KQM | Genshin Impact - sac-procs-on-diona

Ticket Tool BOT Jan 29, 2022 1:59 PM

@Homura - sac procs aoe
As an author, It is your responsibility to complete the ticket
- Take advice constructively
- NO disrespectful behavior
- The ticket will be scrapped if: No activity > 1 week or open for > 1 mo.
- When done, type $close

Theory/Finding/Bug: Title of your submission

Evidence: Explanations with calculations and/or Youtube/Imgur proofs

Significance: Conclusion

Rito Jan 29, 2022 2:00 PM

Rename here

Nice

Homura Jan 29, 2022 2:00 PM

aight

Rito Jan 29, 2022 2:00 PM

Go Get that salmon

Homura Jan 29, 2022 2:01 PM

I don't exactly have video of anything except half my 5th test on diona tho, I do have test results tho

Rito Jan 29, 2022 2:01 PM

Do it based on this

Eh forgot to mention that when your sure that your done type

$close

because it will go into "for review and submission" if you do that

pai Jan 29, 2022 2:07 PM

here's a video to start i guess

Homura Jan 29, 2022 2:08 PM

Finding: Sac Weapon Procs on Enemies hit on the same frame

Evidence: I only have 1 video atm and it's only half a test, but so far I've tested sac bow procs on diona quite a bit and it's always low for some reason I believe it's something related to being able to land all her hits on one target at the same frame. Results so far are:
92/100 r2 hold E in abyss
93/100 r2 hold E in abyss
95/100 r2 hold E in abyss
93/100 r2 hold E in abyss
94/100 r3 hold E in abyss first 50 in this video (https://www.youtube.com/watch?v=113GE8yXMpc)
92/100 r3 hold E on pyro regisvine
78/100 r1 hold E (from friend) in abyss

Significance: It'd adjust energy computations if something is found. (edited)

Shouga

50 r3 hold E sacbow procs

oh yeah, xq on pma hit 87/100 with an r3 sac sword

Rito Jan 29, 2022 2:09 PM

Just put finding not theory/finding/bug

Homura Jan 29, 2022 2:09 PM

aight

Rito Jan 29, 2022 2:10 PM

Rito Jan 29, 2022 2:11 PM

@Homura imo go look at this beidou one and see it frame by frame to see if all enemies are attacked at the same frame

To me it looks like they are

Homura Jan 29, 2022 2:13 PM

ye it looks like the same frame

Rito Jan 29, 2022 2:15 PM

Yeah at frame 185

I'm on phone so I can't see the exact frame it might be 184

Homura Jan 29, 2022 2:17 PM

aight I'm 22 tests into hitting 5 enemies at the same time with r3 sac ayaka, I think my hypothesis is wrong for aoe then this isn't 60% at all

Rito Jan 29, 2022 2:17 PM

Rito Jan 29, 2022 2:18 PM

How many procs?

Homura Jan 29, 2022 2:18 PM

22/23 atm

Rito Jan 29, 2022 2:18 PM

Thats interesting (edited)

Homura Jan 29, 2022 2:18 PM

yeah this hypothesis seems wrong but who knows maybe I'll fail 39 times in a row before 100

Rito Jan 29, 2022 2:19 PM

How many for diona did proc after 50 tests?

Homura Jan 29, 2022 2:19 PM

47/50

Rito Jan 29, 2022 2:19 PM

pai Jan 29, 2022 2:20 PM

k!calc 1-(0.6)^5

Keqing BOT Jan 29, 2022 2:20 PM

pai#3602

Did I do it right?

Input:

1-(0.6)^5

Output:

0.9222400000000001

Homura Jan 29, 2022 2:20 PM

between me and my friend there's 700 tests of sac bow procs on diona over 7 tests, each test was legit under by a lot

pai Jan 29, 2022 2:20 PM

it should proc 92% of the time when it hits 5 enemies right?

Rito Jan 29, 2022 2:20 PM

Yes I believe (edited)

Homura Jan 29, 2022 2:20 PM

an r1 sac yes

Rito Jan 29, 2022 2:21 PM

In video it's r3 tho

pai Jan 29, 2022 2:21 PM

k!calc 1-(0.4)^5

Keqing BOT Jan 29, 2022 2:21 PM

pai#3602

Did I do it right?

