#tc-research

1 messages · Page 42 of 1

noble pendant
#

Yeah venthyr only

thorny fog
#

venthyr has access to two powerful haste buffs so makes sense

#

they even rotate like good bois so it has more uptime

gentle cave
#

How many days/weeks would it take to get back on renown level. When you switch from Fae to venthyr

spiral light
#

I think it takes 2 resets of very active gameplay. Continue at #wow-general

lean talon
#

Okay looks like we had a bunch of bug fixes on the PTR.

  • CtO has been fixed, meaning you don't lose your CtO buffs when the original RtB buff fades through timing out or rerolling (thanks to @thin night).
  • Echoing Reprimand from Stealth doesn't apply Find Weakness for Outlaw and Assassination anymore.
  • Crimson Tempest now properly grants an Alacrity Stack for Assassination.

Still bugged:

  • Envenom still doesn't pandemic
  • Pistol shots from Concealed Blunderbuss still dont Blade Flurry, dont proc combat potency and don't benefit from Opportunity
  • Finishers consuming the animacharged CP with Grand Melee up still don't grant Slice and Dice as if they consumed 7 CP
  • Shiv still doesn't proc Combat Potency and doesn't work with Broadside
light osprey
#

kekw they took a year to do this better buff firemages

boreal patrol
#

Both, i'm in the camp that i usually run stuff locally though because i have a simple VS Code workspace setup that I'm comfortable with and can also use my scripts (that power Hero Damage) if necessary too.

regal agate
#

raidbots is sufficient if you want to only change the apl or spell data (advanced sim).
if you want to work on the simc code(an compile simc) or adjust logs you will need to do that locally or on wcl.

#

if you run a very high amount of combinations on simc its possibly also better to do it local assuming that your hardware is decent

wet mountain
#

Thanks

plush kestrel
#

@lean talon did they fix BTE flurry?

lean talon
#

Ooh forgot to check that will do later good call

west prism
#

What exactly is the bug with BTE flurry?

lean talon
west prism
#

seems to be bugged still

lean talon
#

yup also just tested, its exactly 20%

copper roost
#

is the new venthyr ability going to be worth switching to?

raw knoll
#

you can ask in your spec channel. this is a research channel

merry cloud
#

Tried to optimise Flag + Exsang combo, mostly focused on NS variations of talent setups but it was gain/neutral for Subtr as well, MA is tricky because Ven/Vanish are slightly less often used there with sync for whatever reason

One of the test sims: https://www.raidbots.com/simbot/report/ffbDtnjQkDNHqcT4dNEYE5

Diffs:
https://www.diffchecker.com/j9PqyBSr/
https://www.diffchecker.com/4yqvHpqG/
https://www.diffchecker.com/s5kzyCaw/
https://www.diffchecker.com/f34TZo6T
https://www.diffchecker.com/kEMwhuYH

Raidbots
Advanced Sim - 6,611 DPS - Raidbots
Diff Checker
Diff Checker
Diff Checker
Diff Checker
knotty oriole
#

Pretty significant loss in DungeonSlice unfortunately so probably needs some more logic here to be able to be integrated. I'll look into some things.

west prism
#

It does not seem like Secret Techniques is giving a stack of Alacrity, not sure if this is a known bug

west prism
#

Also seems like Alacrity has a chance of giving you 2 stacks if you play Deeper strat as sub and finish at 6.

bc39c2c55cb4675ff6295ae1ebdff7f2.png
#

Logically it's doing 100% on 5 and 120% at 6 CPs, with the 20% being a chance to give 2 stacks

civic mango
#

Yea, that's always been a thing

#

just like how you can get 2 combo point procs when spending 6

west prism
#

Alright, i asked around and nobody seemed to know. Makes sense either way

golden briar
#

been like that for ever ye, but ST not giving stacks is weird

west prism
#

It's just the same old story with ST tbh

#

Barely any interaction with rogue things

civic mango
#

If they made it interact with everything it wouldn't be much of a secret, now would it

plush slateBOT
granite fossil
#

are there any good guides on understanding sims and learning how to manipulate em to try diff things or is it jus a learn as ya go thing

noble pendant
#

Would be the place to start

granite fossil
#

Thanks, appreciate it

formal turtle
#

Has there been any analysis regarding the changes to Flaggelation and the legos?

regal agate
#

look at the pins in the spec channels

lean talon
#

Another ~20 dps optimization for Dreadblades. Both making the APL only cast Blade Rush at very low energy during Dreadblades, and making the APL cast every Pistol Shot during Dreadblades even without Quick Draw are worth about 20 dps individually, though putting both the changes together doesnt seem to increase the gain, probably because they both achieve similar things individually.

This still brings Dreadblades over the best setup with alacrity, again EZ

https://www.raidbots.com/simbot/report/dMg6vSzh9BybHpvy8dSfYU

knotty oriole
#

Will take a look although still don’t really see how this build is practical unless it is actually significantly better since it’s completely unplayable on many fights. But I would probably want to isolate how that would work with some different builds considering the conflicting changes.

#

Pistol Shot change seems the most potentially problematic

#

In comparison

lean talon
# knotty oriole Will take a look although still don’t really see how this build is practical unl...

Well it gives us a pretty different damage profile, and an actual burst CD we arent punished as much to hold. On shorter encounters and when the timer lines up very well with the CD, it is actually significantly ahead too (https://www.raidbots.com/simbot/report/aVqwzsGVdGbRtD9Sk15d1f 1% over alac on 2min). Probably also makes active stat trinkets significantly better for us.

Unless by completely unplayable you mean the damage it does to yourself? From testing it on dummies that is so little its barely noticeable

Raidbots
Advanced Sim - 7,308 DPS - Raidbots
raw knoll
#

@knotty oriole having an alternative talent setup that allows for a different damage profile seems rather big. The negative hit of dread blades is basically non-existent unless you’re referring to something else making it “completely unplayable on many fights”. Could you elaborate on that?

knotty oriole
#

I mean on dummies it’s not going to be an issue but I can’t imagine taking 35-40% HP worth of damage over 8s is a fantastic idea on prog just to gain 0.2% damage to be perfectly honest. Has some role for sure with the damage profile but always did. At the end of the day half a percent either way is always going to be negligible for noticing either way in most cases.

lean talon
#

40% HP would mean 8 finishers. and i think the damage is reduced by vers, it seems my passive health regen alone outheals it, even without recup

#

and before that the damage profile was not a thing, it was 300 dps worse and had basically no burst

knotty oriole
#

I mean that is why it is not noticeable on dummies but damage is damage.

raw knoll
#

remember lurkers gift was a thing.

knotty oriole
#

You will be taking that damage in addition to anything you already were taking in a real encounter and your healing is already going towards keeping you alive. Can’t really evaluate it in the vacuum like that.

#

Blindly using it on CD on some raid encounters would just be straight inting at times. 😛

raw knoll
#

healers would like to parse. kekw

lean talon
#

idk one use of it is doing 8.8k damage to me. sure sometimes its a bad idea, but especially with recup being so effective at topping us vs small damage i really dont think its a serious problem

unknown.png
knotty oriole
#

Although this is really only burst on the scale of Outlaw. So I wouldn’t overplay it too much. The burst profile on the damage graph really isn’t that much different tbh.

lean talon
#

this is quite the difference. it makes flagellation deal 36k damage per execution on average vs 23k

unknown.png
knotty oriole
#

Not like we don’t already do this kind of thing with certain roll combos anyway. I wouldn’t really say it’s going to stack up burst wise to anyone else.

lean talon
#

and its a consistent CD you can line up where it matters, unlike AR

#

outside of the opener, the alac profile only ever peaks at 7k dps. the dreadblades profile peaks at 8.5k

knotty oriole
#

13k extra damage over 12s isn’t exactly what I’d classify as a huge difference in burst damage tbh. That’s kinda my point. Yes it’s better burst but you’ll still be middle bottom DPS during openers or burst windows comparatively. 😛

lean talon
#

well no thats not what i said

#

flagellation alone does 13k more damage

knotty oriole
#

Well that 1k extra DPS is most of the difference you are describing.

lean talon
#

i really dont get your point. this actually makes outlaw burst as high as assassination IN EXECUTE, it sims a bit higher on 5 minutes and way higher on shorter/better fight times. and it does 8k damage to us every 120 seconds

knotty oriole
#

My point is that yes it is a burstier profile but it’s not like “huge burst” or anything like that. Just trying to be realistic about terminology. After all, the difference in the 5 minute sim is only 0.1% so there’s only so much happening here.

lean talon
#

i mean it doesnt turn us into a convoke spec or a fire mage sure

#

but having that on sludgefist would be great

knotty oriole
#

Generally try to temper use of terms like “significantly” on the TC level for things that will actually be notably different in practical terms and I don’t really think this is quite there yet. But we will see. 🙂

lean talon
#

i didnt use that term, i guess you mean me saying "way higher"?

noble flicker
#

@lean talon I heard you may have been testing whether or not opp procs were worth casting without QD and having certain buffs?

lean talon
#

it has more to do with energy than with buffs

plush kestrel
#

@knotty oriole @lean talon could Invigorating Shadowdust be a legitimate consideration for DB Flag build? Could potentially get the cd down of both flag and DB to 50 seconds

#

I'm not savvy enough to build a sim for it with the conditions for vanish but was thinking a cd reduction build might be worth it with new flag

knotty oriole
#

Would it need any specific conditionals? APL pretty much uses Vanish as quickly as possible. Other stuff should be adjusted uniformly.

plush kestrel
#

Well you would have to use vanish after DB and flag went out

#

Or else the reduction from isd wouldn't apply to them

#

I'm just not savvy in simc or building of any kind. Sorry I'm so naive!

knotty oriole
#

I would probably suggest just taking a look at how it is working now to get a sense of how it’s timing things. Probably also need to double check they remembered to make it work with the new Flag spell lol

plush kestrel
#

Oh true!

knotty oriole
#

Haven’t checked that on PTR yet

plush kestrel
#

It works with current covenant abilities at least Kyrian. Not sure on new flag

knotty oriole
#

I would assume it works but probably should check

plush kestrel
#

Yeah definitely. Thanks!

rocky latch
#

it does work on flag as of right now

rapid fern
#

I hopped on PTR to play around with the flagellation + dreadblades + invigorating shadowdust build, and invigorating shadowdust is currently reducing the CD of dreadblades twice, -40s total, on PTR. I checked the bug tracker and didn't see that listed there.

knotty oriole
#

I think there is a bug for it

#

Blizzard only knows why that makes any sense

#

lol

rocky fern
#

Yep, not sure if it's still in effect but when I was dabbling earlier that was the case. lol Blizz logic.

rapid fern
#

ah, sorry. guess I was looking in the wrong place then

knotty oriole
#

np

plain coral
#

Posting here in case

#

it looks like Shadow Blade is procing off initial Flagellation hit

#

which might be a bug since it doesn't generate CP anymore

knotty oriole
#

This will be simming this way since it's still in the Shadow Blades whitelist

#

Even though it's modifying nothing

#

I would still view it as a bug yes

plain coral
#

this is today

unknown.png
#

is this only Doomblade bugfix ?

boreal patrol
#

probably

plain coral
#

ok ok

knotty oriole
#

100 DPS seems about right

marble salmon
#

Are there Dungeon Slice sims for Outlaw with new Venth and Necro, MA vs Cele vs Deathly Shadows, and BS vs KS?

