set where we pull blades

cb timing is kinda whacky

like 4-5sec off from insta sending flag when its up again

do we hold then to sync with cb and to get 2x cb in the cd window?

• on fights like fyrakk where we have some downtime going from phase to phase

it should be a gain aswell to triple dance opener even tho fight is longer than 5min no?

but what ive noticed already from hitting the dummy

triple dance opener only really works consistently with WF

otherwise you lack too many shd stacks for resets

this are my notes from hitting a dummy for an hour now and what i noticed

appreciate it if i could get some answers to the q under the 3rd point

[golden rain]

Your last 2 questions are why this isn't really a gain on longer fights lengths from what I could tell. You end up having to lose a CB or lose some cdr on blades or lose a CD usage somewhere. Something like fyrakk is a bad place to do this unless your CDs were idling anyway (depends on your push timing). But on shorter fights it doesn't matter because you won't lose a usage

[golden rain]

More simming needs to be done but in these cases I would probably insta flag sod to avoid delay then dance just a bit later to hit that first CB

As long as you are getting 30 stacks of flag reliably should still be good

[crimson vault]

should it be a gain for 3/10 tech stacks burst aiming for 7 finisher in 1st dance?

with 1st finisher sectech and in-mid tornado

[golden rain]

can we push the mirror/fragment/grieftorch custom apl to github

people are making mistakes in their gearing very often because trinkets show as downgrade

[remote sparrow]

i think they submitted a push request

theyre just waiting

[rugged solar]

its soon™️ in simc

[rugged solar]

^https://github.com/simulationcraft/simc/commit/8296d72323c715c86c89484cea497d626c7a428f

[remote sparrow]

I guess to get things started again. using the old APL we had, just deleted the rupture in dance logic. Best simming builds for both hero talents are still dust

trickster

deaathstalker

SF is about 0.1% dps

and GB seems to be the strongest choice out of premed/tfd/RS/Swift

[rugged solar]

ye, it probably is fairly irrelevant

the nerf and it being only stealth is quite a deficit

[remote sparrow]

yep its insanely weak

dust vs db builds for deathstalker are about 7% different

I tried some conditions for refreshing rupture to have enough corrupt the blood stacks but it seemed kinda irrelevant

not sure if any optimizations can be done for deathstalker

trickster im sure theres a lot we can do

[rugged solar]

deathstalker is fairly passive

[remote sparrow]

also planned execution and warning sign are very very close

which was interestin

[rugged solar]

warning sign was never that far behind

[remote sparrow]

yeah its like 0.3% behind now

also fatal intent doesnt seem to have been implemented

[rugged solar]

i am fairly certain it isn't that diffrent to retail, moment

its 0.5% on retail - sim

[remote sparrow]

its a lot more behind for trickster, afaik deathstalker just likes haste

[rugged solar]

can even be a win, if you disable raid buffs - sim

[remote sparrow]

ah interesting

[rugged solar]

haste as a bad stat is a bit of a common missconception

it can be quite decent esp. if gear just like in 10.2 is using very little

[remote sparrow]

makes sense

but yeah i have no idea what to even test for deathstalker

I thought maybe trying to get it to only darkest night on finalitty:evis

but even in practice it was really hard to do it

idk

[rugged solar]

is dust winning or losing currently?

if db is stilightly behind, having db and other amps work could easy shift it a bit

[remote sparrow]

oh dust is winning by abut 7%

[rugged solar]

thats quite a bit

[remote sparrow]

yep.I think with some optimization trickster is gonna be pretty close if not ahead

but both hero talents seem to want dust

[rugged solar]

dust is just good

because you get more finishers

and dances

and the tier set as well as the hero talents give you benefits from it

[remote sparrow]

u get more dances, more cds, more benefit from tier. idk man they just dont understand the problem with dust and keep making the spec worse with the changes...

theyre probably gonna nerf it to 20 seconds so we dont play it.... such a bad decision that would be too...

