#tc-subtlety

1 messages · Page 1 of 1 (latest)

rugged solar
#

Experiments:
Shadowdust:

# around 1.5+% gain
+ actions.stealth_cds+=/vanish,if=!cooldown.sepsis.ready&!cooldown.shadow_blades.ready&!cooldown.shadow_dance.ready&!buff.shadow_dance.up&combo_points.deficit>3&!buff.the_rotten.up
``` Seems to be a better condition for dust, needs checks for `talent.enabled`, **Flagellation** is seemingly not a gain to add to the cooldown list.

__Vanish:__
`(buff.flagellation.up|buff.flagellation_persist.up|!talent.flagellation)` seemed to be a gain in prev. iterations, but now neutral.

__Lingering Shadows:__```diff
+ actions.stealth_cds+=/variable,name=shd_combo_points,value=combo_points=2,if=buff.lingering_shadow.up
```Is a ~1% win for specific builds, needs further investigation, and seems to be not a gain anymore with newer apl updates
#

Some rly min/max style changes / experiments

- actions.cds+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2
+ actions.cds+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2&!buff.premeditation.up```0.1% gain for very specific builds, no change to almost everything. Potential to gamble a bit with tea sync with something like `&(cooldown.vanish.charges>=2|buff.thistle_tea.up)` but still not more than 0.1% dps max.
Similar for AoE
```diff
-  actions.stealth_cds+=/vanish,if=(!talent.danse_macabre|spell_targets.shuriken_storm>=3)&!variable.shd_threshold&combo_points.deficit>1
+  actions.stealth_cds+=/vanish,if=(!talent.danse_macabre|spell_targets.shuriken_storm>=3&!buff.silent_storm.up)&!variable.shd_threshold&combo_points.deficit>1
```Is a marginal win on very large target counts
#

Shb optimization for DS & Sepsis

- actions.cds+=/shadow_blades,if=variable.snd_condition&combo_points.deficit>=2
+ actions.cds+=/shadow_blades,if=variable.snd_condition&combo_points.deficit>=2&target.time_to_die>=10&(dot.sepsis.ticking|cooldown.sepsis.remains<=8|!talent.sepsis)|fight_remains<=20```
Small gain for ST with sepsis, small optimization for Dungeon slice.
beween 0.2 and 1.5% damage gain, depening on fight style.
sim:
<https://www.raidbots.com/simbot/report/pd2Cxpkx7QDUM4QrMieJJB>
#

Rotten improvement

- +actions.stealth_cds+=/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
+ actions.stealth_cds+=/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&!buff.the_rotten.up
- actions+=/call_action_list,name=finish,if=combo_points.deficit<=1|fight_remains<=1&variable.effective_combo_points>=3
+actions+=/call_action_list,name=finish,if=combo_points.deficit<=1+buff.the_rotten.up|fight_remains<=1&variable.effective_combo_points>=3

Earlier finishing with rotten + don't dance with the buff
~0.3% gain
sim: https://www.raidbots.com/simbot/report/1eVENDkF7qAnECM6hARYfc

Note:

- actions.build=shuriken_storm,if=spell_targets>=2+(buff.lingering_shadow.remains>=6|buff.perforated_veins.up)
+ actions.build=shuriken_storm,if=spell_targets>=2+(buff.lingering_shadow.remains>=6|buff.perforated_veins.up|buff.the_rotten.up&combo_points>=4)```
ended up neutral overall, even tho it seemed logical
#

Small correction, to account for Improved storm:```diff

  • actions.stealthed+=/shuriken_storm,if=variable.gloomblade_condition&buff.silent_storm.up&!debuff.find_weakness.remains
  • actions.stealthed+=/shuriken_storm,if=variable.gloomblade_condition&buff.silent_storm.up&!debuff.find_weakness.remains&talent.improved_shuriken_storm.enabled**Trinket optimization, pre pull use:**diff
  • actions.precombat+=/variable,name=algethar_puzzle_box_precombat_cast,value=3
  • actions.precombat+=/use_item,name=algethar_puzzle_box**two target cleave improvements:**diff
  • actions.build=shuriken_storm,if=spell_targets>=2+(buff.lingering_shadow.up|buff.perforated_veins.up)
  • actions.build=shuriken_storm,if=spell_targets>=2+(buff.lingering_shadow.remains>=6|buff.perforated_veins.up)```it makes sense to use storm on lower lingering stacks
#

Small improvement to Cold blood handling
thanks to Kojiyama#4176.

- actions.finish+=/secret_technique,if=buff.shadow_dance.up&(buff.danse_macabre.stack>=3|!talent.danse_macabre)
+ actions.finish+=/secret_technique,if=buff.shadow_dance.up&(buff.danse_macabre.stack>=3|!talent.danse_macabre)&(!talent.cold_blood|cooldown.cold_blood.remains>buff.shadow_dance.remains-2)```
#

Thinks i will check sooner or later:

#

General:
This channel is for discussions mainly focused around APL/Simulation improvements and everyone is welcome to contribute.

The pins contain experimental apl changes, hacks or ideas to explore.
Topics will be kept up to date, and if something gets added to the default apl, it will be removed.

Initial pins do contain most of the findings/discoveries collected recently, which often break other builds/talents or need further refinement.

To find the latest version of the apl, you can look at: Github - APL

Note: the channel will be spammy at time when iterating over changes.

valid robin
#

there should be a lot of low hanging fruit to solve there

#

since its literally wasting 60% of the cp gain from the talent

rugged solar
#

yes, seems like something you can get quite a bit more out of it

rugged solar
#

@shrewd geyser, @golden bay, @fading bronze
Grieftorch:
actions.cds+=/use_item,name=manic_grieftorch,use_off_gcd=1,if=gcd.remains>gcd.max-0.1,if=!stealthed.all
sim: https://www.raidbots.com/simbot/report/usMmtwcarcGu2tjrP5MzF9
Putting torch on a off gcd use does opt the biggest dps increase, also it seems to make other optimizations seemingly irrelevant given that you only lose 1 gcd.
(jhon did send me a sim with off gcd use)

#

not sure how reasonable a perfect off gcd use is

#

but it seems doable

rugged solar
#

testing on dummy, off gcd use does seem fine

#

its funny how the trinket is off gcd usable^^

valid robin
#

It’s very easy, just slam it instantly

rugged solar
#

yes, you can just kind of press both buttons together

#

it seems like a easy change, maybe something to add with the other changes

valid robin
#

I assume you still don’t want to use it during blades or symbols

#

Since ur just wasting a few seconds each use

rugged solar
#

you waste one gcd

valid robin
#

Ye

rugged solar
#

what seems to make the shb lines fairly irrelevant

#

but moment

valid robin
#

Massive kekW

rugged solar
#

the optimizations kekw

#

2 is slightly worse when running it 3 times in general

#

some times 0.1% behind

#

it just adds a check for shb

#

3 is just a try to get better use at the end of the fight

#

but seems all fairly irrelevant

#

"just send it"

shrewd geyser
rugged solar
#

i feel like torch will stay a meme for a while

#

with the "thanks for dying" weakauras and stuff

valid robin
#

It might survive to the next tier

#

Simply cus of the reset mechanic

#

And the ability to have a 400k dmg nuke

#

Is more valuable than simple dps numbers

rugged solar
#

possible, it is quite op

#

but also thats the reason it is so good or fun

shrewd geyser
#

pretty sure that happened with draught of souls in legion which was a pretty similar trinket, people were still using in ToS

rugged solar
coarse laurel
#

If you're gonna macro it to some spam ability it's probably worth checking if [nochannel] is worth it on the GCD ability to keep from accidentally interrupting

rugged solar
#

@dire glacier did Dm me for Tea improvements.
I did only try some small adjustments, so maybe there is more to gain.
After trying multiple lines, the only improvement that i got from a "quick" look was:

- 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)
+ 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.remains>=4&!buff.thistle_tea.up&spell_targets.shuriken_storm>=3|!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)```
Maybe there is more gain somewhere, syreldar iirc has a line change with some improvements.

The changed line is only a minor upgrade, ideal conditions are multiple targets.
the sim is on 4 targets.
sim: <https://www.raidbots.com/simbot/report/a7vwhoiL5uXWpKU5T1ER7T>
#

Things i tested:

  • consider DM stacks
  • consider Shadow dance length for single target (multi target is in the above)
  • consider the remaining Shadow Dance Cooldown for Single target
valid robin
#

Nice, is there any gain in holding off building cps or finishing when dance is 4-5 seconds off being ready?

#

Sometimes I find myself finishing at 2-3 seconds left on dance cd which is wasted cdr, and ideally you would save that energy for dance instead

#

But idk if there is any real gain to do this, there is potential for slight energy overcap, but what does the sim say?