#

or can someone lead me in the right direction to SIM what I want? I just dont know how to sim v 9.0.5 without 9.0.5 being out

maiden cargo
#

hey guys, thank you so much for all the hard work you are doing in here its really helpful and amazing of you all

lean talon
#

quick +30 dps ST damage gain for bonespike. First of all, removed the slice and dice condition which didn't seem to do anything. I guess it was initially put there to avoid overcapping energy from bonespike early in the fight and prioritize SS over it?
The damage gain is from adding a condition to use bonespike to reach max CP exactly when Opportunity is not up.
Might need to be rephrased for AoE situations since this just assumes that bonespike gives 2 CP, which is fine for ST.

(new line: actions.build+=/serrated_bone_spike,cycle_targets=1,if=!dot.serrated_bone_spike_dot.ticking|fight_remains<=5|cooldown.serrated_bone_spike.charges_fractional>=2.75|combo_points.deficit=2+buff.broadside.up&buff.opportunity.down)

https://www.raidbots.com/simbot/report/tBRQAihusN4JgBTALKpkCN

Also this is only for Outlaw but should be a gain for Sin and the other spec too (adjusted for seal fate for sin i guess, and without the broadside rule)

knotty oriole
#

Well the SnD condition was primarily because the damage was so low it wasn't worth dropping the AA damage for

#

Not surprised it's non-relevant now though at 4x damage

#

CP threshold logic seems reasonable

#

But yeah will need more work for AoE

#

It's pretty huge loss atm

#

for DSlice

#

-1.2%

lean talon
#

yea because it actually makes it use bonespike worse for aoe, overcapping a lot of cp lul

knotty oriole
#

I'll probably have to make an expression to return the active bone spikes

#

Will take a look at that tonight and see if I can find something that works for aoe

#

I mean could slap a target threshold on it but I don't think that's optimal

#

Shouldn't be too difficult to add an active_bone_spikes thing

#

Also think I'd split these lines

#

Because I think it's best to let the cycle_targets logic run first

lean talon
#

yea. although in most aoe situations you would probably just spread the bonespikes out asap anyways, or not?

knotty oriole
#

And then just fallback use on the main target for the CP condition

#

If there are no valid cycle targets

#

Because this is likely circumventing multi-dot opportunities

#

Since it'll return true immediately on the main target

#

I think I'm gonna split that whole line apart

# 
actions.build+=/serrated_bone_spike,cycle_targets=1,if=!dot.serrated_bone_spike_dot.ticking
actions.build+=/serrated_bone_spike,if=fight_remains<=5|cooldown.serrated_bone_spike.charges_fractional>=2.75|combo_points.deficit=2+buff.broadside.up&buff.opportunity.down)
#

is probably just universally better

#

for MT

#

Yeah just splitting it helps a lot

#

Gets halfway there anyway 😛

lean talon
#

yea i dont really thing its very realistic to think about using it to exactly cap CP in aoe scenarios. maybe having a blade flurry sync would be benefitial now though with it doing actual damage

knotty oriole
#

@lean talon So I added a few expressions last night to help with this kind of thing

#

active_bone_spikes is now a global expression

#

Also updated cp_gain for the serrated_bone_spike action to take into consideration the proper bleed gen, so for like that like you can do cp_gain>=x

#

And not have to worry about broadsides or bleeds on sbs lines themselves

#

Also updated the "bleeds" expression to return the updated count for target bleeds for Doomblade on the Assassination side

#

I had some luck locally last night with

actions.build+=/serrated_bone_spike,cycle_targets=1,if=variable.blade_flurry_sync&!dot.serrated_bone_spike_dot.ticking
actions.build+=/serrated_bone_spike,if=variable.blade_flurry_sync&(cooldown.serrated_bone_spike.charges_fractional>=2.75|combo_points.deficit=cp_gain&buff.opportunity.down)|fight_remains<=5
#

Which gets pretty close but it's still behind in DS

#

Was gonna thrash it a bit more with raidbots today

#

It still may be acceptable to drop like 0.1-0.2 to have this logic in for DS though if it's just reset timing edge cases

fallow comet
#

does anyone know why well-placed_steel:5 isnt working in simcraft?

regal agate
#

you used a - instead of a _

fallow comet
#

i tried that as well

boreal patrol
#

wellplaced

fallow comet
#

ah ok

boreal patrol
#

dashes are discarded during tokenization

regal agate
#

you can also just use the id instead: 238:5

fallow comet
#

OK thanks, is there a list with the ids on them? i dont know where to find that info

merry cloud
lean talon
mystic jolt
#

Does the apl try to snipe sbs on low health targets in dslice to get extra resets?

regal agate
#

I attempted sniping logic early but i don't think any spec atm tries to actively enforce sniping

mystic jolt
#

I imagine it'd be similar to any MFD sniping lines. Might be able to boost some dslice sims but I'm not as well-versed in simc code yet to worry about it too much, just food for thought!

regal agate
#

a modified mfd line was the very first thing we had as condition for it.
But it evolved over time, you need to consider that you always compete with other builder.
We atm use TTD (target time to die) conditions to estimate when its worth to use or not,
you can lower those values and see how it works out.

mystic jolt
#

I might try and fiddle around with it later today

regal agate
#

but reduce the output by ~3.5%

unknown.png
#

possible that limiting the condition to a more specific use case could make it a win.
Also possible that outlaw/assassination has more potential for gains.

mystic jolt
#

I'm wondering if when someone has MFD + sbs, the competing condition causes a lower output specifically (speaking from an outlaw perspective ofc)

regal agate
#

the sim does not use mfd

mystic jolt
#

ok

#

For sub specifically there might need to be a check for if FW is up and/or if sbs will generate more or an equal amount of cp as sstorm in aoe, so there may be a gain for 2-3 targets but once you start getting to higher target counts (5+) sstorm will just overtake sbs because of cp generation. I don't see that in the apl cast condition in your sim but I may just be missing it

knotty oriole
#

I mean obviously this is increasing the SBS DPS but

#

It's likely it is having a negative effect on other things

#

So I think it's more about finding time window limitations that don't give up other sources of damage that may be more beneficial

#

Find Weakness uptime on that variation is lower for example

#

By a notable amount seemingly

regal agate
#

i did test limiting it to a target cap, it migrated the loss but did not opt a win

mystic jolt
#

Which is where my consideration for having a FW condition comes in, I'm trying to think of how to implement that though

regal agate
#

i am with koji, i think the better approach would be to try to find the "special cases" when it would make sense to use sbs additionally

knotty oriole
#

It also loses like 3% Black Powder DPS

#

Which offsets almost all the gains by itself

#

So there may be some target count thresholds or other things that might limit it enough to be a gain

#

Definitely tricky for Sub though

regal agate
#

the problem for subtlety is rly shuriken combo because it becomes so efficient with multiple targets

knotty oriole
#

Yes I mean

#

The average SS cast is 4.6k

#

SBS DD average is 3.9k

#

So almost certainly this is not worth when considering FW uptime at high target counts

#

Maybe at some lower amount

mystic jolt
#

Yea, I imagine it'll only show any potential of being at least equal dps in 2-3 targets where each has FW up already

#

and the sbs dot rolling

regal agate
#

i assume that looking at outlaw/assassination has higher potential if you want to invest time into it.

mystic jolt
#

what's the code for checking if FW/sbs is up on multiple targets (rather than just the current one)?

knotty oriole
#

fw_targets

#

But think that's only available on black powder action so might not be available in the context you're talking about

mystic jolt
#

Ah, ok

knotty oriole
#

Might be able to make it global

regal agate
#

i don't think he needs fw_targets in that case.
Just checking find weakness in the cycle target condition should be suefficient like we do with shadow strike:
simplified version of what we use for shadow strike:
actions.stealthed+=/shadowstrike,cycle_targets=1,if=debuff.find_weakness.remains<1

#

we already use cycle_targets for one of our bone spike conditions
so its something you get "for free" when modifying that one (like i did in the sim i posted earlier)

twilit granite
#

how is sbs ptr ver implemented on simc? Does flurry all bleeding dots?

#

Or isnt implemented yet

plain coral
#

I did not understand your question 😄

twilit granite
#

On ptr, rn, if you SBS 1 mob, the dot flurries 85 dmg, if you SBS 2+ targets the dot still flurries 85

#

On simc, does it flurry all dots or only one of them?

lean talon
#

Can you post evidence of the dot flurrying on the ptr, because it doesnt behave like that on live

twilit granite
#

My ui is a mess, will try

boreal patrol
#

The dot does not flurry. Are you sure you're not just seeing flurries of auto attacks?

twilit granite
#

I tested it with @sullen grotto a few hours ago

#

It was doing 82-84 dmg (after autoattacks flurry)

#

88 sometimes

lean talon
#

A simple log would help a lot. I can also hop on ptr and test it in like 20min

mystic jolt
#

I did see something like that when I was messing around on ptr, but I wasn't sure if it was the dot or autos doing it

plain coral
#

I'm testing rn and I don't see the dot flurry at all

twilit granite
#

Same, is not flurrying anymore

#

I guess they fixed it? Had to take logs damn

#

The dots hit for 211, 211*0.4 = 82

#

But can't see that 82 anymore

#

Everything is the same, same gear same casted spells

plain coral
#

yeah but it's also what your off hand weapon is doing

#

so I think you might have see a simple off-hand attack flurying

twilit granite
#

No, i had my offhand rmoved

#

to test that particularly

plain coral
#

even if you remove your OH weapon, you will punch the dummy you know 😄

twilit granite
#

Uhm is not punching for me 😛

#

ah i removed my MAIN hand, not the offhand

lean talon
#

it doesnt flurry this is me standing next to two dummies with the dot on them and flrury up

unknown.png
plain coral
#

if the dot would flurrying, it would flurry even if you are not activly auto hitting the dummy

lean talon
#

my auto attacks flurried for about 80-90 so i guess thats just what you saw

twilit granite
#

Yeah idk, is not flurrying anymore 😦

#

But i remember i removed a weapon just to avoid that second autoattack

#

so idk what it could be if it wasnt the dot

noble pendant
#

Maybe the flurry, was in our hearts the whole time thonkang

twilit granite
#

Damn 😦 i feel sad now

#

was too good to be true

lean talon
#

i mean you attack even without weapons equipped at all right. they definitely didnt just hotifx something on ptr

boreal patrol
#

pro tip when testing abilities on dummies. there's a certain angle where you can still use any ability but won't auto attack. i use that quite a lot

#

should be somewhere around 45 degrees facing away from the target

twilit granite
twilit granite
#

(arena tournament)

unkempt crescent
#

That is still possible

plain coral
#

@knotty oriole
Btw I found a small gain allowing Envenom to be casted freely during Flagellation (basically considering Flagellation the same way Envenom consider Shiv or Vendetta)

The change is basically
actions.direct=envenom,if=effective_combo_points>=4+talent.deeper_stratagem.enabled&(debuff.vendetta.up|debuff.shiv.up|debuff.flagellation.up|energy.deficit<=25+variable.energy_regen_combined|!variable.single_target)&(!talent.exsanguinate.enabled|cooldown.exsanguinate.remains>2)
Seems to show a small ~15dps gain

Also tried to get rid of this wierdness :
https://i.imgur.com/jZCn3bN.png
But was not able to find any gain, it seems however that the Envenom change sorta made this edge case even rarer.

https://www.raidbots.com/simbot/report/6XUmNk4KYL57jS1Tk8mys6

Imgur
Raidbots
Advanced Sim - 7,354 DPS - Raidbots
knotty oriole
#

Pretty tiny, but still worth tossing in.

plain coral
#

nice

#

I thought it would make sense of removing the pool condition for envenom during Flag but I expected a better gain tbh

twilit granite
#

Is there a call to start a sim with an active buff?