[rugged solar]

thats what i am concerned about overall

i don't think 10/20 works well

[remote sparrow]

I dont think 20 is even enough to actually make it worse

just makes it clunkier

[rugged solar]

without dust felt fine, but i think it is a bit too dance starved

dust as stupid as it sounds felt a lot better to play on raid testing

[remote sparrow]

oh I know. was the same for me

its really unfun and clunky

to send an empty sod + sectech

also not sure why but tornado vs veiltouched seem to sim identical for some reason

oh 1 more i guess deathstalker optimization would be to storm after consuming a mark for clear the witnesses

[rugged solar]

tornado is optimized as a filler

so maybe it leads to just enough additional cdr to have a better allignement somewhere

[remote sparrow]

might be

for trickster its about 0.9% loss

for deathstalker its even

well at this point probably best simming single target build for deathstalker and your dslice build are less than 1% different in their single target damage lmao

[rugged solar]

lol

to be fair the capstone/cantrip make deathstalker maybe a good option in m+

[remote sparrow]

yea the defensive node is also a lot btter

and the utility node

its aoe is not very good but neither is tricksters so eh

[rugged solar]

the BF equivalent seems rly good

i wish it was without auto attacks

[remote sparrow]

Idk somethings wrong with that node

like its bugged in a way that i cant tell what it is

i know that 1) it double dips armor 2)doesnt work with coup 3)when u proc nimble with backstab it doesnt cleave the UB that you proc

but something else is wrong with it

i did 4 keys today, would go to a pack of like 7 mobs

pack would last for litearlly 10 seconds

so i have form for 100% of the pack, i would evis

and my evis would end up higher than nimble

[rugged solar]

it is possible that damage calculations are not correct

[remote sparrow]

its hard to get an exact idea becuse its reduced by armor on the other targets for a 2nd time, but reduced less if that target has FW on it

then theres 5% variance too

it just becomes too hard to tell exactly how much less damage its doing

[rugged solar]

i would need to look at logs

but assume a lot of damage is not attributed correctly

subtlety has a lot of things like finality, damage amps, shadowed finisher, ...

[remote sparrow]

i was going through sapp's logs and there was instances were nimble just was not doing damge when it should have

like autos should be always cleaving but nimble was doing 0 damage during that time

[rugged solar]

its a repeating problem of things not working with what you would assume it does

[remote sparrow]

but yeah, im sure theres quite a bit of optimization we can do for trickster both in st and in aoe

[rugged solar]

esp. in cleave/aoe

because you 100% won't use bp on 3 targets

[remote sparrow]

eleem and I did the math

it should be higher than bp until 11 targets

but thats assuming no bugs

[rugged solar]

my rough estimate was between 9 and 13 targets

depending on implementation

my hope was it to be lower, so bp would stay relevant

11 seems too high, but i don't rly think the talent should be nerfed

[remote sparrow]

bp is still relevant for larger pulks, when you cant be in melee to evis, downtime when no form

[rugged solar]

yes absolutely

you just need a weakaura to know when its up

not ideal

[remote sparrow]

oh u need a lot of new thnigs to track

fazed and form are seperate in duration and activation

[rugged solar]

i keep it to "no new things"

and will add things if they rly seem to be needed

in my weakauras

[remote sparrow]

u prob also need to track how many UBs you are going to get after sectech

[rugged solar]

i did some very experimental changes to the apl with armin.
They are nowhere finished, but you can use this apl as a baseline if you want:

[remote sparrow]

Darkest nice doesnt seem to be triggering as often as we'd expect even with the line to force max cp. Ill look into it to see whats happening

[coarse laurel]

Yes because of the change to Deathstalker's Mark not being able to be refreshed in the latest build it prevents it from triggering/being wasted. You guys will actually need to test how this works in-game probably to confirm.

[remote sparrow]

im not sure I follow, how does the mark not being refreshed cause the # of total triggers to be this low over 5 minutes?