#

At the time of dance almost being ready again, you won’t have any lingering stacks

rugged solar
#

i think people don't play subtlety correct atm

#

i saw this a lot in logs

#

it is natural but also something hard to notice

#

a lot of times the delay betweend ances is up to 10+ secs even in better logs

#

there are 3 or even 4+ finishers between

valid robin
#

Well

#

I only talk about when dance is still on cd

#

But very close to being ready

rugged solar
#

oh sry, i might missunderstand

valid robin
#

We are talking like 3-4 cps of cdr wasted because of this, but it’s not a big thing

#

I dont think the sim is considering that either

rugged solar
#

no, the sim would just send a finisher

#

and dance after

#

even if the cdr is higher than needed

valid robin
#

Ye which makes sense if you have 2 charges

rugged solar
#

i mean what could be reasonable

valid robin
#

But ur just wasting cdr with 1 dance

rugged solar
#

is to test if finishing earlier is a win

#

like e.g. you on 4 and it would be enough to cdr dance

#

you finish-dance

valid robin
#

Or just not build or spend

rugged solar
#

instead of building one more

valid robin
#

Maybe

rugged solar
#

i am fairly certain not building is bad

valid robin
#

If played correctly you shouldn’t have a lot of energy

#

Because you have spammed gloomblades

#

At that time

#

So what’s the loss in not building for 3-4 sec

rugged solar
#

energy conditions could work ye

valid robin
#

If you don’t overcap

rugged solar
#

fw uptime, snd uptime(potentially), rupture uptime(potentially), lost gcds

valid robin
#

Ye I mean you have to play better with maintenance

#

If you want to do this

#

Snd is unlikely to drop tho

#

But perhaps rupture

#

It’s just something I have noticed recently when playing, that I am sending some wasted cdr finishers and it feels bad

rugged solar
#

energy levels are rarly rly low

#

but they are low enough to potentially allow energy to not overcap

valid robin
#

Well tea is also used

#

In cases where it might not need to be used

rugged solar
#

yes ofc

#

wdym

valid robin
#

Well if ur going to wait for dance cd

rugged solar
#

instead of tea

valid robin
#

You might not need it

rugged solar
#

just waiting for some secs?

valid robin
#

Tea is quite hard to find use for on st

#

You have to spam ur brain out to drop enough in energy

rugged solar
#

yes

#

thats correct

#

subtlety is fairly agressive with builder use

#

but lets see first

#

if that is actually reasonable

#

seems to not rly change much

#

so, under waht condition do you want to wait?

#

if dance cooldown is low

#

and ?

valid robin
#

You have 0-1 cps

#

It happens when you have like 10 sec left on dance and then do a max cp finisher

#

Now you have a 5 sec cd on dance and have to either spend a bunch of energy on building and overcap your dance cdr or hold

#

Currently it’s just building and wasting cdr

steel jasper
#

It probably depends on the fight timer, because losing 1-2 seconds of cdr each dance doesn't add up to much until you have 30+ dance casts to gain an extra cast with no cdr loss. Using a 7 cp finisher gives 4.9 sec of cdr, so if dance has 1 sec left on CD you waste 3.9 seconds of cdr. If I did my math right, you'd need 15 casts of consistently missing out on 3.9 cdr to miss out on an extra dance charge, and that's about the worst case scenario (assuming dance is sent on CD with no finishers used before it is)

rugged solar
#

the loss is likely from globals

#

the apm go down slightly

#

you might save a little dance cdr

#

but you lose casts of spells

rugged solar
#

time to purge some pins ^^

#

guess its fairly irrelevant

rugged solar
#

even off gcd use or connserving energy for puzzle box isn't a gain

rugged solar
#

alright, pins are back to a sane state

rugged solar
#

ill cross post this, because the discussion started in a diffrent channel

#

Stealthi: but then in this scenario that i showed how would it then not just be better to do a small rupture (that the fits the window until you wanna rupture into dance) over just smacking a 7cp one that gonna get freshed in 5 sec anyway?

ivory kestrel
#

sameu : we can do a dance cd <15 secs or so and begin tackling diff cps of non finality rupture if you want to try that
although maybe <10 or so could be better as that is more or less 2 finishers
might be able to sneak 1 evis more
and not lose pandemic value possibly

rugged solar
#

it might be that you can do a lower cp rupture

#

in this specific case

ivory kestrel
#

yea what i meant is that if you have to renew rupture(non finality) and dance is x secs off cd to be able to do a lower one maybe

rugged solar
#

it atm has a 10 or 12 sec window

#

at least neutral now

#

slightly above target error

#

let me re-run with high iterations

#

so, it seems like fairly irrelevant to optimize further

rugged solar
#

just for context, i think that might got lost a bit
stealthi did point to this sequence asking if it could be further optimized

ivory kestrel
#

and the conclusion is that even if it would be a win to use a lower cp sometimes other times maybe you are unlucky or whatever and lose dance cdr or cant get the finality rupture and dance cdr to line up well so it is neutral

valid robin
#

@rugged solar DK came up with an idea that might be worth trying for builds that don't pick TFD

rugged solar
#

sure

valid robin
#

it's to dance at around 3-4 cp

#

then gloomblade

#

in dance as first gcd

#

(only for ST)

#

similar to tfd but without.. tfd

#

by pre-building cps

rugged solar
#

think this should do what you want

#

^ a bit tighter, but still same result

#

oh wait, thats with tfd

#

hmm

mental dirge
#

fuu

#

both of those sims "test" is playing TFD

rugged solar
#

yes

#

i just realized

#

i did copy the wrong build

#

same result

#

with the more stricter condition

mental dirge
#

its not doing what I had in mind lol

rugged solar
#

hmm i see

#

moment

mental dirge
#

just playing as usual kekman

rugged solar
#

ye

#

its because

#

it falls through

#

easy to fix

#

might make more sense to do 4 or 5

#

not 3 - 5

#

but could end up too strict

turbid pulsar
rugged solar
#

ye, seems so

#

the above does it well (the -0.7 sim)

#

^ but can end up doing 2 gloomblades

mental dirge
#

If raising the restriction to 4/5 causes a 2.7% loss i assume somethings breaking in the sim lol

valid robin
#

I assume it still has the gloomblade at 5 cp in dance condition

#

Which means you might gloomblade too much in dance

turbid pulsar
#

i could def see there be some krangled shit going on

rugged solar
#

it seems to reliably follow the idea

turbid pulsar
#

dont know if relevant but it does use Tea alot in dance

rugged solar
#

that should be a net positive

#

because it would mean more damage during high dps

rugged solar
#

^ it is not worth it to use BS during Lingering on 2 targets

- actions.build=shuriken_storm,if=spell_targets>=2+(buff.lingering_shadow.remains>=6|buff.perforated_veins.up)
+ actions.build=shuriken_storm,if=spell_targets>=2+(talent.gloomblade&buff.lingering_shadow.remains>=4|buff.perforated_veins.up)
turbid pulsar
# rugged solar it seems to reliably follow the idea

on the topic of this one Fuu, the "Test" has 3% lower dance uptime, so it must almost be something krangled with the threshold for combo points it wants to do this, cus Essentially if played correct, its the exact same as just playing With TFD, was thinking if we made the CPS needed from 3-6 maybe it could help it out

it basically lost 1.3 shadow dance casts over the sim duration to this

valid robin
#

Nice find

rugged solar
#

3-6

#

moment

#

i think its more a conequence of introduced delays

turbid pulsar
#

i had an idea for it, would it be possible to make it do something like this.
if at 3-5cps it would do dance+gloom and if at 6-7cps it would do dance+Finisher?

#

so you dont have it hit a threshold where it would just hold dance

rugged solar
#

it would do that by default

#

but doing it up to 7

#

moment

#

basically what the sim above does is to dance-gb on 3,4 and 5

#

and dance-evi on 6

#

highering it to 7 will likely close the gap of the 0.3%

#

^ in general doing gb on 3, 4 and 5 instead of only 4 and 5 seems dps neutral

turbid pulsar
#

the Test one is not doing it tho

#

it just does Dance+strike everytime

#

atleast the bottom one you posted does not do it

turbid pulsar
#

if we could be able to eliminate something like this happening as well it would get some gains

valid robin
#

Should prob limit gloomblade to 1 time per dance

turbid pulsar
#

yeh

rugged solar
#

agree, i think it would make sense to have a option to check if you already used one spell

wicked locust
#

fuu, that lingering find is for just backstab, or also gloomblade?

coarse laurel
valid robin
#

But ur not delaying anything

#

Sure the bot might be

#

But ingame you dont

wicked locust
#

ok thats what i thought, but sometimes my brain just doesn't process the dense lines well haha

coarse laurel
# valid robin Sure the bot might be

Not really sure what this means. Any sort of changes to sequencing will result in some sort of delay if they didn't before. Can't just put extra constraints on something and assume it's always going to align perfectly? A 1.8% difference in benefit seems pretty expected for something like this. This is definitely small enough a player wouldn't even notice doing this themselves.