#

I am thinking in actions.precombat+=/slice_and_dice,precombat_seconds=1, but i guess it wont work without combo points

#

Want to sim and encounter with SnD up prepull and rupture already on target

twilit granite
#

^Managed to fix it.

#

Why shadow blades damage is that high when adding raid_events+="/vulnerable,cooldown=70,duration=12,first=0.9?

#

It seems it has a weird interaction with vulnerable event

plain coral
#

maybe it double dip on raidbot

#

like it would start with an 100% increased value because SS has increased value and then this value is increased by 100% again such that it deals 4 time normal damage

#

I'm unsure it work this way on live tho probably a simc bug

plain coral
#

@knotty oriole
I noticed randomly using fixed_time=1 sim over 300s that your recent change appears to be a big loss over exactly 300s : I just removed the
effective_combo_points>=4 condition from flagellation.

What's even more wierd is that it's a small win as long as you add variance in time lenght. This mean it is probably a big win for some specific fight lenght, and big loss for some others.

I'm trying to investigate what happen in this 300s sim but I think it is delaying Flag to build up 4combo (maybe delaying for like 10sec) which might end up loosing a cast of flagellation.
If we can fix this edge case it might be a bigger win overall.
https://i.imgur.com/XqWvitx.png
https://www.raidbots.com/simbot/report/5QKvuKcjqU7K1LbawY7nbY

Imgur
Raidbots
Advanced Sim - 7,352 DPS - Raidbots
#

Yeah ok that's it !
In exactly 300s sims, on some iteration the combo point condition is not true when the engine usually run 2nd Flag at 1:30. So the engine is building up 4cp, but once at 4cp the Flagellation sync cooldown isn't true anymore (missing maybe 10sec ?) So it delay Flagellation to 2:00 to sync with Vendetta.

But the correct move would have been to ignore the Combo Point condition and just fire Flagellation at 1:30, because over 5:00 you gotta fire on CD or you loose a cast.

My suggestion is that the cooldown sync condition needs to have absolute priority over the CP conditions.

#

https://i.imgur.com/vMJbh4J.png over 4:30 your change seems to be a solid gain. With ~0.3% over other time lenght around 4:20 and around 5:30.
But I think it's counter blanced by some potent loss around the 300s marker.

Imgur
plain coral
#

I don't know if you already dived into it, but it seems it's an edge case that only happens with iterations that last exactly 298 to 301s

knotty oriole
#

Generally this is why we don’t use pure fixed length sims though. There are always edge cases but the APL is intended to be generalistic. There are plenty of APL rules that are a loss at a very specific fight length but that is usually not something that matters. The reality is that the “lose a cast” line is always going to be fuzzy in reality because nobody knows the exact fight length to the second. I believe there is an interpolated_fight_remains nowadays that introduces error on purpose to simulate this.

plain coral
#

I think I found a fix, but it's so edgy that it doesn't change anything

knotty oriole
#

May look at what happens if you use that on the line instead

#

Of the exact fight_remains

plain coral
#

I just wanted to be sure that we were not in the case where a bigger win was smothened by this edgy case

#

but looks like it's not

twilit granite
#

It seems its doing x2 times damage it should do.

plain coral
#

I don't know :/

#

Koji might look into it if he think it's relevant enough

twilit granite
#

Nah its not since it works for everything else, is just that case

#

If I extend the fight duration the bug dissapears

plain coral
#

that's wierd

twilit granite
plain coral
#

because you are not using SB during Vulnerability in this case

twilit granite
#

Right, good point

#

It seems its a SB issue as I suspected

#

Well not sure because rest of the dmg looks weird too

plain coral
#

yeah I think Shadowblade is double dipping on Vulnerability in sims.
If it's indeed the case that's something related to simc core and I can't really help you

twilit granite
#

Okay thanks 🙂

boreal patrol
#

Yeah, Shadow Blades was double dipping off target mods in sims. This seems to be from an old game bug change I made end of BfA that we missed during SL prep when it was fixed.

#

Fixed. Means sub sim deeps might go down a bit due to Sinful. Lovely.

plain coral
#

Ty good catch @twilit granite

twilit granite
#

@plain coral actually it was because of you xD with the sludgefist profile we made two nights ago

plain coral
#

would have guessed ye 😄

orchid flame
#

random question about the apl: do we know if anyone has tried to use charges of SBS "smartly" to get ~15 seconds of agi buff uptime with emeni? this could potentially be a gain when aligned with cooldowns (opener would be easiest to measure).

noble pendant
orchid flame
noble pendant
#

It showed no gain

lofty halo
#

Hi im new at simming but I have a cursed question so sorry if i missed something obvious.
https://www.raidbots.com/simbot/report/xyaQsk6XMehxJAfmSaAryo
https://www.raidbots.com/simbot/report/mkxGjhEWU61a1XacLdcozi
These are two sims on my char comparing greenskin and blunderbuss with celerity. On the second on with lower dps I modified the APL to keep skull and crossbones as a single buff since it synergizes with blunderbuss. Although i was expecting the relative dps difference between the two leggos two change as a result of the APL change but it is the same as with the standard priority. This seems weird right? Not sure if im missing something

Raidbots
Top Gear - Sendnis - 5,727 DPS - Raidbots
Raidbots
Top Gear - Sendnis - 5,618 DPS - Raidbots
noble pendant
lofty halo
#

hmm i was expecting a bigger change cause i think blunderbuss should proc roughly 1.2-1.6 times as often as greenskin unless you have broadsides where it procs 0.8 times as often but maybe thats a smaller difference than im giving credit for

knotty oriole
#

The difference in uptime is only like 3%

#

Rerolling is of very minor consequence nowadays given CtO

maiden cargo
#

noticed a quirk with the marrowed gemstone buff from the bonesmith heirmir soulbind

so, with the soulbind you should gain marrowed_gemstone_charging for up to 10 stacks and then marrowed_gemstone_enhancement triggers, with a 60s cooldown

i think something is causing the charging buff and the enhancement buff to overlap at the same time, and when it does happen the charging buff remains at whatever stack it was until the next enhancement triggers

https://www.raidbots.com/simbot/report/dsfPmUnazid5Wm6ustqimD like this sample sequence, the enhancement triggers early and the charging buff gets stuck at 1 stack. it happens later in the sample at 4 stacks. note that this is a dungeonslice sim but ive seen it happen in patchwerk as well

unknown.png
#

on further inspection i think the enhancement buff is triggering normally but the charging buff is being "shut off" too late

knotty oriole
#

Not totally sure since I didn't work on any of the Soulbind stuff

#

But I can pass it along

raw knoll
#

you can actually continue to cancel aura the charging 10-12 seconds prior to CDs being ready to try to stack the crit chance buff with the cds

plain coral
#

@maiden cargo have you been able to check if this also happen in game ? Because I know BotE passif effect use to work this exact same way sometime

#

I don't remember if it has been fixed or what, but I do remember that you could get a couple of stacks before the haste buff was triggered and you kept these stacks

knotty oriole
#

This might not actually be incorrect looking at the sim, it seems maybe it adds 1 extra stack (so it's 10 instead of 9) to make checking it easier

#

But I'd need to debug it

lean talon
#

https://www.raidbots.com/simbot/report/bRzm6Xtx1hYzFLWXTeHdRF there is a decent gain to be had by spacing out bonespikes if you run Emeni so you dont clip your LBE uptime. buff.lead_by_example.remains<=3 is by far the best threshold, i tested 1, 2 and 4 as well.

Gain is slightly bigger even in a 3 target patchwerk sim even though it should only ever be relevant in the opener here.

https://www.raidbots.com/simbot/report/5krWKSNdA1omBPig7RDecV (3T 3min)

https://www.raidbots.com/simbot/report/x3LvaXrKqj4tDLkdKreBFF (3T 90 sec)

Mostly irrelevant for ST sims since the APL rarely if ever casts more than 1 bonespike within 5 seconds anyways

https://www.raidbots.com/simbot/report/79kHV1HvmiwuD44dCxg6nJ

noble pendant
lean talon
#

whats the sin condition? wonder if we can just steal that and its better than mine

noble pendant
#

serrated_bone_spike,cycle_targets=1,if=master_assassin_remains=0&(buff.slice_and_dice.up&!dot.serrated_bone_spike_dot.ticking|fight_remains<=5|cooldown.serrated_bone_spike.charges_fractional>=2.75|soulbind.lead_by_example.enabled&!buff.lead_by_example.up)

boreal patrol
#

Both sub and assa have it. I think it was not worth for Outlaw when i looked at it in earlier SL

noble pendant
#

yeah sub has serrated_bone_spike,if=cooldown.serrated_bone_spike.charges_fractional>=2.75|soulbind.lead_by_example.enabled&!buff.lead_by_example.up which is basically the same thing

lean talon
noble pendant
#

Imo at least

solar swift
#

!bug

plush slateBOT
solar swift
#

IS is still bugged? -40 sec for DB and flag, is it possible to sim it w/o bug?

#

bugs=0 doesnt work

loud temple
#

does raidbots account for the dps from the set bonus from the covenant hall? (I know its bad, but im just out growing the gear)

boreal patrol
#

Shadowdust bug is not implemented in simc and no, it does not sim the covenant set.

solar swift
plain coral
#

might have vanished twice between both DB/Flag cast

solar swift
#

hm, true, thx)

plain coral
#

I cannot do it rn, but if any tc-er want to check about opener things especially with no lust (or no lust on pull) :
My statement is that we starve in energy with current opener when using Deeper Stratagem, waiting for 3sec during the opener before garrote and it feels not good.
My suggestion is : Mut > Rupt > Mut > Gar > Flag > Snd + Vend > Shiv > Vanish + Mut > Env > Mut

#

Or maybe even Mut > Rupt > Mut > SnD idk but not having bleed for the first 6sec of the fight really feels wrong without Vigor

west prism
#

Bonespike isnt affected by Shadowblades, which is very likely a bug

merry cloud
#

Openers typically make very little difference if any in the overall fight dps

shrewd crater
#

Hi, did anyone try if its worth to save 1 Charge of SBS with Emeni's LbE for MA specifically, since LbE is 5s and MA 4s? Or to sync it somehow/hold one for the other?

noble pendant
boreal patrol
#

Quick request: Can some Sub Rogue who rolled Venthyr please check something for me? Looking at spell data they still have leftovers of Shadow Blades in the effects. I think that having Shadow Blades up makes the tooltip of Flagellation in the action bar change to say 2% Haste (without any practical effect). Would like somebody to confirm or deny this.

noble pendant
boreal patrol
#

during Shadow Blades?

noble pendant
#

oh

#

yes popping shadow blades indeed raises it

boreal patrol
#

thanks for confirming that

#

buff should stay at 1% though, right?

noble pendant
#

I don't see why it would go to 2%

boreal patrol
#

yes just making sure. thanks

noble pendant
#

pleasure

plain coral
boreal patrol
#

right

#

not any big deal now.. but it might become one in case they change flag at some point. also confirms that i read the data correctly

maiden cargo
#

noticed the sim still holds trinkets for master assassin atm

#

for outlaw

plain coral
knotty oriole
lean talon
#

I just noticed that the damage of Sudden Fractures (necrolord conduit) is reduced by armor in simc which leads to it dealing 30% reduced damage in sims. Tested it in game (vs dummies and open world mobs) and it is not reduced by Armor and deals the same damage as the normal dot ticks. Not sure if this is a recent fix or if it has never been reduced by armor

unknown.png
west prism
#

How much % dps do you think that is?