[coarse laurel]

Because we don't allow Darkest Night to be consumed if the debuff is already on the target

Based on Whispyr's testing a bit ago, that appears to be how it works in-game as well, which may be unintended but I figured that's how it would work after the limit was put in

[remote sparrow]

so if there is a deathstalkers mark on the target already, and I also have the darkest night buff, if I finished with maximum CPs the darkest night buff will not be consumed and the effect will not happen?

In game, if the target has a mark, and I have the darkest night effect, when I do a max cp evis on the target, it will first consume 1 stack from the target, consume the darkest night (with its effects) then reset the stacks on the target to 3.

[coarse laurel]

@hearty finch was seeing different things when he mentioned it to me, so idk if you guys can grab some logs or something

[hearty finch]

Yeah I’m out today but the combo was get capstone -> reapply mark with a strike -> darkest night can’t be consumed at all

[remote sparrow]

Ill test it again, but on sub, you can enter dance with darkest night up and 0 marks on the target

press strike -> get 3 marks

press evis at full cp -> consume 1 mark -> goes down to 2 -> consumes darkest night -> resets back to 3 marks

ill try and log it one sec

https://www.warcraftlogs.com/reports/rTMC8FNBR9YdAZgh#fight=last&type=damage-done&source=13&start=6685&end=28408

evis are at 12/14/16 (this one gives the darkest night buff)

strike at 20 applies 3 new marks

evis at 21 consumes darkest night buff

crits and does extra damage

ends darkest night buff

also kinda insane that this is a randomly generated char name but 2 rogues with this named existed at one point. what are the odds kekw

[coarse laurel]

Is that just a logging issue or did the Shadowstrike only apply 1 stack initially?

I assume it's a logging issue

[remote sparrow]

I think its a logging issue, it applied 3 on my screen

[coarse laurel]

Ok

yeah not sure why Whispyr was seeing something different here. I wonder if there is some other case where it bugs

[remote sparrow]

Ill let you know if I can replicate what whispyr saw! I know his bug tracker is correct though, with teh marks being consumed by builders etc

[hearty finch]

second video is a bit jank

oh are my bars weird

[remote sparrow]

did u rupture instead of evis?

[hearty finch]

yeah just saw that

MY BARS ARE FLIPPED

aight it's fine then

fuck stealth bars

[coarse laurel]

I'll make some changes

[hearty finch]

blizzard please let me turn this shit off

[remote sparrow]

https://tenor.com/view/joe-biden-smile-joebiden-donaldtrump-gif-5341877189215905052

[hearty finch]

nah I'm cooked af for that one

[coarse laurel]

Does the Darkest Night crit buff always work on the Shadow damage portion?

(It does in SimC, just wanted to check)

[hearty finch]

ye

[remote sparrow]

ill have to double check, it didnt give the shadow part the amp I think last week

could have been fixed this week ofc

[hearty finch]

I'm digging but its looking like the crit works

https://www.warcraftlogs.com/reports/mAZhap1zYd2x9PBN#fight=last&source=8&type=damage-done&pins=0%24Separate%24%23244F4B%24auras-gained%240%240.0.0.Any%240.0.0.Any%24true%249072560.0.0.Rogue%24false%24457280^0%24Separate%24%23909049%24casts%240%240.0.0.Any%249072560.0.0.Rogue%24true%240.0.0.Any%24false%24-196819&view=events

can ctrl+f "darkest night fades" on that

it's not a large sample size but I saw a couple more on my details scroll as well

[hearty finch]

it didnt give the shadow part the amp
as in like it didn't crit or what

[remote sparrow]

no like, evis does 60% more damge and crits, but im not sure if the shadow part is getting the 60%. I think it didnt last week, but maybe its fixed now

[hearty finch]

in that log it looks okay

30% of 867 is 260

30% of 764 is 229

5% from veiltouched

[remote sparrow]

its hard to tell, because the shadow part also has about 30% shadow amps on it right? from dd db vt

[hearty finch]

it's pretty close

idk if I had db

vt/dd gets it in the ballpark though

idk

[remote sparrow]

one second im gonna do a test without vt dd db

its less than 25%

[hearty finch]