#

This is always going to be the potential downside of running 1 SD charge compared to 2. We've ran into plenty of this issue with Shiv/Toxic Blade over the years with Assassination. Lots of "logical" things have mostly just been ignored due to single charge constraints vs. just sending on CD.

coarse laurel
#

2% difference is only something like half a second per dance difference over that iteration length, but does potentially add up

valid robin
#

Well what i meant is you would plan for it

rugged solar
#

the pin if you mean that

#

i did check it because somone asked me, and i forgot to check for backstab initially

#

simulationcraft would need to offer me a apl option to check if i used a spell already for a DM stack to avoid use

#

it would allow some more micro optimizations

#

overall it does fairly well play around stacks

wicked locust
#

yea i was just asking because someone made a comment in the general channel that was a little ambiguous

rugged solar
#

basically

#

if you run gloomblade you can maximize with gloomblade after dance down to 30 stacks

#

with backstab this does not hold up sadly

#

^ this pin is also in addition to that with strike-evi-storm-gb in dance on 2 targets

#

i am not sure if backstab is worth it

#

in this specific case, but def. something to look for when exploring it further

wicked locust
#

"backstab: it aint worth it!"

rugged solar
#

well

#

i did also for science test it with "infront of the boss"

#

and it was even worse

#

the "bonus damage from behind" mechanic shows age and idk why it is still a thing

coarse laurel
rugged solar
#

nice

rugged solar
#

Grieftorch:

- actions.cds+=/use_item,name=manic_grieftorch,use_off_gcd=1,if=gcd.remains>gcd.max-0.1,if=!stealthed.all
+ actions.cds+=/use_item,name=manic_grieftorch,use_off_gcd=1,if=gcd.remains>gcd.max-0.1,if=!stealthed.all&(trinket.2.cooldown.remains>0|trinket.1.cooldown.remains>0)

^ quick fix to torch because it is usually not the best first use in double trinket setups

rugged solar
#

not sure if there isn't a more elegant way to check if the other trinket is on cd

coarse laurel
#

Should be able to do something like

backstab,if=variable.gloomblade_condition&talent.danse_macabre&!used_for_danse...

#

Or

if=action.shadowstrike.used_for_danse
On Vanish lines, etc.

rugged solar
#

nice

rugged solar
#

the thing is

#

this is somehow a loss on st

valid robin
#

The sample is real funky in the st sim

#

Its doing strike into a billion off gcd cds and then instant sec tech + cb

#

After that it also uses tornado mid dance

rugged solar
#

changing tornado use to consider dance is +-0

#

however

#

this would still use tornado in dance

#

hmm not sure

#

maybe it needs more agressive changes

valid robin
#

Maybe

rugged solar
#

trying some things atm

#

probably want to use tornado-flag in the opener

#

that sounds reasonable

valid robin
#

Yep

#

Or flag-tornado-dance

#

Like how we played it in SL

past fable
#

The storms that shuriken tornado produces doesnt seem to give the 4p bonus nor do they seem to consume the 2p bonus

rugged solar
#

yes, the tornado ticks don't interact with most things

#

that is the in game behaviour too unless something changed

past fable
#

Huh? Im saying the sim behaviour doesnt seem to line up with the ingame behaviour

#

There is no builder in that sequence i linked, so there should be no way to proc the buff i highlighted (the 4p).

#

So it mustve procced from the tornado in the sim, which is not consistent with the ingame behaviour from my testing

rugged solar
#

oh interessting

#

ill check in game in a moment

past fable
#

See the buff bar in the top right

rugged solar
#

ye

#

probably needs to be corrected in the sim them

past fable
#

It seems to be benefitting from the 2p aswell in the sim

rugged solar
#

mostly a hack but a win is a win

#

XD

#

hmm

#

rupture-dance might have been better there

#

because this delays rupture to 10 secs

#

more hacks but rupture good

#

oh, this is not what i assumed it does

coarse laurel
past fable
#

Seems like it ye, it benefits from it without consuming it

coarse laurel
#

Ok

rugged solar
#

seems like most of the gain is in rupture

coarse laurel
#

I'll look at the Tornado thing today

scenic schooner
#

whats the line change?

rugged solar
#

actions+=/rupture,if=buff.flagellation_buff.up&combo_points>=3&!dot.rupture.ticking&time<=5

#

horrible line

scenic schooner
#

hmm

#

so kinda forces a rup inside flag?

rugged solar
#

it forces rupture as follow up finisher yes

#

it might be worth to re-visit the skip_rupture line for it

coarse laurel
scenic schooner
#

unrelated to ur other test

#

but still a 1.1% gain in st with m+ talents

#

forced nado to be before dance

#

well forced dance to be delayed till u hit nado actually

#

and made it so it uses gloomblade to build durning nado

#

to preserve the premed buff on strike

rugged solar
#

ah nice

scenic schooner
#

this is also ur spreadsheet profile so if u put ur rup hack on top of that its prob even more of a gain

rugged solar
#

the sequence looks nice

scenic schooner
#

and then in dance it should hit gloom twice

#

while u get the cp from nado

#

oh wait no it somehow got 6cp within the first tornado tick

rugged solar
#

ye, idk how tho

scenic schooner
#

did it just get 2 shtech procs in 2 globals? or how did it get the cp?

#

cause there is only one tornado tick there right?

#

wtb tornado ticks and shtech procs on sample sequence lol

rugged solar
#

shb is active

#

so it gets 2 per tornado tick

scenic schooner
#

true im dumb

rugged solar
#

still a big jump from 0-6 in one global

scenic schooner
#

it has 2cp after rup for some reason

#

so it would be 1 crit tornado tick + a shtech proc

valid robin
#

Well if it’s with lust and wft

scenic schooner
#

from 2 to 6

rugged solar
#

^ lust and wft ye

valid robin
#

You can get lucky with sht

#

And get it on 3rd hit back to back

scenic schooner
#

well servers are back up guess simming time is over

#

(started queue at 50)

rugged solar
#

it might be good to wait

#

to get the tornado fix in

scenic schooner
#

i mean

#

even if i wanted to

rugged solar
#

haha ye

rugged solar
#

seems like there was more traffic than anticipated

rugged solar
#

currently playing around with DM conditions

turbid pulsar
#

did you get to trying the BP stuff on specific 2 targets?

rugged solar
#

this sequence on 2 targets seems interessting to try out

#

not yet, but would also fit into that

#

the above is no damage win/loss

turbid pulsar
#

is that with TFD?

rugged solar
#

yes

#

i need to check without too

turbid pulsar
#

yeh

rugged solar
#

but even being 0.1% seems nice

#

and it would fit into the sequence well

#

as it has 2 finisher

#

bp seems to be not worth it

turbid pulsar
#

can you try that BP test with the dungeon build for me

#

and the 2 target build

rugged solar
#

hmm, i think its not doing what i wanted

#

moment

#

one thing btw. as i think this is a bigger question we didn't answer before

#

not uing GB after it created a stack seems damage neutral

#

but i just started toying around with the new danse condition

#

okay... the above bp test is stupid

#

because it has no bp

#

talented

#

this looks promosing

valid robin
#

If you want easy gains you can prob look at the left side

#

Of our tree

#

PV can be optimized, dust and rotten too prob

#

Rotten can get massive increases, it’s still wasting 60% of the cps

rugged solar
#

yes, def. something to look at

valid robin
#

But you might have to get creative with rotten

#

And rethink symbols usage

rugged solar
#

probably

valid robin
#

Like you have to use it post premed and at 0-1 cps

#

Or before premed

#

Atm it doesn’t consider cps at all I think

rugged solar
#

the problem i found last time i looked was that added delays for symbols ended up fairly meh

#

but yes, def. something to optimize

#

just not the thing i atm look at

valid robin
#

Well it’s throwing away 40 cps atm

#

I think even delaying would be worth

#

40 cps is like the entire value of tfd

rugged solar
#

naming is not super telling

#

^^

#

need to check sth

turbid pulsar
#

can u give me tldr, does it seem worth it to BP on 2 targets for a DM stack (currently writing up Raszageth stuff)

rugged solar
#

it seems so yes

#

but i am still testing

#

ye, no hurry

#

as mentioned in the other channel, if there is something specific you or push want to highlight i can add it to wowhead raid tips too

turbid pulsar
#

when im done and ive sent it to push we can share it so you can see if its something you want up as well

rugged solar
#

sure, anything i can add is nice

#

this looks promising

#

(still just testing ability sequencing, so fairly hacky)

#

with diffrent talent setup

#

the latest is with tornado, the one before is without

#

just to show that this is fairly hacky, this would be the result on pure st

rugged solar
#

but lets put this aside and look into pv/rotten then

rugged solar
#

idk why but the dust improvements do show up fairly low with this build

#

i assumed it was due to shb + stiletto

#

so i removed it from the condition

lilac crescent
#

@rugged solar any more info on SnD usage w/ premed

rugged solar
#

haven't tested it

#

but should be easy to test

scenic schooner
#

ended up beeing a gain

#

2 target cleave

#

on M+ talents

#

idea was to not build or finish at lower cp durning nado on 2target cleave

#

condition was a gain but was still not doing what i wanted it to

rugged solar
#

what was the context of this

scenic schooner
#

nothing in particular i was just trying random stuff i wanted to try a while ago while waiting for servers to get back up

#

involving tornado use in general

rugged solar
#

oh ye tornado

#

i remember

scenic schooner
#

if u got some time can u check where i fucked up on this one tho?