lean talon
#

hard to say because sudden fractures is kind of weird, it can crit but doesnt scale with broadsides so you cant just assume that it does the same average damage as the normal bonespike hit. if this would make it deal the same average hit (right now it deals 35% less damage in sims) it would be about 15 dps in the sim i did (which was 3t patchwerk). not huge i guess, but its actually not far behind sleight of hadn for example)

maiden cargo
#

i want to say dummies and overworld mobs can have armor value weirdness so maybe confirm in an instance just to be safe, could be wrong tho

lean talon
#

hmm maybe, i thought dummies might not have armor but overworld mobs definitely. ill zone into an instance i guess

lean talon
#

same deal in a mythic 0 dungeon

unknown.png
#

this is the damage difference in simc btw, average bonespike dot hit does 306 damage, average non-crit of sudden fractures does 201

unknown.png
west prism
#

It matches my regular dot dmg on sire hc as well

843f9b16b6f9d5ace3374f1cc699cae5.png
#

Did not know it can crit though, interesting.

lean talon
#

yes it can also be parried if you stand in front of the boss lul

west prism
#

Some things are just so weird man

#

i dont understand

lean talon
#

okay just did a key just to test it out and sudden fractures alone did exactly 100 dps, the average hit was actually higher than the average hit of the original dot (SF average hit 442, bonespike dot average hit 342). In a dslice sim, sudden fracture does 42 dps for me rn.
Also worth noting that this was a random +10 mists key i did for valor so mobs lived VERY short which should lower the value of SF

knotty oriole
#

Honestly can’t really see how it bypasses armor tbh but should be able to hack something in.

lean talon
#

its only about 30 dps behind SoH for me right now in a cleave fight

unknown.png
#

it also deals the exact same amount of dps in a dslice sim as in a ST sim while the normal dot gains atleast a little bit of damage

knotty oriole
#

I mean DPS isn’t really meant to be compared across the sim types due to DS downtime.

lean talon
#

yea but SF is directly dependent on the original dot right? so if the dot does more damage, shouldnt SF also do more

knotty oriole
#

Hard to say. I’d have to check all the whitelists to see if they are identical.

knotty oriole
lean talon
#

i get that there are differences (i mentioned earlier that SF is not affected by broadside for example), im just thinking that the reason the dot does more damage in dslice is because of higher uptime (114% instead of 98%), which you would assume should also make sudden fractures deal more damage

knotty oriole
#

if I were to spitball a theory randomly without looking at the sims yet, I think it's skewed by higher Broadside uptime from the restealth Ambushes

#

but I'll look at it later

#

Most problematic missing whitelist thing is actually that they forgot to add it to Assassination Mastery

#

Probably because it's direct and not periodic

lean talon
#

well it doesnt count as a dot at all, it just hits the target with a second hit x% of the time they take damage from it. really weird implementation

#

yea exactly

knotty oriole
#

Yeah but logically it should match the values and the fact that it's direct is kinda just jank

lean talon
#

yea agree

knotty oriole
#

Well had to change some SimC core code for this because Sudden Fractures is literally the only spell in the game set up like this. 👌

lean talon
#

ty once again koji PiratePoggers

west prism
#

I can't find the cheat death bug in the tracker, shouldn't it be? Because it is a bug right? (picture example of where it deals full dmg after the absorb part of the talent)

a7d5ff5e9bdc73a59e8235d0989896e1.png
lean talon
#

SBS spelldata suggests that the dot is not hasted, and in a 300 second sim at 98% sbs uptime i get 98 ticks of the dot, which means it does tick exactly every 3 seconds in sims. In game however it is hasted

unknown.png
#

just to proof that its not just a wrong tooltip and actually does tick faster

unknown.png
boreal patrol
#

pushed a fix

boreal patrol
knotty oriole
#

As far as I can tell nobody had previously identified any purpose to this flag but by process of elimination it seems to be the case

#

No idea why they are not using the standard stuff, but alas

lofty halo
#

Hi i noticed that the APL for outlaw doesnt take advantage of grand melee by using a finisher other than S&D when its up and S&D needs refreshed. I am not good a lua script but something like this should fix it i think
actions.finish=slice_and_dice,if=(buff.slice_and_dice.remains<fight_remains&refreshable)&!buff.grand_melee.up

willow knoll
#

Well is that an actual dps increase?

#

Can u do a sim comparison showing this to be an increase

lean talon
#

i simmed this a bunch and it wasnt a noticeable gain

lofty halo
#

i mean its probably small but its obviously a dps increase

#

i see no reason to not include it

boreal patrol
#

this looks like it would avoid casting SnD altogether if you happen to roll it in opener

#

and otherwise, that situation shouldn't even really happen?

lofty halo
#

most o the time, sometimes if you dont get grand melee for a while you have to refresh mid fight but its rare

#

extremely small difference but it is technically more optimal

willow knoll
#

Well this rule is only really when u get gm in the first roll

#

So you could just do Fixed_rtb=3

#

And check if it ends up being better

lean talon
#

it would only result in a dps gain at all if you got GM as your first roll AND often enough so that you would never have to refresh it even without casting SnD until the end of the fight

#

which is literally never

#

if you would be dropping SnD and having to refresh it at some point in the fight all you did was postpone the SnD

lofty halo
#

i dont understand why we need to verify this is a dps gain. it is like a 0.2 dps gain or something dumb but its so obvious

willow knoll
#

Is it a 0.2 dps gain?

boreal patrol
#

even then, i don't see why SnD in opener with GM up should be a loss? GM is applied on top so you just get some nice extra duration for free

lofty halo
#

it is + some amount of dps yes

lofty halo
#

but you are correct that in theory it could be fine either way, but by not pressing snd you prepare for the best case better

lean talon
#

and in that ULTRA RARE case it gives you one extra dispatch

#

which is barely noticeable in itself

lofty halo
#

is there a downside im not understanding? I realize its a small difference but i figured it would be nice to optimize where we can even if its small

lean talon
#

cluttering the APL

willow knoll
#

It could also just not be a dps increase

#

For example syncing killing spree with vanish for deathly shadows seems to "obviously be a dps increase"

#

But it just mathematically isn't

#

Making setting up the 45 second snd right away is better than letting it build with gm

lofty halo
#

actually i was talking to you the other day about syncing killing spree and MA and killing spree was equal to blade flurry with celerity but lost me dps with MA and your explanation was that it was because having MA made the APL try to use killing spree at all which made some sense but doesnt that mean that the MA sim would do more damage if it didnt press killing spree at all?

#

cause in your example about deathly shadows the difference is that killing spree isnt good enough to press, otherwise the most damage you fit in deathly shadows as possible is an obvious increase

#

so that must mean you can do more damage by not pressing KS

boreal patrol
lofty halo
#

wait i dont think that logic follows since the chance of rolling a bad buff is way higher than rolling a 5

#

also sorry if im coming off as combative im just trying to understand

boreal patrol
#

just like i think the chance of carrying a GM through the entire fight is lower than having to use SnD

lofty halo
#

but there is no opportunity cost with the GM example

lost echo
#

but wouldn't there be a slight increase of using a SnD anywhere other than the first finisher, since you'd have bte on cd and would be resetting that a bit

#

if you could avoid the loss of uptime with GM

lofty halo
#

ok i tried some sims and i think something is wrong with the modified APL i wrote so dont use that one lol. Its somehow resulting in lower S&D uptime

noble pendant
#

Limiting the usage of a skill usually results in lower usage of that skill

lofty halo
#

well yes but im talking about buff uptime. buff uptime should still be close to 100% except the opener

#

im assuming something about the logic i wrote is causing the sim to not rebuff when it needs to all the time which is unintended

lean talon
lofty halo
#

but if you press dispatch or bte instead, you get the snd buff. if the sim does that it should be the same uptime

lean talon
#

no if it gets GM in the opener, it will never manually cast SnD at all even after the first snd from GM drops

lofty halo
#

yeah thats what im saying, thats not the logic i was trying to write so i wrote the logic wrong

lean talon
#

correct

lofty halo
#

how would you write is so it does it correctly?

lean talon
knotty oriole
#

probably want

if=buff.slice_and_dice.remains<fight_remains&refreshable&(buff.grand_melee.down|buff.slice_and_dice.down)

lean talon
#

i think when i simmed that before the reason why it was such a big unexpected loss was because it would end up not casting SnD during the short 5 second CtO procced GMs which is kinda awkward, when i made it check for 3 or more seconds of GM it ended up neutral again

lofty halo
#

oh you think it was like clipping finishers off the end of GM and whiffing?

lofty halo
lean talon
lofty halo
#

ah that will do it too. Lua script is weird

grave wharf
#

(this isn't lua)

lofty halo
#

oh its probably just custom syntax for config for the sim huh? It kinda looks lua esque so i assumed it was based on that or something

orchid flame
#

any ideas when we'll see the sbs tick fixes in? (hasted and no armor reduction iirc). if i can help in some way lmk

maiden cargo
#

armor is already in nightly iirc, hasted will be in next nightly

torn dome
#

github has all the commits if you want to follow

orchid flame
#

ahh right sorry, i'll check there. thanks

torn dome
#

So if you use a local simcraft, its already there after a rebuild.
If you use raidbots its whenever Seri updates, which is usually daily

boreal patrol
#

if you click on the Git hash on the right side of a raidbots report, it opens the history of what is already included in the sim

#

easy to use to check whether a commit was already in nightly or not

plain coral
#

@lofty halo Sometimes small change that seems to be "obvious" gain are indeed obvious gain, unfortunately some changes have unexpected consequences (like collateral damage) on some "edgy iteration" where your change is going to make the engine end up doing something insteed of something else in that very particular iteration will end up being a massive loss, but is drowned in the thousands of iteration and it just annihilate your 0.1% win and end up being a 0.1% loss or neutral change.

While it is theorically possible to fine-tune every change and check for every edge case, it will take a colossal amount of time to search for every edge iteration that prevent what seems to be a win to actually not produce good results. So at the end we just don't want to add a cluster of APL line to manage edge case just to see very marginal gain, especially since these kind of change usually have to be check'd again every single time a major patch comes out.

plain coral
lofty halo
#

@plain coral thanks for your response, i think this explains really well the disconnect I was having yesterday when trying to discuss this. This makes a lot of sense

wet mountain
#

Why does the APL use KS with stealth, instead of syncing it with vanish?

#

Atleast that's what it seems

Screenshot_from_2021-03-14_12-46-08.png
willow knoll
#

Because the APL is optimize to "just use KS with MA up" and then "hold ks for when vanish is back up" it doesn't care if its stealth or vanish the MA comes from
The command instructs syncing it with vanish just for ease of use

wet mountain
#

I see - I guess it doesnt significantly affect the DPS even if the first time you have KS up you will need to wait a few secs before you can vanish

lean talon
knotty oriole
#

Yes in practice this actually doesn’t particularly matter. Just do whichever is easiest for you.

#

I like doing out of Stealth typically since mobs are less likely to be doing some murdering frontal right away. But realistically you can do it either way.

#

I think there are technically some very slight advantages to doing out of Stealth for cto buff utilization which is why doing with Vanish instead of Stealth is technically a small loss in the apl.