Input:

1-(0.4)^5

Output:

0.98976

Rito Jan 29, 2022 2:21 PM

98

pai Jan 29, 2022 2:21 PM

oh wait i'm doing something wrong

Homura Jan 29, 2022 2:22 PM

wdym? that's right

pai Jan 29, 2022 2:22 PM

why does r1 have a higher chance than r3

Homura Jan 29, 2022 2:22 PM

oh I thought u did r1 and then r3

Rito Jan 29, 2022 2:22 PM

Me too lol

Homura Jan 29, 2022 2:23 PM

you subtract the chance of failure, not the chance of success

pai Jan 29, 2022 2:23 PM

oh wait i'm pepega

Rito Jan 29, 2022 2:23 PM

Homura Jan 29, 2022 2:23 PM

so it's actually 1 - (1-.6)^5 for an r3 sac

pai Jan 29, 2022 2:23 PM

well it still works, now it's r1 then r3

Rito Jan 29, 2022 2:25 PM

To me it looks like it dosent matter if you hit the enemies at the same time instead it might just be how many enemies you hit we can try it by testing it with beidou on 1 enemy then comparing it to beidou vs more than 1 enemy if I'm not pepega

Homura Jan 29, 2022 2:25 PM

wait actually, your sucrose strong enough to 1 hit floor 1 enemies with e?

Rito Jan 29, 2022 2:26 PM

Idk

I'm not home so can't test

Homura Jan 29, 2022 2:26 PM

that should mathematically be equal to diona's odds

pai Jan 29, 2022 2:26 PM

mine is level 80 with level 6 talents so probably

Homura Jan 29, 2022 2:26 PM

unless ofc diona's weird

Rito Jan 29, 2022 2:26 PM

Yeah

Yeah yours should one shot

Homura Jan 29, 2022 2:26 PM

5 electro slimes

lvl 25

I doubt killing the enemies would factor into this right?

Rito Jan 29, 2022 2:27 PM

Yeah we can just do these tests against one enemy and then on like 5 enemies in floor one we should see a dramatic difference in terms of proc counts

Rito Jan 29, 2022 2:28 PM

It wouldn't

pai Jan 29, 2022 2:28 PM

100 tests each?

Rito Jan 29, 2022 2:28 PM

50 can work tbh

I wont do it tho

Homura Jan 29, 2022 2:28 PM

bruh

this is my 238th sac weapon test of the day

Rito Jan 29, 2022 2:29 PM

Damn

No need to do it for alot of characters tho

The passive is the same for sac weps

Homura Jan 29, 2022 2:30 PM

ye

Rito Jan 29, 2022 2:30 PM

We can only do Sucrose

Or beidou

pai Jan 29, 2022 2:31 PM

beidou on electro slimes

Homura Jan 29, 2022 2:31 PM

currently I'd divide into diona/aoe that hits on the same frame/sgt that hits on different frames or just 1 hit in general

pai Jan 29, 2022 2:31 PM

that's probably a good idea

Rito Jan 29, 2022 2:32 PM

Diona is basically a mix of the two

Homura Jan 29, 2022 2:32 PM

diona is just different

Rito Jan 29, 2022 2:32 PM

Your attacking one enemy 5 times

But yeah good idea

This tc can end fast

pai Jan 29, 2022 2:32 PM

i think emc hits enemies on different frames too

Rito Jan 29, 2022 2:32 PM

Dk about dat

You wrote the guide

I'll check back here tomorrow see what yall found rn it's late

Gn

Homura Jan 29, 2022 2:34 PM

gn

Homura Jan 29, 2022 2:51 PM

aight Ayaka r3 sac sword, floor 1 killing 5 electro slimes with 1 E is 97/100

I mean that's still kinda low but also it's not 92 or 94

that's 303 tests from me today

legitimately had one skill issue per hundred test so I'm at least consistent with that

hmm... would lisa pressing E 5 times in a row count as the last thing? or should 1 E 1 hit be a different category?

pai Jan 29, 2022 3:14 PM

i just got 42/50 with r2 on sucrose (edited)

Homura Jan 29, 2022 3:16 PM

wait you hit all 5 of them right?

pai Jan 29, 2022 3:16 PM

yup oneshot all 5 every time

uploading the video

Homura Jan 29, 2022 3:16 PM

damn that's incredibly low

pai Jan 29, 2022 3:43 PM

https://youtu.be/dbYao5ego1Q

EmilysPaimon

Sucrose R2 Sacrificial Fragments on 5 enemies - 50 tries

Homura Jan 29, 2022 11:08 PM

btw where is this from? maybe I should do these per test?