[ivory kestrel]

maybe it still has the krangled aura

of -30

shadow evis is another spell anchored to evis

same thing happened with coup

for a bit

[hearty finch]

in the log it's like 32-33%, but I guess if it was 30% times all the multipliers it should be closer to 50

[ivory kestrel]

yea usually it is a bit higher than 50 but it kinda checks out if instead of 3x goes to 17

[hearty finch]

so many interactions that can break man

jenga ass spec

[remote sparrow]

yep. legit cnat even test these things

5% vt... cool did i get variance diff'd? does it work? who knows

[hearty finch]

wait does shadowed finishers have variance?

residuals shouldn't have variance I thought

[remote sparrow]

No it doesn’t, I just meant as a more general comment

I’ll do some more testing and math to see what’s happening

[ivory kestrel]

at least it is auto crit so easier to test but still

fuq

[coarse laurel]

The issue is Darkest Night is totally scripted as far as I can tell so the 60% damage buff is kinda hard to verify what it applies to

[past fable]

I can Recommend buying a level one dagger from a dude in orgrimmar for testing since then its only the 5% variance

[remote sparrow]

okay so this is a normal evis distribution between main hit and shadow hit

about 36%

the darkest night one is abotu 23%

so we can probably assume its bugged and the amps isnt being applied

[ivory kestrel]

well that is better than worst case at least

[remote sparrow]

also new bug, they fixed cb working with coup (all 6 hits), but now it straight up just doesnt work with evis anymore i cant

[ivory kestrel]

thankfully you dont use it on evis but that one sure is funny

[remote sparrow]

this is without any amps. no vt, no dd, no db

both tests were withot ayn amps

[coarse laurel]

11.682 Player 'baseline' eviscerate hits Player 'Fluffy_Pillow' for 2231995.377228188 physical damage (crit)
11.682 Player 'Fluffy_Pillow' decrements Buff deathstalkers_mark by 1 to 2 stacks
11.682 Player 'baseline' gains Buff deathstalkers_mark_buff (stacks=1) (value=-2.2250738585072014e-308, time_duration_multiplier=1)
11.682 Player 'baseline' performs Action deathstalkers_mark (183.11866244249998)
11.682 Player 'baseline' deathstalkers_mark hits Player 'Fluffy_Pillow' for 147919.47110262 shadowstorm damage (crit)
11.682 Player 'Fluffy_Pillow' loses Buff deathstalkers_mark
11.682 Player 'Fluffy_Pillow' gains Buff deathstalkers_mark (stacks=3) (value=-2.2250738585072014e-308, time_duration_multiplier=1)
11.682 Player 'baseline' refreshes clear_the_witnesses_1 (value=-2.2250738585072014e-308, duration=12.000, time_duration_multiplier=1)
11.682 Player 'baseline' consumes 33.25 energy for Action eviscerate (149.86866244249998)
11.682 Player 'baseline' consumes 7 combo_points for Action eviscerate (0)
11.683 Player 'baseline' loses Buff darkest_night

Think that should be correct

[remote sparrow]

looks reasonable! I guess for now, maybe we can make the shadow part not gain the benefit for more accurate testing (but they'll probably fix this since it seems like a bug?)

[coarse laurel]

Also

[remote sparrow]

up to you

[coarse laurel]

Given the ordering here

What happens if you consume the last stack with a max CP Eviscerate? Instead of the 2nd one?

Since the consume seems to be happening first, does it immediately benefit?

[ivory kestrel]

consuming last stack only gives you the buff to proc the big boi

[remote sparrow]

so as in -> target has 1 mark, u do max cp evis. does this evis get the darkest night buff?