#

But this is not really a concern in a practical sense

lean talon
# knotty oriole I think there are technically some very slight advantages to doing out of Stealt...

i think its actually the opposite. Using KS after vanish instead of the initial stealth allows you to get a finisher into both MA windows (Ambush > flurry > build > finish first window; ambush > KS > Finisher 2nd MA window) while if you KS in the first window you are forced to go Ambush > flurry > KS. I remember I simmed this and send it to you and it was a gain, just not big enough to implement

knotty oriole
#

Well that depends a bit on if you are opening with Ambush instead of Dispatch I guess, but that's a little more fluid in a real dungeon. With MA you wouldn't want to Ambush on the Vanish if opening with KS though, you'd Vanish into Dispatch

rough nova
#

side by side comparison of main/mangle vs lethal poisons only places it at +/- >1% what do i trust? wowhead or my own simcs

#

i prefer the synergy of maim/mangle and the fact if i forget to apply my poisons im not punished as badly, don't say you haven't, you ain't a rogue unless you been half way through a fight looking over your numbers and why you doing so badly only to realize you forgot to coat your blades with poison 🙂

civic mango
#

Not so much a TC question as it is a spec question. The TC channels are mainly to optimize APLs or simc implementation. However, sims are specific to YOUR character, and wowhead (and any guide) recommendations are based on generalized sims, so it's always better to use sims of your own character to make gearing decisions. If you feel like something is not simming properly, then that's what you can come to this channel with questions about 😄

rough nova
#

oh nice

#

well on that note

#

i love your Docker work

#

and running parallelized instances of Simulacraft help me crunch massive work loads

#

🙂

#

you guys should import logs into Splunk or and ELK stack and make some pretty images with graphite 🙂

#

@civic mango do any of you guys work actively on SimC dev work I'd love to get involved

#

@civic mango i actually wanted to help you guys with running docker containers and making reports and computations run quicker

civic mango
#

If you're interested in contributing to the upkeep of specifically the rogue APL, then most everyone involved in that hangs out in this discord. If you want to do backend support for simcraft or raidbots themselves, then the best starting point would probably be the specific simc or raidbots discord servers

knotty oriole
golden briar
#

i'd say >50% but ok 😄

regal agate
#

@rough nova koji and mystler work actively on simc, if you want to help just ask in the simc discord, help is always welcome. invite: https://discord.gg/tFR2uvK

young steppe
#

I did some simming with IQD and noticed it's not being synced with Vendetta in fights in which we wouldn't lose a use (5 min for example). Not sure how useful info this is but just forcing it to sync is an upgrade on 5min sims

unknown.png
granite fossil
#

is there a way to change how many people are hit with LBE?

noble pendant
#

shadowlands.lead_by_example_nearby=x defaults to 4 normally, and defaults to 2 in dslice

granite fossil
#

would i put that in an advance sim or somewhere in the desktop app

noble pendant
#

advanced

granite fossil
#

thanks. still kinda new to manipulating sims. appreciate it

noble pendant
#

has everything you need

mystic jolt
#

Is there any threshold for posting optimizations here (i.e. if it's a negligible/low increase but still an increase don't bother)

knotty oriole
#

I mean feel free to point out anything that pops up at high precision sims

noble pendant
#

^

knotty oriole
#

Sometimes we might say it's too small for the effort but can just let us sort that out

noble pendant
#

as long as it's out of margin of error and not ridiculously complicated or convoluted

knotty oriole
#

I mean if it involves a lot of conditions for 0.1% probably not gonna bother but

#

It just depends

#

Related topic, @gleaming fern is gonna be so happy with my SBS Assassination optimizations last night.

mystic jolt
#

Trying to look for more stuff but I doubt I'll find much

knotty oriole
#

SBS is tricky

gleaming fern
#

Will I tho? monkaHmm

knotty oriole
#

Having wasted (YES WASTED FFFFFFF STUPID ASSASSINATION DSLICE/HAC SIMS) 4 hours on this crap last night

#

You better

gleaming fern
#

😄

#

So whats the result?

knotty oriole
#

It was like +2% on DSlice and +7% on HAC lol

gleaming fern
#

Same as before?

#

Hey

#

That sounds worth it!

knotty oriole
#

+0.3% on ST

gleaming fern
#

!!!

#

Is it LBE sbs improvements or just general SBS stuff?

knotty oriole
#

For Lead from Example anyway

#

Syncing with Vendetta

gleaming fern
#

😮

#

yuge gains

knotty oriole
#

The AoE stuff is a combination of some rule changes

#

About MA and targeting priority

#

Allowing SBS to be cast during MA on targets without DoTs yet

#

And allowing SBS to be cast in the last global of MA if using a generator

#

And burning

#

Also prioritizing putting it on the shorter-lived targets more specifically

#

Since this involves MA windows and all kinds of stuff, there were a lot of false positive conditions and it was hard to work everything out

#

That's why it took so long to try to find the "true" rules that worked

mystic jolt
#

is there some condition to check how many combo points you'll get from an sbs cast in the apl?

knotty oriole
#

Since there was some stuff that worked accidentally

#

cp_gain will work

#

On the SBS line

#

e.g. if cp_gain>4

mystic jolt
#

So to check to make sure I won't overcap, it would be cp_gain<combo_points.deficit?

knotty oriole
#

<=

mystic jolt
#

yea was just about to correct that haha

knotty oriole
#

Anyway @gleaming fern yeah mostly it was just stuff that changed with 9.0.5 since the damage of SBS itself makes it more valuable now to prioritize resets (by putting on shorter rather than longer-living targets) and also allow MA in some situations. Still doesn't always beat Mutilate during Vanish windows which was the main hiccup. That's what took a while to figure out. Assassination CP gen/vanish windows are tricky.

#

So ended up going with the final global only, but it still seemed better out of stealth to put on targets without SBS up yet

# 
# Apply SBS to all targets without a debuff as priority, preferring targets dying sooner after the primary target
actions.direct+=/serrated_bone_spike,if=variable.use_filler&!dot.serrated_bone_spike_dot.ticking
actions.direct+=/serrated_bone_spike,target_if=min:target.time_to_die+(dot.serrated_bone_spike_dot.ticking*600),if=variable.use_filler&!dot.serrated_bone_spike_dot.ticking
# When MA is not at high duration, use SBS to apply Lead by Example during Vendetta, otherwise keep from capping charges
actions.direct+=/serrated_bone_spike,if=variable.use_filler&master_assassin_remains<0.8&(fight_remains<=5|cooldown.serrated_bone_spike.charges_fractional>=2.75|soulbind.lead_by_example.enabled&!buff.lead_by_example.up&debuff.vendetta.up)
#

Was the final version

#

Which isn't too complex in the end.. I just had to try about 100 variants to get there 😄

#

Still might be some stuff hiding here but this seemed a decent enough place to stop for now

#

Gains apply for both MA talent and MA legendary so it may lead to some slightly different setups, hard to say

#

Didn't sim top gear combos or anything with it

gleaming fern
#

Mmmmh interesting stuff for sure

wet mountain
#

How much gain is that @knotty oriole ?

maiden cargo
mystic jolt
#

actions+=/variable,name=energy_regen_combined,value=energy.regen+poisoned_bleeds*8%(2*spell_haste) is the % here a mod operator or a percentage (assuming the former)? and what is it for?

noble pendant
#

Division

mystic jolt
#

oh what

noble pendant
#

The / character is already used already as a break

#

So % takes its place as a divisor

#

That line specifically calculates the energy regen of venomous wounds

mystic jolt
knotty oriole
#

energy+variable.energy_regen_combined<5 seems like it would never get hit

mystic jolt
#

whoops that should be 50

noble pendant
#

Yeah I was a little questionable on the energy thing

#

Doesn’t seem necessary considering the energy conditions on the generator variable

knotty oriole
#

I think this is functionally quite similar to something I was looking at earlier which was just to use during vendetta

noble pendant
#

But low cost generator during shiv makes sense

#

Very efficient cp

knotty oriole
#

Which is a similar 0.2% gain and I'm fairly sure that's where most of this comes from

#

Is just forcing it into the opener

#

Need to come up with a way to make it not bad in aoe though

noble pendant
#

I haven’t seen gains from forcing it into the opener

#

Maybe I clowned my tests though

#

Very possible

knotty oriole
#

actions.direct+=/serrated_bone_spike,if=master_assassin_remains<0.8&debuff.vendetta.up&dot.mutilated_flesh.remains>2
was a similar gain

noble pendant
knotty oriole
#

Problem is burning charges before raid_events.add.in blah

noble pendant
#

Oh yeah

#

Would need to figure out some timing nonsense

#

Yuck

#

I call not it

mystic jolt
#

I'm just looking for small stuff 🙂

noble pendant
mystic jolt
#

Yea, it's been fun trying to find little niches here and there while also figuring out apl code nuances

#

Is there a way to make a function?

knotty oriole
#

Not really

mystic jolt
#

Darn

knotty oriole
#

One thing probably need to look into is how it fares with EP since Blindside effectively buffs Mutilate quite a bit

#

Seems possible some non-Blindside specific stuff could be worth looking at

plain coral
knotty oriole
#

fwiw this variable is getting replaced

#

with energy.regen_combined

#

and doing it in code

#

So that it can approximate stuff like Dashing correctly

#

I added it yesterday and using it for the new priority rotation logic of avoiding cycle DoTs at energy.regen_combined>35

#

Which seems to be roughly the magic number atm

lean talon
#

The APL currently casts KS during Flagellation, which devalues KS a lot, especially on pure AoE sims (Dancing Steel simming higher than KS with Celerity on short AoE simsfor example (https://www.raidbots.com/simbot/report/vjAYYev4ucpf2xHGY47SQ9 ) ). I've found a fix for this for those pure AoE sims, however it ends up being a loss on dslice and I can't find a way to fix it for both.

https://www.raidbots.com/simbot/report/iBH8wgN1aLWLywa23tFFsU (90 second 5 targets with Celerity)
https://www.raidbots.com/simbot/report/eV1ABqsh76Cbh8MQsBfTUD (90 second 5 targets with MA)

https://www.raidbots.com/simbot/report/8Hxa46NtMtkxwBzbyamck6 (dslice celerity)

knotty oriole
lean talon
knotty oriole
#

I just commented out the Covenant line on that same profile and it was -1.8% or something

lean talon
#

hm. i mean still seems wrong to go flagellation > killing spree, and its a decent gain for that sim too

knotty oriole
#

Wonder if maybe there's something more restrictive.. like only during flag if at low energy or something

#

It's too bad KS channel duration is not hasted 😦

lean talon
#

true, though i think during the haste buff its fine its more so that you waste 2 seconds during the debuff where you dont generate any CP

knotty oriole
#

Yeah that's why I wonder if an energy condition might help

verbal latch
#

Not that Im jazzing good but you would need energy to get cps and finisher, so if you were low on energy, KS would be ideal as you have twos to get energy back to build up the stack but what do I know, nothing.

lean talon
#

I just took another look at the prio of Slice and Dice and between the eyes and atleast for Necrolord, using BtE before SnD is a gain again. I remember @knotty oriole theorized it was mostly a Kyrian thing to prioritize SnD so should probably double check if that is still the case.
Forcing BtE at 6 CP is also small gain. I also noticed that the APL loses some BtE casts by refreshing SnD as soon as it enters pandemic, even if BtE is up and it could cast BtE first and refresh after. I put in a condition to attempt fix that, too (slice_and_dice,if=buff.slice_and_dice.remains<fight_remains&refreshable&!cooldown.between_the_eyes.ready).