pai Jan 29, 2022 11:46 PM

not sure but you can probably find a similar one here https://atozmath.com/

CQLQRS Jan 30, 2022 12:20 AM

u can probably take a look at this

https://library.keqingmains.com/evidence/equipment/weapons#sacrificial-weapons-procs-on-multiple-enemies

Weapons

nass did a sac vid

Homura Jan 30, 2022 1:36 AM

lemme just put all the currently data in one message so that it's easy to find here:
(Diona)
92/100 r2 hold E in abyss (theoretical 96.875%)
93/100 r2 hold E in abyss
95/100 r2 hold E in abyss
93/100 r2 hold E in abyss
94/100 r3 hold E in abyss (theoretical 98.976%)
92/100 r3 hold E on pyro regisvine
78/100 r1 hold E (from friend) in abyss (theoretical 92.224%)

49/50 r1 fav bow in abyss, 89.4% crit rate (theoretical 97.859% from itzsomebody's math)

(XQ)
87/100 r3 in abyss (theoretical 84%)

(Sucrose)
42/50 r2 5 enemies hit in abyss from pai (theoretical 96.875%)

(Ayaka)
97/100 r3 5 enemies hit in abyss (theoretical 98.976%)

(EMC)
47/50 r3 in abyss (theoretical 93.6%)

(Sara)
49/50 r3 4-5 enemies hit in abyss (theoretical between 97.44% and 98/976%) (edited)

btw I am allowed to add both of diona's r3 tests together like this right? pretty sure this is pretty good evidence of something being off on her specifically (if I had like more video)

also important number here is binomial probability right?

I think I'll do more sucrose tests even tho it's not really too weird

adding Nass' 114/137 on kazuha on 2 enemies with an r3 sac to the data here and I think it looks pretty safe to say that aoe or hitting multiple enemies on the same frame is within expected results, so diona herself is really just weird

currently the only people with the possibility of hitting the same enemy more than once per frame is diona, klee, aloy, and emc(?) so maybe I should test those too

Homura Jan 30, 2022 2:24 AM

aight emc functions normally in 50 tests, I have ran out of hypotheses for why diona doesn't do sac bow very well, it'd have to be sac bow specific or diona specific

pai Jan 30, 2022 2:30 AM

could test fav to see if it's a sac bow thing or a diona thing, if no one has done it yet

pai Jan 30, 2022 2:30 AM

Homura Jan 30, 2022 2:34 AM

who's be a good bow user to test it on tho? probs still floor 1 so sara and fischl are immediately disqualified

maybe childe?

Homura Jan 30, 2022 2:46 AM

EMC runs
https://youtu.be/f0T14Pxa7uA

Shouga

EMC 47 50 r3

pai Jan 30, 2022 3:20 AM

i mean you could either test fav on diona or sac on someone like amber/ganyu that does aoe dmg

Homura Jan 30, 2022 3:29 AM

I'll do both later, should diona have like 100% cr?

pai Jan 30, 2022 3:31 AM

that would probably make it more consistent since it's just 1 rng thing instead of 2

Rito Jan 30, 2022 4:24 AM

That's like 84% damn it's low

Tbh I think it's obvious now. But to make sure we need to see procs on 1 enemy

I currently from these tests believe that the passive activation requirement is not limited to one enemy and its not about in what frame you hit an enemy. Its quite possibly just how many times you hit enem y/ies

But still looking at these chances their low

But like if you have a r1 let's say sac claymore and test with one beidou e you should theoretically get something around 40% of the tries to proc (edited)

Which would already be very far from the 90%+ tests you got

And then comparing ayakas r3 results with diona r3 results it's 97/100 compared to 92/100

Which is close to each other

With diona your doing hitting one enemy but 5 times and with ayaka your hitting 5 enemies but once so it's like a 1Ã—5 and 5Ã—1 situation

We would guess the results should be around each other and it appears to be like that

Idk gimme your thoughts

Homura Jan 30, 2022 5:04 AM

AoE results atm includes a 42/50 r2 sucrose, 97/100 r3 ayaka (both of these are 5 targets), and, from Nass' tests, a 114/137 from r3 kazuha (2 targets). Ayaka and Sucrose were kinda low but it's not too rare to get tests of those numbers. Nass' is pretty much spot on. AoE seems like it functions fine.