#

i was trying to go for it to just finish on low CP durning tornado

#

or not build at all durning it

#

it seems to keep building sometimes tho and idk why

#

actions+=/call_action_list,name=finish,if=buff.shuriken_tornado.up&combo_points>3

#

i added this line to the apl

#

the line is above the call to build

#

but somehow its still building durning nado even if above 3cp

rugged solar
#

@lilac crescent snd refresh seems to be low enough to not matter

#

what i could test is vanish use

#

to consider snd

#

for context:
premeditation does only extend up to 47 secs max, this means that premeditation can be wasted in some cases

#

(47 is the max pandemic of snd 5 cp)

lilac crescent
#

as expected

#

I knew it didn't happen often enough to be more than say 0.2% or so

rugged solar
#

i did try to:

  • not use snd if dance comes up soon
  • not use snd if duration would be to high
    in the above, but it just isn't happening enough
#

to only vanish when you would be able to get full premed benefits

#

also damage neutral

lilac crescent
#

I suspect it's because it happens once over 5m

#

at the most

rugged solar
#

snd just isn't somehting to worry about too much

lilac crescent
#

It's just swapping finishers to different positions essentially

#

the duration doesn't matter

#

instead of evis > full dance > snd it would just become snd > full dance > evis

rugged solar
scenic schooner
#

ok ty fuu feelsloveman

rugged solar
#

but i think ther is more to be done with tornado, just a bit split between what to test first ^^

scenic schooner
#

gonna try it out later, currently prepearing a physics exam...

#

it is not fun

scenic schooner
#

but ye tornado use is rly weird atm

rugged solar
#

ah nice

lilac crescent
#

speaking of subterfuge

#

has it been looked into since talent tree reshuffle

#

instead of long arms

turbid pulsar
#

for the ST build, its a dps loss to take it or atleast was

#

since its completely useless when we Vanish in dance, cuz gloom does to much dmg comming out of dance kekman

rugged solar
#

because in general

#

it was a damage gain to ignore stealth applied from subterfuge

#

if talented into it

lilac crescent
#

over long arms

#

it does seem like it's just a useless effect but was just checking if it was because it took up a talent

rugged solar
#

i did not change the apl to ignore subterfuge

#

because i think that nobody would play it like that if talenting into it

#

but it is popular in pvp

lilac crescent
#

makes sense different game really

rugged solar
#

subterfuge is utility for them

#

so yes

wicked locust
#

i woke up in the middle of the night thinking about dust. i know it doesn't play nicely with the apl so i figured maybe talking about the theoretical way it plays would be a potential starting point for you better apl folks to hack at when you get bored.

i figure that the actual play of the setup would be based around a 1-min cd cycle of vanish/sepsis/flag. then you'd aim for getting a dance in both the building and finishing side of the flag window. with rotten and tfd you've got a pretty set dance window of [dance+symbols+storm]->[evis]->[gloom]->[sec tec]->[strike]->[evis], which should be a 5-6-5 minimum. then you need (i think) to spend another 10cp and use a vanish in the next 6-12 seconds to get dance/sec tec back up for the flag buff window.

so off-hand i'm not sure how much cdr you can expect from stiletto for sblades over 1.5 minutes. if it gets under 1 minute that'd probably be really strong. I've got a feeling it'll be over a minute, but maybe interactions with PV/rotten/dance could be strong enough to delay the 1-minute window slightly.

then i think you'd maybe be able to fit another dance between those 1 minute windows.

sepsis/vanish stuff might be able to be lined up to juice the gloom inside of dance.

now that i've typed this all up, my mind is wondering if i'm thinking about the extra vanish charge incorrectly and this all falls apart after 3-4 cycles lol.

valid robin
#

Its not a rolling cd no

#

The 1 min vanish timer breaks apart after a few minutes

wicked locust
#

yea i got done typing and then my brain went "wait not its probably not like that". i guess the window would be a 6-6-5 minimum cp too, since storm would always crit

#

so anything over like 4 minutes it'd all go to shit on

#

i'm not sure if theres a better way to try to leverage dust, since you have to give up so much to get it

rugged solar
#

I am mostly surprised that the apl change for dust seemed to drop from 1-2% gains to neutral ones

#

They are straight forward checking for cooldown to be not ready

valid robin
#

Well there is a lot of cds

#

And it probably ends up fucking it up

#

And not use it

#

Or something

rugged solar
#

possible

valid robin
#

But yea i mean we already have an example of how you can play it

#

Someone just have to implement it in simc to test it out

rugged solar
#
actions.stealth_cds+=/vanish,if=
!cooldown.sepsis.ready&
!cooldown.shadow_blades.ready&
!cooldown.shadow_dance.ready&
!buff.shadow_dance.up
&combo_points.deficit>3&
!buff.the_rotten.up```
#

it checks for sepsis, shb and dance

valid robin
#

It seems like its a bit of a hassle to fix, but someone just has to put time into it and accept that its gonna take many iterations to fix

rugged solar
#

no, i mean the line i linked seemed to work fine

#

thats why i was wondering

#

hmm

#

i mean its a increase

#

but still fairly irrelevant

#

(this is just the above line, with minimal adjustments)

rugged solar
#

how exactly is it that you want to use vanish?

#

(experimental) Dust improvement (with talent checks):

+ actions.stealth_cds+=/vanish,if=(!cooldown.flagellation.ready|!talent.flagellation)&(!cooldown.secret_technique.ready|!talent.secret_technique)&(!cooldown.sepsis.ready|!talent.sepsis)&(!cooldown.shadow_blades.ready|!talent.shadow_blades)&!cooldown.shadow_dance.ready&!buff.shadow_dance.up&combo_points<=1&!buff.the_rotten.up|fight_remains<30

https://www.raidbots.com/simbot/report/jmmLKerj1QUjFSukMEYDLc
best case: 0.2%, but usually neutral

rugged solar
#

^ i think thsi is a solid baseline to go from

rugged solar
#

btw i found a higher simming combo than your one with dust @valid robin

valid robin
#

Really?

#

Nice

rugged solar
#

this one

valid robin
#

Hmm

rugged solar
#

but its not intuitive because it does not take sepsis or flag

#

83.2k

#

vs. 82k

valid robin
#

I assume this sims better because its not as complex

rugged solar
#

but the above

valid robin
#

And actually functions in simc

#

Not because its better ingame

rugged solar
#

is a straight up loss for it

#

so i need to look into that

#

maybe it is worth to add seceet to it

#

for that combo

#

or keep the use during dance

#

from dm

valid robin
#

Yes

rugged solar
#

lets see

valid robin
#

But

#

If it sims 83k its prob decent

#

Ingame

#

Since rotten is wasted

#

Idk how much % dps is missing from rotten or PV

rugged solar
#

you could prbably skip rotten too

valid robin
#

Wouldnt surprise me if its like 3% tho

rugged solar
#

and just go with tfd

valid robin
#

Rotten is better than tfd

rugged solar
#

lets compare

valid robin
#

Rotten is literally just a better tfd

rugged solar
#

well tfd also sims worse

#

guess so

valid robin
#

Rotten is actually not a bad talent

#

Dps % wise

rugged solar
#

i still think the main issue is pv

valid robin
#

Wouldnt surprise me if its one of the stronger capstones if it got full use of the cps

rugged solar
#

i could see the left side being competitive

valid robin
#

Well PV is bad yes, but the sim is also playing pv very bad

rugged solar
#

if you had a highway from sepsis

valid robin
#

Remember when pv was played like shit in SL

#

Its value increased by like 100%

#

When fixed

rugged solar
#

well yes

valid robin
#

We cant say anything about the strength of the left side until its actually like played properly

rugged solar
#

but there was a diffrence

#

because pv was simply not utialized/wasted

#

and pv stacks had a higher impact

#

you would only need to wait for 1 gcd to utialize it

#

if you now generate 5 stacks

#

you need to wait 7+ gcds between dances

#

to get rid of them

valid robin
#

Ye its not missing half its value this time. But its still wasting its potential and not by not consuming or overcapping, its just not shadowstriking enough