https://www.raidbots.com/simbot/report/pZi3gFi4mWDkDQH63mdmZJ

knotty oriole
atomic wyvern
#

There isn't any new research. This would be a question for #assassination, not here. Mastery is just bad for sin outside of aoe situations no matter what legendary you choose, it would take some borrowed power effect or a huge buff to mastery to change that

knotty oriole
#

Mastery has no interaction with Doomblade at all

#

I mean sure

#

But it doesn't modify the Doomblade DoT at all and doesn't really change anything rotationally

lean talon
knotty oriole
knotty oriole
#

actions+=/variable,name=finish_condition,op=reset,if=cooldown.between_the_eyes.ready&effective_combo_points<5
With also doing the SnD/BtE flip in the finish list

#

This seems to be a global gain from both of the things you observed at the same time so think that's likely what I will put in

lean talon
knotty oriole
maiden cargo
#

i noticed, for outlaw, if using master assassin and if marked for death is talented, the sim will always vanish > dispatch

#

i'm confused at what that has to do with marked for death being talented

#

rather than marked for death being off cooldown or something

#

my attempts to change it were a loss so it clearly seems correct, just wondering why tho

spiral light
#

Further optimization can be achieved most likely by moving the MFD after the Envenom in Vanish during the Shiv window (and possibly syncing MFD and Shiv), but that's still being tested.

noble pendant
#

good find, I'm currently taking a look at some other mfd things, will add on

#

should be noted at least that the line removal is a significant loss for necro at first glance

#

also would need some changes for dungeonslice considering it has a prep for add events

merry cloud
#

I had some work done on aligning shiv and mfd

noble pendant
#

might be an edge case kind of thing

merry cloud
#

it was dps neutral or small gain, at least seems like a viable strategy

spiral light
merry cloud
noble pendant
#

yeah that handles basically everything I was checking lul

knotty oriole
#

So what was the thing with the prepull thing for MfD? Was that only Venthyr?

spiral light
#

I'm still trying some stuff out, ie. holding MfD for Flag for certain scenarios. Edit: Holding MfD for Flag on pulls where you use MfD in pre-pull or with movement is a marginal gain, nothing major.

knotty oriole
#

But also @merry cloud since you also removed it from the precombat, your change is about +0.2% or so then?

#

Since your profile is also Venthyr

merry cloud
#

I presume so

knotty oriole
#

Gonna try yours with my Necro profile

#

Looks like -0.3% with the full profile for Necro, -0.2% with Precombat added back in

#

So I think this is Flag-specific interactions most likely

noble pendant
#

agree

spiral light
#

Most definitely.

knotty oriole
#

Which is fine, that makes a little more logical sense I think

#

Considering the mechanics

#

Also, I think Lorentz your APL version doesn't have some of the SBS APL optimizations recently in it, so that's changing the opener sequence slightly. I'll need to check to see if that impacts

noble pendant
knotty oriole
#

Seems to make it even worse for me.. like -0.6% 😄

noble pendant
#

ouch

knotty oriole
#

MfD is already like 1.8% behind for Necro though so I suspect there's no saving it anyway 😄

random fog
#

Rogues can use bows. Do you think that we will have SWP 2.0?

unknown.png
tall mantle
#

Well data says it costs focus. Not sure where we would get that.

noble pendant
#

Equipping bows also disables all of our main abilities, sin cant mutilate, sub cant shadowstrike, outlaw wouldn’t be able to dispatch

#

So unless that ability magically does like 5k dps

#

Doubtful you would run it

#

And if it did do that much, they would nerf it very quickly

unique moat
#

and 5k dps in 9.1 gear will be shit tier dps most likely

knotty oriole
#

As mentioned, this ability costs Focus so no, we can't use it. It's not really worth talking about further.

lavish zenith
#

can someone tell me qualitatively how ss at 4 with skull is more worth than finish at 4? if u ss at 4, u have a 65% chance to waste a cp. if u finish at 4, u have a 20% chance to waste a cp. so, the intuition is 1 more cp dispatch provides more damage than 0.45 wasted cp and whatever wasted energy that entails right? because when it comes to skull + broad, and shift the convo to ss once more at 3 or finish on 3, u have a 65% chance to waste 2 cps vs 40% chance to waste 1 cp. so if the ss on 4 w skull is remotely close on the sims, i feel like logically it could sway the other way to favor finish on 3 seeing as the stakes are worse w skull + broad. can any og tc'ers offer thoughts on this? also mainly curious if have we data from specifically simming finishing on 3 w skull + broad in 5t cleave?

maiden cargo
#

Pretty hard to answer that other than let the sim take the wheel with determining what's optimal for damage, energy, and restless blades efficiency

#

But there's so many factors involved I don't think you can "napkin math" the reason why

knotty oriole
lost echo
#

because finishing at 5 gives you the 1 cp guarunteed, but finishing at 4 has a 20% chance to not give you that cp

#

i forget what the passive is called

knotty oriole
#

Hmm. I am not really sure I would call that "waste" though

lavish zenith
#

yeah that's what i was referring to

knotty oriole
#

That is just a refund conversion rate, over time that will still come out to be the same ratio of CP you "put into it"

lavish zenith
#

I suppose that's true yeah

#

didn't really think abt it that way

lost echo
#

wouldn't it be a different ratio to the energy you put into it tho

knotty oriole
#

Yes, finishing at lower CP is really only energy inefficiency

lost echo
#

since a 4cp dispatch and 5cp dispatch cost the same amount of energy

#

ye

knotty oriole
#

Or sometimes cooldown window inefficiency (say for MA)

civic mango
knotty oriole
#

I mean, somewhat but I think energy and GCD inefficiency are just two opposing sides of the same coin

#

If you are Energy capping it is GCD inefficiency but the Energy won't matter

#

If you aren't Energy capping it is Energy inefficient but the GCD doesn't matter

lavish zenith
#

So, couldn't find anything abt this on the apl. if u are playing ma/ks on ST, ks supposedly only worth if u ks for energy regen. if u ks at full energy on st, it's a loss (as far as i understand from what i've read from community figures here). so suppose ur going into a boss in m+ w/ ar+lust. u obv want to vanish for the damage, but my question is, should i killing spree for energy regen from like 100-145 and cancelaura it right before i hit max energy. because i for sure hit max energy before even half of the killing spree is done. and so in my head, i've had a dps gain for half of the ks, and a dps loss for the second half where i was at max energy. this line of thinking correct or wrong?

#

because i've never seen other rogues cancelaura their ks like 2 hits into the ks, despite hitting max energy. i started doing this a few days ago but figure i should ask here just to confirm it's in fact a dps gain and i'm not an idiot for cancelling my ks like 2/5 of the way thru lol

lean talon
#

without MA its just a 10 dps loss to send it on CD versus following an energy rule. so basically can just use it on CD without MA too

lavish zenith
#

hm ok but guy says it's extremely bad unless ur using it while regenning energy. i mean he said "very ppl know how to use it properly on st", so i figured he was alluding to this. 10 dps loss to just send it on st doesn't really seem like it's that hard of a rule to follow

#

is it a meaningful difference if i use 2 seconds of ma to do the ks vs using the ma window for normal globals and using the ks after the ma window?

#

ie (regardless of them getting temporarily desynced) it's better to use ma for amb bonespike disp mfd bte or something vs amb ks disp

lean talon
#

yea its just 10 dps, and thats WITHOUT ma, with MA you wanna use KS on CD

lavish zenith
#

cuz i feel like using the ma window for non ks on st is good so long as u sync up ur next ma ks for aoe after the boss

lean talon
#

you mean intentionally desyncing KS from MA? i dont really see the point, and im pretty sure KS is our ability with the highest damage per execution time so

lavish zenith
#

ok ok just checking

#

was thinking amb bs disp mfd bte and then ks after would be more dam than amb ks bte or something along those lines (disp ks mfd bte etc)

lean talon
#

looks like dispatch and bte are slightly more damage per global cooldown than KS, ambush and bonespike are not though. However killing spree also makes you proc more main gauche and instant poison during the 2 seconds which is not taken into account in that

knotty oriole
#

Using KS during the MA window at some point is generally optimal. As to if you do that from the stealth opener or vanish it's up to you. That doesn't matter. But mid-fight Vanish should always be synced with KS. It's a pretty big gain to do that.

#

Really the main reason KS is able to be competitive and beat Blade Rush for DS sims

#

Without MA and syncing, Blade Rush is probably just better

lean talon
#

you mean on ST? cuz on aoe KS is just always better

#

oh you said dungeon slice. KS wins without MA too

knotty oriole
#

I mean depends on what you mean by wins

#

If KS is even remotely close to Blade Rush damage, Blade Rush is just better

#

Blade Rush doesn't get you killed and has potential on larger uncapped pulls

#

KS has to effectively be much better than Blade Rush to actually be usable

#

imo

lean talon
#

its 1.5% better without MA for me. whether thats enough to take the higher risk you have do decide for yourself

knotty oriole
#

Sure, but I think my point is that KS gets almost double that advantage from MA syncing over Blade Rush vs. using Celerity so it's a non-trivial difference in the relative value

#

If you aren't syncing KS with MA you basically lose half the gain over Blade Rush

#

Running KS for +3-3.5% probably makes a lot more sense than running it for +1-1.5% imo

#

But regardless, syncing with MA is pretty large for KS

lavish zenith
#

yeah true. and even still i'm probably taking br. i'm just trying to think of some stuff ahead of time for the great push

#

my team just hates ks and like if i die even one time from ks over the course of 3 dungeons then it wasn't worth it

knotty oriole
#

KS DPET is definitely the highest overall

lean talon
#

well this was about single target

lavish zenith
#

well we were talking specifically st here

#

no aoe considerations at all

#

that's why syncing wasn't really important here to me

knotty oriole
#

Well you still sync MA even in ST

#

Definitely the MG/Poison procs during KS make it worth compared to other abilities during MA

lavish zenith
#

wel yeah that was my question. which takes precedence. syncing ma and having ks active while at 100 energy, syncing ma and canceling ks the second i get to 95 energy, or just sending ma and saving ks for when it would be fully used for energy regen.

knotty oriole
#

The other rule about using it in ST is that you always use it when BtE is up

#

If you aren't running MA

#

Current APL threshold for triggering KS in ST is energy.regen*2+15 so basically 2 seconds of whatever your passive regen rate + 15 cushion. Haven't really explored canceling but I think it'd probably be of pretty minor value

#

But keep in mind the APL totally ignores the energy condition when MA is active

#

If you are running MA, you don't care about energy at all

lean talon
#

i tried making it cancel but i dont think it worked. just scrapping the energy rule altogether was only a loss of 10 dps though so i doubt cancelling would be worth it really

knotty oriole
#

Energy rule only gets used for ST and only without MA

lean talon
#

yes

#

thats what i simmed

lavish zenith
#

ok so ks actually is braindead on st

#

i thought there was something special i was missing since it was said to be hard to use in st

lean talon
#

actually it was only 3 dps

#

so the energy rule literally does nothing

knotty oriole
#

Pretty sure it's just because it's not really getting hit, not because it does nothing

lean talon
#

its 0.7 more killing spree casts if you remove it

#

either way the rule doesnt help for anything and just confuses people when they think they should delay KS during lust on pull, or have to think about energy in some way imo

knotty oriole
#

Totally depends on your haste and energy regen in general as to how much of a gain it is. It's 10-15 DPS for me.