Single target multiple hits also seems to function well enough as 47/50 r3 emc and 87/100 r3 xq are also pretty much spot on (xq is high but iirc it's not too rare to get that much difference on him). So it's fine.

So pretty much the only problem is legitimately diona or sac bow now. AoE and multiple hits on the same enemy really doesn't seem to have an effect, so ye currently the question is now is this a diona issue or a sac bow issue.

Rito Jan 30, 2022 5:05 AM

Homura Jan 30, 2022 5:05 AM

this is Ayaka's test, iirc the important number to look at is binomial probability, and I mean it's low but that's acceptably low imo

Rito Jan 30, 2022 5:06 AM

Yeah

Homura Jan 30, 2022 5:06 AM

diona's test tho, it's like not even a percent of ayaka's percentage

Rito Jan 30, 2022 5:06 AM

Yeah probably a diona issue, I mean why would the wep act differently when your equipping it on diona specifically?

Homura Jan 30, 2022 5:07 AM

tbf, all of our data is sac sword, greatsword, and fragments atm

it maybe a sac bow thing

or a diona thing

Rito Jan 30, 2022 5:07 AM

Yeah other weapons function as they should

Homura Jan 30, 2022 5:08 AM

the tests that need to be done are if Fav proc rates on diona are normal, and if sac bow rates on sara/childe/etc are normal

Rito Jan 30, 2022 5:08 AM

Sounds good then we should see the problem

Hopefully lol

Homura Jan 30, 2022 5:08 AM

if both of those are normal tho....

idk what to do anymore

Rito Jan 30, 2022 5:10 AM

If their both normal wouldn't it be a diona issue? Since we are considering that weapon is working as it should for childe, Sara etc..

Homura Jan 30, 2022 5:11 AM

if they both work normal, it's a sac bow on diona issue

Rito Jan 30, 2022 5:12 AM

Yeah that's what I mean

Homura Jan 30, 2022 5:12 AM

then ye

Rito Jan 30, 2022 5:12 AM

If that's the case we should just leave it be tbh (edited)

Homura Jan 30, 2022 5:13 AM

wait do y'all got better artifs I can't hit 100% cr with diona on fav bow?

Rito Jan 30, 2022 5:13 AM

Idk if I do

I can check in 5h when I get home

Homura Jan 30, 2022 5:14 AM

hmm...

ye I'll just skip that for now and test like idk childe/amber E on floor 1 with sac bow

oh wait venti exists

Rito Jan 30, 2022 5:15 AM

Can hold e hit 5?

Homura Jan 30, 2022 5:15 AM

on floor 1, pretty easily

Rito Jan 30, 2022 5:15 AM

then it would work if you can now get 100% cr

Homura Jan 30, 2022 5:16 AM

oh I'm testing sac bow first so no crit

Rito Jan 30, 2022 5:16 AM

Homura Jan 30, 2022 5:16 AM

the diona test I'll figure that out later

Rito Jan 30, 2022 5:16 AM

Hope we'll figure this out

Homura Jan 30, 2022 5:25 AM

aight these slimes don't like amber or venti, probs gonna do sara on another floor

Rito Jan 30, 2022 5:30 AM

Slimes died too many times

They want vengeance

Homura Jan 30, 2022 5:43 AM

aight, I kinda did some oof tests with sara and missed 1/5 peeps a few times, I don't think I ever missed 2 things at the same time tho

given that 1-.4^4 and 1-.4^5 are like 97.4ish% and 98.9ish%, Sara's test of 49/50 is still spot on

I should probs stop until later today I think I'm too tired to not have skill issues rn

Rito Jan 30, 2022 5:47 AM

So better and more sensful Results compared to diona

This is sac bow r3?