#

Cus of the gloom in dance apl

rugged solar
#

yes

#

i need to double check that def. esp. with danse

valid robin
#

Also you can prob play around the pv stacks at high lingering

rugged solar
#

because when i added that, danse was not relevant

#

gb in dance might be perfectly fine

#

because gb damage value is already on average higher

#

and esp. if you have max stacks, might be worth to trade off 1 gcd

#

but also i think it is atm not smart with it

#

so likely that ther eis improvement

#

possible

#

but ill check dm first, dust with dm might be interessting

valid robin
#

If ur ever bored there is probably a ton of stuff to do with the left side apl

#

Goes for anyone with an interest in sub

rugged solar
#

i think there is in general a lot of yack shving you could do in the apl

valid robin
#

YOU can be the guy that creates the next meta build. Contribute today and be forever known as the maker of builds

#

^ ad

rugged solar
#

haha, i love how much you like the left side

valid robin
#

No i just want my sim to be some kind of truth

#

I dont like the talents

#

I just want to be able to sim it and say, ”yes this is how weak/strong this is”

#

Instead of thinking ”this doesnt work at all”

rugged solar
#

fair, i hink you overestimate how much their is to gain

#

but its just a waste of time to discuss that, better to focus on trying what actually works ^^

#

wonder how dm + rotten is vs db + rotten

#

okay, its not great

#

db + rotten seems like the sweet spot, probably also due to lingering

#

well also because you would not skill the dance damage amp, what is a big loss

rugged solar
#

the line change is also neutral on regular builds

#

so something i can easy add

rugged solar
#

alright, back to looking at what the sim does, was sidetracked in dms and #subtlety

#

might make more sense to put vanish a bit lower in the cd prio

#

this shouldn't happen

#

hmm

#

i think its not that but just using it again after vanish because it gets it ready

#

ah ye^^

#

another dust build that sims even higher

#

86k

#

and it makes sense

#

rotten and sepsis both are weaker capstones

#

the strong 3 are dm, db and flag

rugged solar
#

(test = the change)

#

Dust improvement:

- actions.cds+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2
+ actions.cds+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2&(cooldown.secret_technique.remains>=30|!talent.secret_technique)```^ up to 0.6% dps gain depending on build, seems to be damage neutral in worse case
rugged solar
#

^ would be a sim with the flag build above and the above change

#

all flag windows seemed to ended up with high stacks, however 2nd one seemed a bit worse with 20 satcks only

rugged solar
#

but quite a bit of a loss on some builds

#

basically, it seems to be neutral in DM builds, but else a loss

#

to "not use GB/BS"

#

so, what would be ideas to improve the use

#

what would be interesting to test for pv?

wicked locust
#

weirdly subterfuge might be good for some setups

#

my weird rambly thoughts from last night, i was thinking that with subterfuge and sepsis you'd be able to get like 4 stacks of pv for the second dance window

#

extending out the vanish charges is probably neutral, but might be a more natural style or line up with fights

#

maybe aim to have symbols of death active for the full dance window if you have vanish charges?

#

basically if we're going to reset something completely via dust it should be a gain to line it up with something right?

rugged solar
#

probably if you wouldn't play with dust

coarse laurel
rugged solar
#

i haven't looked at sepsis much after the change

coarse laurel
#

Having done a lot of work to attempt to make it work for both Outlaw and Assassination as well, I'm pretty unconvinced there is any real place for it in any build right now. It just kinda sucks.

rugged solar
#

does the sepsis buff count for stealthed.all?

#

the new one

coarse laurel
#

The stealth change to Sepsis doesn't do much to really help it be good

#

It does

rugged solar
#

nice

coarse laurel
#

Technically it should be best for Assassination given the special coding that allows Sepsis to trigger Improved Garrote

#

And even then it's miles behind SBS and not even remotely viable

rugged solar
#

sepsis seems to be tuned too low for its cooldown

#

compared to other capstone talents

coarse laurel
#

The best way to think about the Sepsis buff is that it's 1/3rd of a Subterfuge and Subterfuge is not even a DPS gain most of the time as a talent, let alone competitive so "doubling" the value to 2/3rds of a Subterfuge is still pretty pointless.

rugged solar
#

i assume it could work for subtlety with a highway to dust

coarse laurel
#

Only way Sepsis will be viable is to giga buff the damage

rugged solar
#

half the cooldown double the damage

#

sounds good

coarse laurel
#

Considering the Imp Garrote interaction and Assassination being able to Shiv the bonk

#

And it's still not good

#

Sub and Outlaw have no chance

rugged solar
#

one of the streamers had the idea

#

to vanish with ma for the last sepsis tick

#

but that seems not that great

#

due to improved garrote

coarse laurel
#

That would be kinda pointless because you'd double up the Imp Garrote benefit and basically lose the extra Imp Garrote proc anyway

#

Gain more damage just letting it go off randomly

#

Sepsis just isn't good

wicked locust
#

if pv didn't suck it'd help sepsis out a lot for sub as well

coarse laurel
#

Even Assassination syncing with Shiv it only ends up contributing like 1.5-2% damage which is quite weak for an active and gets beat by many passive talents

rugged solar
#

the pv design change rly is the biggets issue

#

but push did say he has good ideas how to improve pv

#

in the sim

#

i see some problems

wicked locust
#

make consuming pv stacks give vanish cdr 🙂

rugged solar
#

because vanish is fairly bound to dance with dm

coarse laurel
#

Also Sepsis doesn't scale with multiple targets so it's not only fairly low benefit in ST but just doesn't apply to many situations either

rugged solar
#

so the entire concept of "don't use vanish when on high pv stacks"

#

would introduce a lot of delays

#

and value from vanish

#

similar, you can't just delay dance for 7 gcds

#

just to get rid of pv stacks

#

i mean you can, it just does not sound super good

#

there was the idea to "make pv play better around lingering"

coarse laurel
#

And for Sub I assume that a raw Shadowstrike without any of the Stealth benefits is of pretty limited value

rugged solar
#

but isn't that the case by default?

#

ignoring subterfuge was a damage gain

#

last time i looked

#

i just never added that to the apl, because realistically

#

if somone specs into it, he will play around it

#

and it is easy to see if its a win/loss with talent combos

#

but maybe thats the wrong approach

wicked locust
#

since the apl ignores rotten, does factoring it in change optimal dance stuff? like if you have rotten and tfd and a 6 sec dance you really only have room for 1 ss if you don't want to waste a bunch of cp

coarse laurel
#

My problem with the left side of the Sub tree is there are elements that seem to really fight each other a fair bit, which not only makes oddities in the APL (as mentioned) but I just don't think they all have elegant solutions because they all have some sort of inefficiency built in.

wicked locust
#

then you could leverage vanish/subterfuge to build pv stacks for post-dance lingering spam

coarse laurel
#

Like you have this odd dynamic where you have some things that want you to build around using CDs going into Dance (TfD+Inevitability+Imp Shadow Dance) and others that skew in wanting to use your Symbols after dance (Lingering+Rotten+PV)

#

Just feels like some innate inefficiency being built in here where you kinda just have to pick one angle to ignore

rugged solar
#

we have 2 or 3 lines that care about rotten and adjust

#

there is just maybe room for improvements due to cp loss

coarse laurel
#

If you optimize for Symbols into TfD+Inevitiability to get longer Symbols and amp your Dances you will lose out on getting juiced Gloomblades coming out of Dance with Rotten+Lingering. If you optimize using Symbols on your last finisher into big Rotten you'll get fewer extensions from Inevitability and lose Dance damage in general.

#

There's not a lot of cohesiveness in what the left side "wants" you to do

#

Have always felt it was just a grab bag of mechanics that didn't really work together all that well

#

Especially with only single dance

wicked locust
#

ah i see, its just using the rotten proc outside of dance

#

even if i give the build tfd 😄

#

unless i mucked this up,, which is possible

rugged solar
#

the first dance would want you to use it outside

#

or what do i miss here

#

you would else overcap with dance-gb

#

similar to premed

wicked locust
#

its doing a "use the rotten buffed gloomblade" directly into dance and sec tec finish

rugged solar
#

moment

wicked locust
#

which is way overblowing cp

rugged solar
#

yes

#

i thought you mean the other way around

#

let me check

#

moment

wicked locust
#

you have to adjust the build around a bit to get tfd, something like this i'd assume BUQAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAhgkkICJSSSQCBJJJIhikElIkUiQSSkWAAAA

rugged solar
#

this sim should have it

#

but that seems fine

#

oh, i see what you mean

#

if it uses symbols in dance

#

it can still apply the 4/5 cp gloomblade rule

#

so thats not good ye

#

but easy to filter

wicked locust
#

i had this pop up in one i was messing with, seeing if i can get it to occur again (maybe its a flag thing?)

coarse laurel
#

Oh, and before I forget to mention--the other issue with Sepsis is also the points allocation, right?
I mean given it's below the break, you need all 10 points to get 2 capstones.