#

I've seen you fight over less 😉

lavish zenith
#

can u guys do me a fav ❤️

#

can u swap it to 2 min, lust, potion, no monk buff, no war buff, no wf, double sharp? <33

#

want to see if it is worth more like 30 w those conditions

#

cuz then that's actually starting to get sizeable

lean talon
#

this is all assuming no MA james

lavish zenith
#

ohhhhhok

#

ah if i'm not running ma i'm never playing ks

lean talon
#

if you are running MA you just wanna vanish and KS on CD always no matter your energy

lavish zenith
#

ok

#

so it's free asf

#

ty boys

atomic wyvern
#

Generally we don't put rotations that require long pull timers into the APL because slowing your raid down for a marginal dps increase for you isn't worth it

upbeat sandal
noble pendant
#

it is a gain to have a stupid long pull time and snd pre-pull, and some logs already do that with critters to log

#

obviously like osh said, things like that don't go into the APL

#

and openers in general don't really see much difference in dps from moving like 2-3 gcds around

upbeat sandal
noble pendant
#

hmm yeah, that's what I said

upbeat sandal
#

me, as a spanish native speaker don't understand everything as good as other people haha but that's all, no off topic

knotty oriole
#

Opener optimizations like this are still extremely small in terms of total DPS

#

All you are "wasting" in this case is 2-3 CP and a little energy. Over an entire pull that averages out to probably less than 5-10 DPS or something.

#

Not only does this require a large prepull timer but it also requires a tome on every single pull, which is just not realistic for all but super extreme cases

solar swift
knotty oriole
#

The energy condition is just to match the logic for when Envenom will be used

#

So it serves no purpose without the Shiv check

#

I'm gonna need to double check these profiles because I'm finding it a bit skeptical that CT is worth using in ST over Envenom in ST and Doomblade within Shiv at all gear levels. This was checked quite a bit already before.

atomic wyvern
#

A couple of us in the assassination channel did some sims with using CT in ST and found that it was the same dps to use it and to not use it, as necrolord at least. Someone who was venthyr had different results

solar swift
#

mb its only for venth, coz flag

atomic wyvern
solar swift
#

it's bs+ma build

knotty oriole
#

If it's Venthyr, that makes sense to some degree. I'm going to guess it's a Haste threshold.

solar swift
civic mango
#

Not exactly a TC thing, but were we aware that dodging anything causes a main gauche proc, not just with evasion active? I just tested it multiple times after MGing something when i dodged while riding past mounted

unknown.png
remote granite
glass rune
#

Wasn’t this added to compensate for Riposte being removed

civic mango
#

the thing is, it procs the mastery hit 100% of the time

#

any melee dodge procs a MG hit

remote granite
#

Yeah, the wording implies its only supposed to proc during evasion and not outside

#

the bug here would be that it procs on any dodge

willow knoll
#

Lol

ruby crag
#

does changing soulbind on raidbots change covenant ability too ?

#

like can I sim venthry without swaping

regal agate
#

it would not make sense to not change covenant ability too.
This is also the wrong channel for this kind of questions.

ruby crag
#

sorry about it

crisp granite
#

Hey maybe it has been tested already, but the GS rule in the apl is to use PS only if you need energy, or if you just need 1 cp, wouldn't removing the needing one cp rule for guile charm be a gain, as more sinister strikes makes the buff build up faster

civic mango
#

You could make an APL that does that and see :))

crisp granite
civic mango
#

Copy paste the apl line for pistol shot for me here? I'm at work on my phone

crisp granite
stiff roost
#

when using zoldycks it appears to be a dps gain of about 20 on ST patchwerk (50 in execute patchwerk) to remove the !shiv and energy deficit restrictions on CT. This means that its worth keeping the CT buff up all the time even in shiv windows. due to the reduced gain on normal patchwek this may only be a gain when actually in execute phase. Sims done using my character so maybe worth investigating more with other setups

noble pendant
#

It's actually specifically a zoldyck AND venthyr thing from what I've seen looking into it

knotty oriole
#

More likely a haste thing, but that is mostly the same difference at this gear level

#

That being said I would highly advise running comparison sims in the same advanced sim because otherwise the error margins are difficult to compute to compare results

noble pendant
#

yeah I'm redoing the sim in one

#

it's margin for other covenants to be fair

knotty oriole
#

Probably the gain is restricted to when the Venthyr haste buff is up if I were to guess

noble pendant
#

yep

#

it's a significant loss for non-zold setups as well

knotty oriole
#

Plus the fact that Zold doesn't work on Envenom for reasons defying understanding

noble pendant
#

this is a doomblade necro for example

knotty oriole
#

Obviously in execute the DPET valuation shifts a bit due to Envenom not getting the execute bonus

noble pendant
#

yeah

#

I think the gain is specifically isolated to execute during venthyr haste buff

#

doomblade venthyr shows a loss as well

#

I remember looking into it as well, actually as an attempt to see if we can get rid of CT single target

#

and I found the same stuff with zold/venth

#

might be worth revisiting with gear from next tier

knotty oriole
#

Seems like adding |runeforge.zoldyck_insignia&target.health.pct<30

#

is essentially equal

#

So yeah, seems mostly restricted to that case

#

But looks like just flag buff is +0.3% even without Zold, just execute is about +0.1-0.2% without Venthyr but with Zold.

#

So really mostly the flag buff though

#

Which basically just is some sort of haste DPET threshold

noble pendant
#

Some optimizations for AoE sims regarding dying enemies for assassination. There's 2 main changes, one regarding rupture casts on dying targets, and the other involves casting CT slightly before pandemic in aoe. Here are some example setups. The gains vary from about 0.8% to 1.3%. Legendary/Covenant/Talents don't have too much of an impact, but obviously if you're not running CT, half of this gain won't affect you. It is completely neutral on 1-3 targets, and then starts having very minor (0.2-0.3%) gains in 4+ sustained aoe. Other combinations are left as an exercise for the reader, but have been tested :)

Dslice https://www.raidbots.com/simbot/report/qdAJWYHtCD5J3WVrtjhW9m
HAC https://www.raidbots.com/simbot/report/mPCskSiaFFn9s3CoNqLTfA

civic mango
#

if i wanted to reduce the number of mfd casts in a dslice sim because i felt like it was getting an unreasonably high number of casts within the time frame, what would be some good methods to do that

#

I obviously don't want to reduce the mob counts

#

perhaps adding a minimum time to die requirement? Such as if mfd becomes available when a target is about to die, it must have a ttd greater than .5 seconds, or maybe even as high as 1 second, in order for it to be eligible for sniping

noble pendant
#

time to die seems reasonable enough

civic mango
#

marked_for_death,line_cd=1.5,target_if=min:target.time_to_die&target.time_to_die>.5,if=raid_event.adds.up&(target.time_to_die<combo_points.deficit|!stealthed.rogue&combo_points.deficit>=cp_max_spend-1)

#

?

noble pendant
#

should be able to change the time to die to like 3

#

and it would reduce the casts significantly I would think

civic mango
#

marked_for_death,line_cd=1.5,target_if=min:target.time_to_die,if=raid_event.adds.up&(target.time_to_die<combo_points.deficit|!stealthed.rogue&combo_points.deficit>=cp_max_spend-1)

#

This is what's in the APL

#

the 1st line i linked is my alter

noble pendant
#

oh

#

I don't think 0.5 would do much tbh

civic mango
#

Perhaps, i can try it and see, i'm mostly asking if that edit would have the desired effect

noble pendant
#

I think it would

civic mango
#

but i guess, again, i can just try it

noble pendant
#

changing MFD around is always pretty fickle from my experience

civic mango
#

I think i found it

#

yea i did

#

even if i copy the APL exactly as it is in the default profile it still has an issue with that line lol

noble pendant
#

that's odd

civic mango
#

direct APL copy

noble pendant
#

double if statements in one line makes me nervous

#

escalate it to @knotty oriole

knotty oriole
#

The line_cd was already added for the purpose to prevent “spamming” of MfD

#

You can just increase that if you want

civic mango
#

and that line basically gives mfd a 1.5 second icd?

knotty oriole
#

Yes basically

civic mango
#

any idea why advanced sim doesn't like that line?

#

and ignores it

knotty oriole
#

It doesn't stealth because that is a patchwerk fight style not ds

#

Fight Style: Patchwerk

civic mango
#

i fixed the stealth part, i didn't have the invulnerable raid events

knotty oriole
#

Has the raid events added but that doesn't add the out of combat time

civic mango
#

marked_for_death,line_cd=1.5,target_if=min:target.time_to_die,if=raid_event.adds.up&(target.time_to_die<combo_points.deficit|!stealthed.rogue&combo_points.deficit>=cp_max_spend-1)

#

This line gets ignored

knotty oriole
#

I'm not sure I follow

#

Sounds like you're not adding it to the action list

civic mango
#

somehow i deleted that

noble pendant
#

kekW oh man I completely missed that

civic mango
#

yea me too

knotty oriole
#

That's why they pay me the big bucks!

civic mango
#

ok I must be completely missing something else, because now it's just running two copies with the default APL, instead of using my copy with the modified one

lavish zenith
#

i don't really know if this merits any kind of apl change, but i noticed at least once or twice per dungeon (in the big pull dungeons like halls, there would be instances where mobs would be dying at sufficiently fast pace such that if i were doing sbs -> 6cp disp -> sbs -> 6cp disp, i would still overcap charges on my bonespike. like i would be wasting charges. i noticed that 1 sbs (3 instances of damage, all flurrying onto 5 targets, was significantly more damage than 1 6cp dispatch flurrying onto 5 targets), so on these certain pulls where i could predict based on mob hp that if i did sbs -> 6cp disp like normal, i'd be wasting charges, i changed to just spamming sbs and overcapping cps instead of wasting my sbs charges, so long as i didn't waste too many and could make sure i'd be going into the next pack w 5 charges. this happened much more frequently w spiteful. also, i don't know the specifics of how the volcanic plumage anima power works, but i noticed i had significantly more volcanic plumage damage than my teammates (im assuming it works w crit and vers, but also if it's proccing off of instances of damage, you would be upping the probability to proc this thing by opting for more applications of sbs vs. just dispatching). i'm not sure mathematically which is better, this was just my intuitive sense of looking at the damage values (to overcap cps and not waste sbs charges when flurrying onto 5t, or to disp for cdr and waste sbs charges), but i thought i'd mention it here for anyone that's interested that knows what to do w this information.

knotty oriole
#

Looked like Volcanic Plumage is just 4 RPPM so I don't think it should matter too much what you do, although it'd be affected by Celerity/BtE/Guile/etc.

#

Have spent a fair amount of time trying to do SBS optimization. Sniping (low duration prio), dumping, reverse sniping (high duration prio) and largely it didn't seem to make any notable difference. Spent a lot of time swapping logic around without really getting anywhere.