[coarse laurel]

Yes

[ivory kestrel]

nope it has to be the next

[remote sparrow]

i dont think it will one sec will test

[coarse laurel]

Given that the Mark is decremented before the Darkest Night buff seems to be handled

[ivory kestrel]

ah that is because of a shadowstrike interaction

[remote sparrow]

nope does not get it

[coarse laurel]

(Which is why it goes 2->3 rather than 3->2)

[ivory kestrel]

you can shadowstrike when you have 0 marks and you get 3

then when you consume darkest night the buff goes from 2 after spending it on evis to 3

[coarse laurel]

No my point is about the ordering when you Eviscerate to begin with because a Darkest Night cast trigger can still decrement and still proc

And that appears to happen first

[remote sparrow]

yes, mark is consumed first by evis. then darkest night refreshes mark after the evis goes through

[ivory kestrel]

yea this is a failsafe to prevent double nights

from happening

no idea if it is intended or just a guardrail

[remote sparrow]

also the shadow part is 100% bugged

[ivory kestrel]

yea that is krangled

[hearty finch]

Hey the crit works though right

[coarse laurel]

Alright should be updated

[remote sparrow]

yep. crit works

[remote sparrow]

thank you! will test now!

[remote sparrow]

This is a log for testing nimble on 2 dummies. the 2nd dummy never had FW.
https://www.warcraftlogs.com/reports/48hMD3QjCkyJaVtc#fight=last&type=damage-done&start=501380&end=523185&source=4&eventstart=512068

couple of things going through this. 1) I dont think blades replicates UBs damage

2. Main evis and autos seems to be doing 34% of damage replicated instead of 40 , shadow strike and shadow evis seem to be replicating 27% of damage instead of 40

and ofcourse nimbles damage is not working with blades

nimble also not working with coup

[coarse laurel]

This is because of target modifiers being reversed and lack of Find Weakness on the second target, which is the same as Blade Flurry. So this is fairly expected.

[coarse laurel]

Not totally sure about this one though, unless I'm missing something. In those logs:

[remote sparrow]

oh you are correct, I think i just missed it on those. nice catch!

[remote sparrow]

this one im still a bit confused why evis / shadow evis / strike are all different.

i would assume both physical hits (evis /strike) to have same modifier

and then sahdow evis be different

but also, if the physical parts (evis/strike) are being reduced by armor, and then nimble is getting reduced by armoor again, its double dipping in armor right?

[coarse laurel]

Nimble not working with Shadow Blades is actually kinda funny because I think probably it should, since Shadow Blades damage doesn't Flurry.

How this works with Blade Flurry is that the base (non-mitigated) damage is replicated which effectively makes the target modifiers (armor, Monk buff, etc.) on the primary target irrelevant

So if your main target has FW on it, you can expect Flurried damage to secondary targets to be slightly less than the normal ratio

Since the damage will be reduced by the higher armor amount

[remote sparrow]

so if physical hit does 100 damage and target has FW, then nimble says the originial hit must have been 120 (to get the roughly 17% reduction from armor after fw), takes 40% of that (48), applies armor to it (30% reduction) and so the replicaated damage becomes 33.6. very close to what nimble is flurrying in logs rn

makes perfect sense thank you for explaining

[coarse laurel]

To use made-up numbers since I cba to look up armor formulas right now
If enemies normally have 30% armor, and you reduce the main target to 20% armor due to FW you get:
10k base damage * 0.8 armor => 8k damage
This then will Flurry for
10k base damage * 0.4 flurry * 0.7 armor => 2.8k damage

Shadow calculations are less likely to be different since we don't have and magic damage taken debuffs

[remote sparrow]

this is the part that doesnt make sense to me still.

this shadow evis damage got flurried for about 26%

[river folio]

I think this dynamic has existed forever right? Like even in vanilla classic this is how blade flurry worked. I do not recall how it worked but I think there were cases where you would hit a primary target with normal armor and if a tank with extremely high armor next to them took damage the blade flurry would just eat through their armor or something

[remote sparrow]

based on our calculation from earlier, it should be roughly 33.6 for 100 damage hit, so about 33-34% will be flurried

[coarse laurel]

Well that has its own problem where the Flurry of a Shadow damage source will be reduced by Armor

When it wasn't reduced at all on the primary target

[remote sparrow]

its getting double reduced I think

ah never mind, youre right

its just getting flat out reduced

[coarse laurel]

Shadow Eviscerate damage is not truly a residual damage, it's just a second spell with a coefficient that matches up what it should be.