#

That effectively makes Sepsis the same placement as a capstone indirectly

#

But it's not even remotely competitive with capstone tuning

rugged solar
#

sepsis would work

#

if you could skip pv

#

so you could lower your point count to left capstone by 1

#

and get flag + sepsis + dust + rotten

coarse laurel
#

For comparison, Dark Brew is worth like 7.5%... lol and DM is worth about 9%.
And even if Sepsis didn't have to compete with capstones (which it does) it would have to compete with Flag, which is also like 6%...
Sepsis literally isn't even in the picture at all given its position in the tree at 1.8% or whatever it is. Even if it somehow got APL gains, it's not viable.

rugged solar
#

this is why the build would have flag and sepsis

#

if you could skip pv

#

i am fairly certain that pv is fairly low impact

#

for a 2 pointer

coarse laurel
#

Allowing that kind of bypass is probably the only way it would work, but that maybe causes some other problems with the tree to convert a 2 point req into 1 point idk..

rugged solar
#

it would cause problems if the left side had stronger options

#

because you would essentially gain a additional capstone

coarse laurel
#

But yeah I mean at least for Outlaw Sepsis is on the golden path instead of to the side like Assassination/Sub

#

And it still doesn't beat Blade Rush or Ghostly Strike

#

And neither one of those are particularly strong

wicked locust
#

thats still doing gloom->dance on the opener 😄

rugged solar
#

yes

#

is that bad?

#

what would you like it to do

wicked locust
#

doesn't that just waste 4cp from tfd? since rotten gens you uh, 6cp? (4+gloom+seal)

rugged solar
#

wait for gb till after dance

coarse laurel
#

Sepsis is like 2% behind Ghostly Strike for Outlaw... would literally have to double the damage it does it for it to be viable.

#

Which wouldn't even be enough for the Sub tree

rugged solar
#

oh

#

i see what you mean

wicked locust
#

i'd expect it to either finish before dance, or to use a gloom inside of dance, something like [dance+storm]->[evis]->[strike]->[evis]->[gloom]->[evis]

rugged solar
#

ye, might be worth to just use gloom in dance

#

because you waste so much cdr else

coarse laurel
#

(Also, is it just me or should TfD have gone back to the original pre-nerf TfD Azerite design instead of inheriting a simplified version of the nerfed version? :P)

rugged solar
#

wdym

wicked locust
#

i was originally thinking about this in terms of the dust rotation being rough "flag -> dance in flag builder window -> vanish stuff -> dance in flag finisher window", though this does break down after 3 or 4 minutes

rugged solar
#

i mean, personally i think tfd would be interessting in the original form

coarse laurel
rugged solar
#

less cp overhead and critical strike

#

oh first iteration had no cp to it

coarse laurel
#

Before it was decided that Sub was "OP" and had to be nerfed into oblivion

rugged solar
#

yes

#

i mean sub in bfa was sadly all "its too op"

#

got 2 nerfs iirc before mythic opening to traits

#

and then zul

coarse laurel
#

CP thing only came in with the Haste buff yeah

rugged solar
#

and even more nerfs

coarse laurel
#

Original was purely just crit

valid robin
#

I think we can leave sepsis where it should be

#

In the bin

coarse laurel
#

All the homies hate Sepsis 😄

#

If it's one thing every Rogue spec can agree on, it's that they don't want to take Sepsis

#

So it's all good

valid robin
#

Its a boring spell for anyone but assa and for assa its just bad

#

Buff the shit out of it and remove it from sub/outlaw

#

Reduce its cd to 30 seconds

rugged solar
#

45 sec

#

exsang 2.0

valid robin
#

Makes the clip effect do something at least

#

Reset cd

#

Instead of this garbage it does atm

wicked locust
#

the problem with sepsis is that its designed for something rogues aren't designed around anymore, or have better access to.

coarse laurel
#

Or like if they had gone back to the original design of it triggering a real Vanish (and thus having a lot of interactions)

#

Maybe it would be okaaaay?

wicked locust
#

like every spec sort of has gotten ways of getting their stealth ability uses more, because the specs need to in order to not be bleh

coarse laurel
#

Only reason Sephuz bailed on that back when is that it would have forced a lot of specific Covenant+Legendary meta issues, which is not really a factor in the talent tree now

valid robin
#

I mean its doing what, 3-4% dmg for sub + the shadowblades dmg

#

5%?

rugged solar
#

vanish is fairly weak

#

for subtelty

#

outlaw has cto

valid robin
#

Not if you have the left side

rugged solar
#

and assassination has improved garrote

valid robin
#

And it triggers dust

rugged solar
#

oh ye dust

valid robin
#

Maybe realz thinks it triggers dust

#

Idk

wicked locust
#

we're just coming up on more ways to completely hinge left side on dust 😄

coarse laurel
#

Sepsis Vanish did trigger Dust back in the day in SL beta

valid robin
#

I mean dust is the only somewhat interesting talent on that side

coarse laurel
#

Was part of the original balancing concern

#

In the context of this tree layout, it'd be more of a selling point than a concern though

#

Since they are close together

valid robin
#

Would make sense

wicked locust
#

yea i don't dislike dust, its just funny that its like the strongest part of that section by far

valid robin
#

Well i think rotten is numerically strong

#

If it works

#

Since we have so many symbols resets

#

A lot of cps

wicked locust
#

yea make sepsis trigger dust, make pv give vanish cdr, and change swift death to something else, like +crit dmg

ivory kestrel
#

sepsis could do many things but as it is rn is super boring i agree it does not even proc premed

#

but for some reason it does proc improv garrote etc

#

and rotten does feel good because it+premed make dance so reliable

#

big problem is Symbols is 30 secs cd so it gets clipped a lot maybe get a second charge of it or make it 1 min cd but stronger

#

and just idk what to do with PV

#

this vers is strictly worse than SL one too

wicked locust
#

having pv stacks reduce vanish cd by 5 seconds on use or something would be something, dunno if it'd be enough but maybe. if you could get vanish down to 1min cd via stuff then you could set up a sustaining 2-dance window around flag at a 1 min cd

rugged solar
# rugged solar **Grieftorch:** ```diff - actions.cds+=/use_item,name=manic_grieftorch,use_off_g...

seems like this is only relevant for boon, so this is probably a simple implementation:

- actions.cds+=/use_item,name=manic_grieftorch,use_off_gcd=1,if=gcd.remains>gcd.max-0.1,if=!stealthed.all
+ actions.cds+=/use_item,name=manic_grieftorch,use_off_gcd=1,if=!stealthed.all&(!raid_event.adds.up|!equipped.stormeaters_boon)```sim: <https://www.raidbots.com/simbot/report/nYk6nXP1eZDbKRjiUBGxNg>
(~2.4% win for dungeon slice only)
twilit moat
rugged solar
#

because it uses it off gcd

#

you see it if you look at the timing

#

grieftorch use is off gcd, basically used with strike

twilit moat
#

SoD is also off GCD, doesnt simc try to use it between GCDs by default then?

rugged solar
#

yes

#

but symbols use in the apl is not off gcd

#

if you look at the apl condition:

twilit moat
#

ok ty

rugged solar
#

most abilities don't make sense to be used off gcd

#

the off gcd use of torch makes sense because you reduce the downtime

wicked locust
#

would it be a gain on st to check fw time remaining and use the imp shuriken storm buff just before dance? so either storm-finish-dance or finish-storm-dance, since if you don't get a refresh right before dance or on your 1 gloom usage it might be falling off before the end

valid robin
#

It already does that

#

Sort of

#

Thats why its storming sometimes on st

wicked locust
#

yea i know it checks for if its up at all, just didn't know if it was looking at the duration of the buff or not

valid robin
#

Ye i might just be checking for debuff uptime

#

But i think it will be too strict to put a timer on the debuff

#

That would prob mean a lot more shuriken storms

#

And i am unsure if thats good, could be worth looking into tho

wicked locust
#

yea i was just thinking that losing a no-lingering gloom to a storm is probably less of a loss than having fw drop for the last global or two of dance. probably depends a lot on your crit rate too

rugged solar
#

it only checks if FW is up

#

and would storm if its not up

#

possible that it is marginally better to check the uptime

rugged solar
#

Experimental change:
I think my first iteration for the Double Dance apl update is finished.
I threw quite a lot of talent combinations on it and it seemed to hold up (on 1-8 targets).