#

Had some luck for Assassination but Outlaw really seemed very neutral with the different approaches

lavish zenith
#

alright yeah just thought i'd mention. cuz 6cp disp w flurry on 5t is like notably less dam than 3 sbs all flurrying onto 5t. so that's the only reason i mention it

#

like sometimes yesterday i could spam sbs 4 times in a row just back to back over and over and still have 5 charges of bonespike when i finished all the spamming lol

knotty oriole
#

Which legendary were you using?

lavish zenith
#

deathspike

#

that's the whole point like this only applies to deathspike

#

basically 1 global w deathspike on sbs is a lot more dam than the 1 global being spent on dispatch (assuming u throw out 3 spikes and all 3 spikes are flurrying onto 5t) but u don't get the cdr, so at times yesterday when i had the decision, ok do i sbs 6 times in a row, or do i sbs disp sbs disp sbs disp and waste 3 charges of sbs, intuitively it felt like more damage to go w the sbs and just overcap cps

#

because on certain pulls in halls, u can feasibly get ur bonespike on like 12 mobs at once, and when they all start dying u can overcap ur sbs charges really easily if u don't just spam sbs

knotty oriole
#

Yeah, Deathspike is definitely something we'll need APL logic for but I haven't put in anything yet. I think you're correct that there is likely a point where just casting SBS is better than worrying about charges, especially since Deathspike effectively multiplies charges for you (since it generates more potential refunds than it costs) compared to raw casting SBS which is just 1:1

lean talon
knotty oriole
#

I'd kinda prefer to get whatever we have in as a baseline rather than reinvent the wheel. Blizzard really can't balance detailed enough to care about 150 dps or whatever

lean talon
knotty oriole
#

Especially outside of beta

#

Ok, so this is specifically for Deathspike I assume given the charges_fractional stuff?

lean talon
#

I originally split the sbs conditions into seperate lines to track how often which condition is being used but it ended up being a huge difference vs the same conditions in the same line split with |

#

Yes but the fractional was also changed in the 8911 simming one

#

Its the "if more than 2 targets and flurry up, cast bonespike regardless of charges"

knotty oriole
#

I don't really see how splitting the lines would matter other than for the !dot.ticking line

#

That is the only one that evaluates on the target

#

That one was split for the Assassination APL for that reason

#

I was going to put something similar in for Outlaw but there was no immediate gain. However if you're adding a more aggressive dumping condition it makes sense to split them to ensure things go on the correct targets as a priority

lean talon
#

I can only go off what i said back then, dont have the sims anymore 😄

Screenshot_20210604-205129_Discord.jpg
knotty oriole
# 
actions.direct+=/serrated_bone_spike,if=variable.use_filler&!dot.serrated_bone_spike_dot.ticking
actions.direct+=/serrated_bone_spike,target_if=min:target.time_to_die+(dot.serrated_bone_spike_dot.ticking*600),if=variable.use_filler&!dot.serrated_bone_spike_dot.ticking
#

for Assassination for example

#

Before the general dump line

# 
actions.direct+=/serrated_bone_spike,if=variable.use_filler&master_assassin_remains<0.8&(fight_remains<=5|cooldown.serrated_bone_spike.charges_fractional>=2.75|soulbind.lead_by_example.enabled&!buff.lead_by_example.up&debuff.vendetta.up|buff.marrowed_gemstone_enhancement.up)
#

There's no reason to split the other stuff

#

But splitting the ticking thing is a logically different execution order

#

Because you are doing a "first pass" to put stuff on non-ticking targets

#

Then a second pass if that fails for the dump conditions

#

Otherwise the dump criteria could be met on the first target even when the 2nd target doesn't have the dot, for example

lean talon
#

Yea i have no idea why it ended up only being a gain when i split it, was more an accidental find

knotty oriole
#

I'd probably have just used this one by itself

actions.direct+=/serrated_bone_spike,target_if=min:target.time_to_die+(dot.serrated_bone_spike_dot.ticking*600),if=variable.use_filler&!dot.serrated_bone_spike_dot.ticking
But Assassination cares about SBS being on the primary-primary target first because of Doomblade legendary

#

But I guess the only substantial change in your APL is basically "dump down to 1 charge if Blade Flurry is up and 3+ targets"?

lean talon
#

Yea

knotty oriole
#

Makes sense to me

#

But yeah would want to split the ticking lines just to make sure you spread them first

lean talon
#

I guess its kinda alleviated for deathspike since it still hits 2 off targets (and i assume it smart targets in simc?), but yea should definitely still target mobs without debuff whenever possible

knotty oriole
#

Probably the charges_fractional thing should just be inverted to be relative to max charges

#

To support both

spiral light
#

I always thought it was weird how the sim suggests to use SnD at 3+ combo points for Assassination, even though you can maintain the buff even with 1CP. I've gotten the question "Why does the sim use Mutilate after SBS before Slice" in #assassination a few times, so I decided to look into it. In fact, in my own play I always use SnD after SBS even at 1CP and it has never dropped for me, so I didn't see any logical reason for building to 3+ CP.

After playing with it a bit, I found it to be a slight gain to cast it at >=1CP instead of >=3CP on 1-2 targets on Necrolord specifically. I'm not sure why the apl was finding gains for using SnD at 3+ CP previously, but I couldn't find any gains. The gain is further inflated by short fight lengths.

I tried simming most combinations, including Venthyr/Necrolord with various talents, target counts, fight lengths etc.

On DungeonSlice with the usual setup (Zoldyck, Necrolord, EP/Subt/Vig/CT) it's +-0. For 3+ targets it's a very marginal gain, but never a loss.
I also tried simming the apl change as Venthyr, and it was always within margin, and again, never a loss.

I'm aware that it's quite a minor and simple change, but it is consistently performing better on 1-2 targets and is never a loss on other target counts, that's why I decided to bring it up.

Sims are done using my personal character profile.
https://www.raidbots.com/simbot/report/coXtDjwYkSrkbUMLMpHwyc - 5min 1T (0.2% - 11dps)
https://www.raidbots.com/simbot/report/sQz6gFjVFQd49tAyZcXnAZ - 5min 2T (0.2% - 12dps)
https://www.raidbots.com/simbot/report/id9J2FWLSjMFof3Z5JPuFy - 3min 1T (0.3% - 21dps)
https://www.raidbots.com/simbot/report/xyKH4CKHTLVKpAS5mxbHi2 - 3min 2T (0.3% - 20dps)
https://www.raidbots.com/simbot/report/2b5Zpe7N91wbSYfZacn7je - DSlice (+-0)

#

actions+=/slice_and_dice,if=!buff.slice_and_dice.up&combo_points>=1

knotty oriole
#

Well it's worth keeping in mind that the Necrolord opener really isn't the same as it used to be with the changes to the MA and Gemstone conditions. Haste is also different now than what it used to be now that we are better geared, so there's quite a lot of reasons there may be a 0.2% swing. That being said, this was neutral or a small loss with some Night Fae profiles I just tested it against, and many people were Night Fae (rather than Necrolord) at the time this was tested.

#

Multi-target in general uses Envenom quite a bit more aggressively than it used to as well

#

But at this point there is probably no major downside to swapping it over

knotty oriole
spiral light
knotty oriole
#

Yeah np, I'll update this later today

spiral light
#

I would reach my fated 7k sim with this apl change as well, so obviously a very important, personal incentive for me.

#

Great stuff, thanks Koji!

knotty oriole
#

Haha

#

Well at least for buffed sims with WFT, we lost some DPS due to the findings of the Windfury Rank 2 bug a few weeks ago 😦

#

For those fortunate enough to have a Shaman

spiral light
#

Yeah, that was a sad day for Raz sims. sadcat

worthy thorn
solar swift
#

Hi, is it possible to override obedience spell data for 24 sec buff insted of 12 sec?

noble pendant
#

It’s implemented as 24 seconds in sims

solar swift
#

hm, i'll recheck, but seems like same dps as few weeks ago 😦

warped grove
#

it was always implemented as 24 in sims iirc

maiden cargo
#

quick question, what's the difference between energy.time_to_max_combined and energy.time_to_max

knotty oriole
#

For Assassination it includes Dashing and Venomous Wounds per/second estimations

#

Eventually might extend it to include Combat Potency estimates for Outlaw, but that's a bit harder

maiden cargo
#

i see ty

#

i assume effective_combo_points is the same thing, but for say animacharged cp yea?

knotty oriole
#

regen_combined and time_to_max_combined are essentially equal for the other two specs as without, but safe to use either way

#

effective_combo_points basically translates to what the CP value would be if you were to consume the current animacharged CP if it is matching

#

e.g. if you have CP3 charged and you are on that, it will return 7

#

instead of 3

solar swift
#

nightly outlaw build with ptr=1 still doesn't work?

# 
Simulating... ( iterations=100000,
 threads=32, target_error=0.050, max_time=300, vary_combat_length=0.00, optimal_raid=1, fight_style= )
sim_signal_handler: Segmentation fault! Iteration=0 Seed=7680122325395551093 TargetHealth=0```
knotty oriole
regal agate
#

it segfaults due to GC @solar swift

knotty oriole
#

It's already fixed but raidbots builds are paused until he migrates over to 9.1 later

maiden cargo
#

Do you happen to know if Marileth will apply his Humanoid mastery buff when casting Fleshcraft even if you only channel it for say 0.1 seconds

#

Might need an APL condition where you quickly cast fleshcraft but cancel the channel

lavish shuttle
#

on PTR the mastery buff is applied directly on press of fleshcraft

knotty oriole
#

Channel can be canceled early in the APL with a command, but I decided to leave it in initially

#

Because it feels like people won't totally int here most of the time and actually try to get their Fleshcraft absorb so they don't die 😛

maiden cargo
#

fair

knotty oriole
#

But there is a way to do it

maiden cargo
#

ye good to know

knotty oriole
#

interrupt_if=

#

Should work

#

so like fleshcraft,if=x,interrupt_if=buff.volatile_solvent_humanoid.remains>90

#

or something

#

It's implemented so the buff is granted at the start of the Fleshcraft cast in SimC already

#

So that should work

maiden cargo
#

interrupt_if=true

#

😝

knotty oriole
#

lol

#

interrupt_if=ticks>1

#

Would probably be the safest

twilit granite
#

Can you add a custom buff that increases agility by x %?

#

Any work around for that?

#

Wanna compare how agility works vs a flat damage % increase

#

like 10% more agility vs 10% more damage

#

For damage adding a vulnerability event works but no idea about agility

knotty oriole
#

You can add custom procs to gear, which might work.

twilit granite
#

Also not sure how vulnerability interacts with some rogue specs, i guess i am better with damage_done_buff_event_t

twilit granite
knotty oriole
twilit granite
#

or blade flurry

twilit granite
knotty oriole
#

Blade Flurry doesn't double dip from target modifiers

#

So that's not really a concern

twilit granite
#

i.e that vulnerability wouldnt affect the damage that gets flurried

lavish zenith
#

just an idea, but u guys should look at adding 16% crit to the dslice profiles w/ like 70% random uptime. would b more realistic sims. sub and sin probably benefit more than outlaw

noble pendant
#

Where would that crit be coming from?

lavish zenith
#

the seasonal affix

#

the orbs

#

there is a 30% execute one which is pretty solid, but it has to stack up and i didnt find it to be as good as the stat stick except for on ST. was really strong on bosses, esp bosses like stradama etc

#

but i figured that one would be harder to sim since im not exactly sure how it stacks

noble pendant
#

historically affixes haven't been added to sims due to the complexity in downtime etc that it adds, pride mini-bosses weren't added to dslice last season for example

#

but that's probably a question for an actual dev like koji/mystler

lavish zenith
#

well the only reason i thought itd be good is cuz prideful would affect all 3 specs roughly the same, whereas crit really benefits sub and sin and not as much relatively for outlaw

#

like if sub is winning currently in dslice, and u give both specs 16% crit, sub should win by even more (i would think)

#

gotta wait for the 1 handers and sylv daggers to be implemented tho i guess as well

#

i feel like 16% semi permanent stat is also a lot more impactful than like 2 or 3 pridefuls over the course of a 40 minute dungeon

golden briar
knotty oriole
#

It's possible something could be implemented, but I think Navv is investigating trying to parse out the dungeon anima powers still

#

The spell data structure is slightly different from existing stuff

#

Likely if we can get it working there will be some option for triggering them

lavish zenith