So it shouldn't be double dipping I don't think, but the damage will not look right lol

[remote sparrow]

a 100 damage shaodw hit gets 40% flurried (40 damage), then (30% reduction from armor) ~ 28 damage.

so roughly 28% will be flurried

so i guess that ability is working as intended (except maybe nt working with blades)

[coarse laurel]

I don't think I had it working in sims on the shadow portion yet since it was broken at some point, but I'll put that in now

[remote sparrow]

Thank you! that'd be great! once its in, I can try optimizing aoe a bit

[rugged solar]

can someone tl;dr me the last 200 messages?

[remote sparrow]

there was an isssue wiwth darkest night in sims which we fixed, darkest night is bugged with shadow part of evis (makes it cri but not get the 60% buff), nimble is working correctly (its roughy a 33% flurry after armor calculations), nimble doesnt work with blades, coup. we figured out armor redction behaviour of nimble and koji said he'll add it

[rugged solar]

thanks

[remote sparrow]

something eleem noticed.

the first stack of unseen blade happens at 18 seconds.

but the bot has 9 stacks of flawless form at 16 seconds

and 4 stacks of forma t 15

which suggests the evis at 15 was actually a coup

but it didnt show the sim gaining stakcs before it

[past fable]

I played around some and found a couple of small things as gains:

1. Dont need to use more than 4 cp on Snd anymore now that we have cut to the chase available
2. Remove ruptures in dance entirely Rupture line needs more work, the line i used sucked with finality.
3. Change it so the sim lines up the "Imperfect Ascendancy Serum" trinket with blades better, theres still some more work that could be done for this but i have doubts wether this trinket is gonna remain as is since its absurdly powerful and likely to warp the entire spec/APL around it, not sure if we should commit to doing that yet

https://mimiron.raidbots.com/simbot/report/nZYZrnNNGqZQG5h1QrZFhj
This is a non-dust sim with DB

[ivory kestrel]

yea the trink idk what to think but my best performances where using it just after flag before the strike or just as you strike

also might be important to make it use sec tech first that way you can always get it twice in 20s

if you use it second finisher you may not be able to it inside trink

+cb but havent tried that one extensively

[rugged solar]

nice, seems all like good changes

[ivory kestrel]

Also should help slice if we put evis outside dance when snd is less than 5 secs or so duration although 10+ to be sure would be ok too so it does not mess it up before cds

even for deathstalker

[valid robin]

The trinket usage is still relevant no? Like even if they reduce its numbers, it’s probably better to line up with blades

[past fable]

I wasnt referring to just the trinket usage. But also im kinda expecting them to change how the trinket works, not just reduce the numbers

Since the trinket is like 8-10% better than other options for dB setups (despite being lower ilvl)

[ivory kestrel]

yea chances are they completery snap it in two

also to remind that you can get it on hero ilvl so diff vs max mythic trinks is not as big as expected

so yea it has a death sentence now just wonder how it gets executed

[rugged solar]

the trinket is maybe a bit too much over budget

[modern fern]

in vault?

[ivory kestrel]

from delve itself

highest lvl delves drop that ilvl

[modern fern]

wow

[past fable]

Update: The ascendancy serum doesnt have a cast time in sims, thats why the optimal use case is after blades rather than before.
The rupture change was kinda stinky with finality so it needs to be looked into more

For deathstalker i changed it so it didnt use finishers at 6 with darkest night up and i tried different things to align the DN buff with dance but it didnt seem worth it, i still suspect theres some way to make it a gain but i havent found it. Preventing rupture and sectech during darkest night didnt seem to be a gain either, which doesnt make sense to me so theres probably something else afoul
https://mimiron.raidbots.com/simbot/report/5dRenXu4NcBwb9x7NPZegK

[rugged solar]

👍