Homura Jan 30, 2022 5:49 AM

yep

Rito Jan 30, 2022 5:50 AM

Then this would be the expected results for diona ig (edited)

Homura Jan 30, 2022 5:50 AM

probably a diona or sac bow and diona issue

Rito Jan 30, 2022 5:50 AM

Yeah it's just diona most likely

Homura Jan 30, 2022 5:51 AM

I'll try later, not quite sure yet how exactly to do fav bow maths tho

Rito Jan 30, 2022 5:51 AM

Well need the fav bow tests too if fav bow works ok on everyone beside diona, it's a diona issue

Homura Jan 30, 2022 5:52 AM

true

or maybe there're resources on that already

Rito Jan 30, 2022 5:54 AM

We can ask scholars

Well people in tc general say there hasn't been any

Homura Jan 30, 2022 5:59 AM

bruh

pain

Rito Jan 30, 2022 5:59 AM

I feel bad for you doing all these runs

Homura Jan 30, 2022 5:59 AM

I've done 900 sac tests so far

Rito Jan 30, 2022 6:00 AM

Damn bro

Homura Jan 30, 2022 6:00 AM

if I had hard drive space I'd have footage of most of them

Rito Jan 30, 2022 6:00 AM

Like ps4 recording system Is just fucked up (edited)

I do a 720p recording ends up in 360p something

Not to mention the upload time which might take a full day sometimes

Homura Jan 30, 2022 6:02 AM

my 255 gb ssd when a 10-ish minute recording is already 1 gb

Rito Jan 30, 2022 6:03 AM

Homura Jan 30, 2022 11:21 AM

wait a minute...

diona stats on r1 fav bow

49/50 procs

if I'm doing the maths correctly chance to proc should be 1-.4^(5*.894)?

that's 98.336-ish %

so it's pretty much also spot on, but that one test didn't show numbers but it crit ig because fav triggered. all 5 hit, only 2 numbers (non-crit) popped out, fav triggered (edited)

Homura Jan 30, 2022 11:57 AM

almost forgot to send these

https://youtu.be/tHPgAaspY1U

Shouga

Sara 49 50 r3 4 or 5 enemies

https://youtu.be/2vikaPLhSPo

Shouga

Diona 89.4% Crit Rate, r1 fav bow, 49/50

ItzSomebody Jan 30, 2022 12:22 PM

1-(1-0.4*0.894)^5

is the theoretical proc rate for that specific setup (edited)

Homura Jan 30, 2022 12:23 PM

wait a minute

ItzSomebody Jan 30, 2022 12:23 PM

i have a writeup of the derivation here https://www.overleaf.com/read/yqwmprcyrbrf if you want to read

Overleaf, Online LaTeX Editor

An online LaTeX editor thatâ€™s easy to use. No installation, real-time collaboration, version control, hundreds of LaTeX templates, and more.

Homura Jan 30, 2022 12:23 PM

that's 1.775% to proc unless I inputted wrong

ItzSomebody Jan 30, 2022 12:23 PM

doesn't account for what happens if you get cryo-resonance midway between paws or rosaria A1/A4 being applied midway

Homura Jan 30, 2022 12:25 PM

wait damn major skill issue from my inputting damn

ItzSomebody Jan 30, 2022 12:25 PM

if you do tap E instead of hold E, change the ^5 to ^2

Homura Jan 30, 2022 12:26 PM

wait a minute

Homura Jan 30, 2022 12:28 PM

so this test is just extremely lucky or sumn?

ItzSomebody Jan 30, 2022 12:31 PM

o

no 0.4 should be 0.6

sorry my bad (edited)

Homura Jan 30, 2022 12:32 PM

ahh aight that makes more sense now

ItzSomebody Jan 30, 2022 12:32 PM

i forgot r1 was 60% instead of 40% on sac

Homura Jan 30, 2022 12:33 PM

I was thinking it was the same 1-.6 we do for sac

aight ty this is very helpful

Rito Jan 30, 2022 12:53 PM

So fav works as intended for diona?

Homura Jan 30, 2022 12:56 PM

pretty much

I'm legit out of ideas for how to explain why diona hits low AF proc rates for sac now

Rito Jan 30, 2022 2:21 PM

It's a diona with sac bow issue then

Currently

All characters work as intended with sac bow besides diona

While diona herself

Works as intended with fav bow

This leads to the conclusion that it's a diona with sac bow problem

But the real question..

Why diona? How?

Rito Jan 30, 2022 3:06 PM

Yea idk about it too

Been thinking for 30 mins (edited)

Homura Jan 30, 2022 3:13 PM

think there's anything else to add here? I doubt I know enough about the game to say anything about why sac on diona specifically functions weirdly. if not how does one conclude these things?