#below the regular snd_condition
+ actions+=/variable,name=snd_condition,value=dot.rupture.ticking,if=(!talent.shadow_dance&talent.danse_macabre|spell_targets.shuriken_storm<3)
#dance line too long?
- actions.stealth_cds+=/shadow_dance,if=(variable.shd_combo_points&(buff.symbols_of_death.remains>=(2.2-talent.flagellation.enabled)|variable.shd_threshold)|buff.flagellation.up|buff.flagellation_persist.remains>=6|spell_targets.shuriken_storm>=4&cooldown.symbols_of_death.remains>10)&!buff.the_rotten.up
+ actions.stealth_cds+=/shadow_dance,if=(variable.shd_combo_points&(variable.shd_threshold)|cooldown.secret_technique.remains<=9*(spell_targets.shuriken_storm<4)&talent.shadow_dance&(talent.danse_macabre|spell_targets.shuriken_storm<=3)|buff.flagellation.up|buff.flagellation_persist.remains>=6|spell_targets.shuriken_storm>=4&cooldown.symbols_of_death.remains>10)&!buff.the_rotten.up
```damage gain between 0.3%-7% depending on target count and talent combination.

To set expectations right, single target damage for single dance is unchanged, and if there are small increases they only happen with more targets (but don't set high expectations)
Double dance looks at 1-3% single target increases that hold up with multiple targets usually, non DM builds are slightly less consistent.

note: follow up change from:
[#1065728795455266888 message](/guild/124169701881282560/channel/1065728795455266888/)
#

just to hype it up unnecessary, the highest increase i found with the change:

#

just to make this clear, even tho i think people in here likely expected this

#

double dance is not beating single dance ^^

rugged solar
rugged solar
#

Two Target improvements for DM:

-     actions.finish=variable,name=premed_snd_condition,value=talent.premeditation.enabled&spell_targets.shuriken_storm<5
+ actions.finish=variable,name=secret_condition,value=buff.shadow_dance.up&(buff.danse_macabre.stack>=3|!talent.danse_macabre)&(!buff.premeditation.up|spell_targets.shuriken_storm>=7)
+ actions.finish+=/variable,name=premed_snd_condition,value=talent.premeditation.enabled&spell_targets.shuriken_storm<5
- actions.finish+=/cold_blood,if=buff.shadow_dance.up&(buff.danse_macabre.stack>=3|!talent.danse_macabre)&cooldown.secret_technique.ready
+ actions.finish+=/cold_blood,if=variable.secret_condition&cooldown.secret_technique.ready
- actions.finish+=/secret_technique,if=buff.shadow_dance.up&(buff.danse_macabre.stack>=3|!talent.danse_macabre)&(!talent.cold_blood|cooldown.cold_blood.remains>buff.shadow_dance.remains-2)
+ actions.finish+=/secret_technique,if=variable.secret_condition&(!talent.cold_blood|cooldown.cold_blood.remains>buff.shadow_dance.remains-2)
- actions.finish+=/black_powder,if=!variable.priority_rotation&spell_targets>=3
+ actions.finish+=/black_powder,if=!variable.priority_rotation&spell_targets>=3|!used_for_danse&buff.shadow_dance.up&spell_targets.shuriken_storm>=2&talent.danse_macabre
-     actions.stealthed+=/shuriken_storm,if=variable.gloomblade_condition&buff.silent_storm.up&!debuff.find_weakness.remains&talent.improved_shuriken_storm.enabled
+ actions.stealthed+=/shuriken_storm,if=variable.gloomblade_condition&buff.silent_storm.up&!debuff.find_weakness.remains&talent.improved_shuriken_storm.enabled|combo_points<=1&!used_for_danse&spell_targets.shuriken_storm>=2&talent.danse_macabre
- actions.stealthed+=/gloomblade,if=variable.gloomblade_condition
+ actions.stealthed+=/gloomblade,if=variable.gloomblade_condition&!used_for_danse
``` Better sequencing for two targets.
link: <https://www.raidbots.com/simbot/report/h6JPefD2Sv8UMtC5pp45yd>
valid robin
#

Seems low overall, didnt we have a 107k dps 2t sim a while back?

rugged solar
#

the talent combo used for the sim was with tfd

#

and it is 106 -> 107k with two target talents

#

(if we are exact more like 108k)

#

108k however is still quite a bit off from assassination/outlaw on 120k

#

when cross comparing between rogue specs

rugged solar
#

small test, lowers overflow to ~25%

rugged solar
#

testing some more, it seems to be not a gain to use GB as the "consumer" of rotten

#

also some funny ideas like always lining it up with dance are not exactly great

#

^ to go a bit into why this is, the reason is dust

#

symbols is a static cooldown and dance a dynamic one

#

and even tho it seems natural to line things up, there will be a certain distance between the cooldowns that can lead to less casts or delays

ivory kestrel
#

usually what i do is with TFD you get to 2-3 then dance+SoD finish backstab finish strike finish

#

maybe trying to get the backstab/gloom with PV+SoD inside dance works better

#

as you will have 1 of 2 dances with a vanish before every min

#

and on opener

rugged solar
rugged solar
#

i think I covered most things from my pin a while ago now, ofc does not mean individual things can have more room for optimization
referece: #1065728795455266888 message

rugged solar
#

rotten improvments:
Gain: 0.2 - 0.5% dps for dust setups.

+ actions.cds=variable,name=rotten_condition,value=!buff.premeditation.up&spell_targets.shuriken_storm=1|!talent.the_rotten|spell_targets.shuriken_storm>1
#replaces symbols condition
+ actions.cds+=/symbols_of_death,if=variable.rotten_condition&variable.snd_condition&(!talent.flagellation&(combo_points<=1|!talent.the_rotten)|cooldown.flagellation.remains>10|cooldown.flagellation.up&combo_points>=5)
+ actions.stealth_cds+=/variable,name=rotten_threshold,value=!buff.the_rotten.up|spell_targets.shuriken_storm>1
#replaces dance conditions
+ actions.stealth_cds+=/shadow_dance,if=(variable.shd_combo_points&(variable.shd_threshold)|cooldown.secret_technique.remains<=9*(spell_targets.shuriken_storm<4)&talent.shadow_dance&(talent.danse_macabre|spell_targets.shuriken_storm<=3)|buff.flagellation.up|buff.flagellation_persist.remains>=6|spell_targets.shuriken_storm>=4&cooldown.symbols_of_death.remains>10)&variable.rotten_threshold
+ actions.stealth_cds+=/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&variable.rotten_threshold
```<https://www.raidbots.com/simbot/report/4GRPzyMUm2DbPPhSmTqkFX>
rugged solar
#

just to mention this, i need to re-iterate over some of the pinned apl changes given that it seems that some could cause conflicts in dungeon slice (lower dps)

jade garnet
#

any reason why
cooldown.vanish.charges
is returning 1 when i have two charges remaining
and returning 0 when i have either 0 or 1 charges remaining?

#

just to clarify, trying to modify apl within hekili

lilac crescent
#

Just use HeroRotation

twilit moat
#

I did some test sims with a small adjustment in the sub APL and it gave some tiny mini gains:
current: 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.remains>=4&!buff.thistle_tea.up&spell_targets.shuriken_storm>=3|boss&!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)

new:
actions.cds+=/thistle_tea,if=cooldown.symbols_of_death.remains>=3&!buff.thistle_tea.up&(energy.max-(energy+gcd.remains*10)>=100|cooldown.thistle_tea.charges_fractional>=2.75&buff.shadow_dance.up)|buff.shadow_dance.remains>=4&!buff.thistle_tea.up&spell_targets.shuriken_storm>=3|boss&!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)

diff is
energy.deficit>=100
vs
energy.max-(energy+gcd.remains*10)>=100

sims
ST no buffs
default APL: https://www.raidbots.com/simbot/report/hfNNEFFnngCLiPpx7xCjpA
custom: https://www.raidbots.com/simbot/report/cCa4JF7Vo7ZdKiqp2qAy3M

  • ~40 DPS

ST all buffs
default APL: https://www.raidbots.com/simbot/report/vnpr1TtSm7NCA2Qiw2EeTN
custom: https://www.raidbots.com/simbot/report/3FtXa4iRFTBjg73iow9oVw

  • ~20 DPS

dungeon build ST
default APL: https://www.raidbots.com/simbot/report/rvdUm8HqgnmSM4j2VWRQW8
custom: https://www.raidbots.com/simbot/report/1NAMsS8SLreWs4ZPsvGgkP

  • ~20 DPS

dungeon build DSlice no buffs
default APL: https://www.raidbots.com/simbot/report/ukufMXVLcSm4NkN82LkKCS
custom: https://www.raidbots.com/simbot/report/hK2qrjcdKTv2TLLrkoSxvv

  • ~20 DPS

dungeon build DSlice + some buffs (BL, Food, Rune, Flask)
default APL: https://www.raidbots.com/simbot/report/ducerkR7uaJdmje3dy53ev
custom: https://www.raidbots.com/simbot/report/4ChSc9epcytNZxV1nSiL67
= ~no diff

#

please test with your profiles

twilit moat
#

did some more patchwerk 3 boss and 5 boss sims (not sure how much sense that makes) and the custom APL always resulted in some gains except for here:
Dungeon build Patchwerk 3 Boss + some buffs (BL, Food, Rune, Flask)
default APL: https://www.raidbots.com/simbot/report/ojRY6MSvuvpGRC839hxaRx
custom: https://www.raidbots.com/simbot/report/oMTHzS3HLj7K1bYvnTcj49

  • ~40 DPS
rugged solar
#

it might make sense to use
energy.regen_combined
so something like:
energy.deficit>=(100+gcd.remains*energy.regen_combined)
or if we stick to your calculation:
energy.max-(energy+gcd.remains*energy.regen_combined)>=100

#

there is also no problem with this change. The energy overflow is however very low:

#

vs. with the change

#

new = your line
new 2 / 3 = the suggestions from above

#

your simulations show more gain/loss because of the default target error.
you can use iterations=100000 to lower it.

twilit moat
#

Alright, ty for looking into it :)

rugged solar
#

still a nice idea, so rly well done 👍

wicked locust
#

does changing the energy usage check from -100 to -110 shift the energy overflow? I know most of whats there is from the initial tea use

rugged solar
#

the energy overflow does not change

#

because the overflow is mostly from the first use

#

the line above basically does something similar to what you suggested

fickle lava
#

you can probably reduce the overflow if you check for finisher because of energy being -100 ish and then using tea into finisher, so maybe you want to separate the two if you want to check if the above idea works better, but most likely won't matter anyway

rugged solar
#

you could check combo points too

#

that idea seems to be not too bad

fickle lava
#

nice

wicked locust
#

good call, thats probably where the energy overflow was coming from

rugged solar
#

the change: ```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.remains>=4&!buff.thistle_tea.up&spell_targets.shuriken_storm>=3|!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)
  • actions.cds+=/thistle_tea,if=cooldown.symbols_of_death.remains>=3&!buff.thistle_tea.up&(energy.deficit>=100&combo_points.deficit>=2|cooldown.thistle_tea.charges_fractional>=2.75&buff.shadow_dance.up)|buff.shadow_dance.remains>=4&!buff.thistle_tea.up&spell_targets.shuriken_storm>=3|!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)```
#

&combo_points.deficit>=2 is what is added

#

but also, the overflow is not from tea

#

it is similar to the symbols condition just redcing overflow around the tea use

#

as you see

#

tea by iteself didn't get betetr/worse

#

hmm

#

even a bit worse

#

because of the introduced delays

wicked locust
#

hrmm, that 40 feels like a curse on us

rugged solar
#

we lose basically 100

#

in the opener

#

i think some of the energy loss is however from

#

!buff.thistle_tea.up&fight_remains<=(6*cooldown.thistle_tea.charges)

#

basically using tea at the end of the fight

wicked locust
#

oh, drink a tea at the end. didn't know that was there heh

rugged solar
#

oh yes, even tho its not in the rotation pages

#

most cooldowns have a similar condition

#

e.g. shadow dance

#

it is common sense mostly when playing it

fickle lava
#

maybe, but also by not using it before a finisher you probably get to a situation where your second tea will be cast because its gonna be close to hitting 3 charges, especially with lust at the start

rugged solar
#

btw. the above was my 2nd try

#

this was my first

#

by trying to only use tea on 0 and 1 cp for energy

wicked locust
#

lost ~30 less from basic energy regen looks like

rugged solar
#

i mean, i take the 0.2% gain, i will need to check if there is a conflict somewhere but overall is a small change

rugged solar
#

ill document this too, small gain for priority_rotation:

- actions.stealthed+=/shuriken_storm,if=spell_targets>=3+buff.the_rotten.up&(!buff.premeditation.up|spell_targets>=7)
+ actions.stealthed+=/shuriken_storm,if=spell_targets>=3+buff.the_rotten.up&(!buff.premeditation.up|spell_targets>=7&!variable.priority_rotation)```
rugged solar
#

(experimental) also a small gain by only using one vanish per dance on two targets:

- actions.cds+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2&(cooldown.secret_technique.remains>=30|!talent.secret_technique)
+ actions.cds+=/vanish,if=buff.danse_macabre.stack>3&combo_points<=2&(cooldown.secret_technique.remains>=50|!talent.secret_technique)```
rugged solar
#

sry, didn't mean to spam

#

just want to keep the apl tests documented to not lose them

#

the last one is mostly for two targets but does not seem to rly impact single target

lilac crescent
rugged solar
rugged solar
#

seems like i might found a fix to

#

need to test more but already promising

rugged solar
#

poggers test beating the reference

#

(its nothing new, just me trying to integrate all changes in the apl and eliminate bad cases)

rugged solar
#

seems like neutral poggers

#

quite big change

#

i think this should include everything from the pins,
anyone who wants to sanity check feel free to send me if you find something

rugged solar
#

Seems like there is room for improvement for boon + torch still

rugged solar
#

(new = reference)

#

Torch fix v2 (experimental)

- actions.cds+=/use_item,name=manic_grieftorch,use_off_gcd=1,if=!stealthed.all&(!raid_event.adds.up|!equipped.stormeaters_boon)
- actions.cds+=/use_items,if=buff.symbols_of_death.up|fight_remains<20
+ actions.cds+=/use_items,,use_off_gcd=1,if=buff.symbols_of_death.up|fight_remains<20|!stealthed.all&equipped.manic_grieftorch```
past fable
rugged solar
#

nice find

#

0.3% is already something

past fable
#

This was using the preliminary "10.1 talents", with inev+swift death and just swapping macabre for sepsis (so only 6s dance)

rugged solar
#

what is the apl change you did in the test

past fable
#

actions.cds+=/sepsis,if=variable.snd_condition&combo_points.deficit>=1&target.time_to_die>=16&cooldown.shadow_dance.remains<10
instead of
actions.cds+=/sepsis,if=variable.snd_condition&combo_points.deficit>=1&target.time_to_die>=16

rugged solar
#

the change looks fine, but ends up a loss in dungeon slice and mixed on >2 targets

rugged solar
#

monkahmm think i can fix that

#

Eleems change for sepsis

actions.cds+=/sepsis,if=variable.snd_condition&combo_points.deficit>=1&target.time_to_die>=16&(cooldown.shadow_dance.remains<10|raid_event.adds.up|spell_targets.shuriken_storm>3)```
3 targets: <https://www.raidbots.com/simbot/report/h35yua6RhzP3jduxCJEH8t>
2 targets: <https://www.raidbots.com/simbot/report/ahH32VLrsPH9oNQzUW9EWL>
1 target: <https://www.raidbots.com/simbot/report/kJKA6LY6rJ7bLET7e8V8kg>
Dungeon Slice: <https://www.raidbots.com/simbot/report/ft6SJRcZHBveAEhNyKQbqB>
Note: cooldown remains can be adjusted up to ~14 without big impact, lower valu seems beneficial to get higher benefits from DM.
valid robin
#

fuu i have a suggestion for rotten+flag apl

#

as in a complete gameplay loop

rugged solar
#

oh ye i forgot to check the one you meantioned recently

#

i can check it

valid robin
#

well

#

i have a gameplay in mind

rugged solar
#

sure, just tell me

valid robin
#

Shadowstrike + Guzzler
Gloomblade
Flagellation
Rupture
Shadow Dance
Symbols of death -> Gloomblade
Secret Technique
Shadowstrike
Evisc
Shadowstrike
Evisc

  • end of dance -
    gloomblade until 6 cp
    evisc
    vanish
    Shadow dance + Symbols of death
    gloomblade
    sec tech
    shadowstrike
    evisc
    shadowstrike
    shadowstrike/evisc

  • play normally -
    hold 2nd vanish until 2nd flag is up, approx 1 min into the fight
    hold flag for shadow dance as they tend to desync slightly
    only use vanish after using flag AND shadow dance to ensure 2x shadow dance uses with each flag

  • min/max potential -
    you can perhaps delay the 2nd shadow dance during flag slightly to get 2x cold blood in each flag as well.

  • notes -
    this does desync about 3 minutes into the fight (on 4th flag iirc)
    but it falls into sync on 5th and 6th flag and by this time most fights are over anyway.

rugged solar
#

whats guzzler

valid robin
#

the trinket

#

from this tier