I think we've pretty much disproven all my previous hypotheses at this point, so the conclusion is currently we did a bunch of tests but yeah no it's just a diona with sac bow issue

Rito Jan 30, 2022 3:19 PM

Well the issue is definitely unknown to me too, maybe we can ask "bigger" tc minds to take a look?

Homura Jan 30, 2022 3:22 PM

yeah by all means, I'm pretty spent and dumbfounded ngl

Rito Jan 30, 2022 3:22 PM

Ultimate Noob Jan 30, 2022 3:33 PM

@Rito tldr

Rito Jan 30, 2022 3:35 PM

Basically sac weps work as intended for all characters except diona while diona and other characters work as intended with fav weps which leads to why diona specifically

Rito Jan 30, 2022 3:35 PM

Look at this

The cumulative probability compared to ayaka is just..

Way different

Ultimate Noob Jan 30, 2022 3:36 PM

So basically firing multiple have little impact?

Homura Jan 30, 2022 3:36 PM

wdym?

Ultimate Noob Jan 30, 2022 3:37 PM

Nvm ignore Iâ€™m being stupid

Homura Jan 30, 2022 3:37 PM

if it was multiple hits on a single target, emc shoulda been lower too

Rito Jan 30, 2022 3:37 PM

Ultimate Noob Jan 30, 2022 3:37 PM

Just woke up

Rito Jan 30, 2022 3:37 PM

It's alright

Yea like it's just not understandable

I'm dying

This tc needs a super big tc freak

Ultimate Noob Jan 30, 2022 3:39 PM

@Zamo take a look at this

Rito Jan 30, 2022 3:40 PM

Is this the melting everyone in tc mention?

Faranight Jan 30, 2022 3:47 PM

"this works for all cases except Diona E" is definitely something I've dealt with more than once. That ability is extremely cursed in its implementation

Rito Jan 30, 2022 3:48 PM

Diona is built different

Zamo Jan 30, 2022 8:18 PM

I thought this was a known thing

Homura Jan 31, 2022 12:10 AM

wait it's known and documented diona doesn't proc sac bow as much as she mathematically should, and the reasoning for this is also known?

Rito Jan 31, 2022 3:55 AM

No one knows why diona is like that

If anyone is gonna know about how to fix it or what's happening here, it's the devs tbh. (edited)

In the end its probably a issue of "wrong coding" since we've seen that diona works as intended with fav bow and others work alright with sac bow (edited)

pai Jan 31, 2022 4:33 AM

afaik this isn't in the tcl so it'd be nice to have some evidence since the videos are already recorded

Rito Jan 31, 2022 4:46 AM

Yeah who in the world would want to do like 700 tests

Zamo Jan 31, 2022 7:14 AM

that aoe hits proc sac, if that's not it maybe rename the ticket KEKl

But yeah, I didn't really read much, just skimmed over the math.

Homura Jan 31, 2022 7:18 AM

well true

makeway4pK Feb 4, 2022 9:51 AM

could this be attributed to a very short hidden cooldown on the sacrificial passive? tho I don't know if Diona's ðŸ¾ ALWAYS hit on the same frame or not

Rito Feb 5, 2022 5:28 AM

The devs won't say the passive has a x amount of cooldown if they didn't know. So it's not about the passive cooldown; also diona I believe hits the enemy with her E shots at once.

Not sure about always tho

makeway4pK Feb 5, 2022 5:37 AM

i see

pai Feb 6, 2022 6:58 PM

@Homura is there anything else you'd like to test? if not you can send an updated version of the write up then close the ticket

Homura Feb 7, 2022 2:11 AM

Finding: sac bow procs on Diona's hold E do not match the theoretical chances, the reason for this is not known or was not determined in these tests

Evidence:
(Diona)
92/100 r2 hold E in abyss (theoretical 96.875%)
93/100 r2 hold E in abyss
95/100 r2 hold E in abyss
93/100 r2 hold E in abyss
94/100 r3 hold E in abyss (theoretical 98.976%) first 50 runs only, but it's 47/50 so same percentage at least: https://www.youtube.com/watch?v=113GE8yXMpc
92/100 r3 hold E on pyro regisvine
78/100 r1 hold E (from friend) in abyss (theoretical 92.224%)
280/300 r3 hold E in abyss https://youtu.be/gNgc723WkO8

49/50 r1 fav bow in abyss, 89.4% crit rate (theoretical 97.859% from itzsomebody's math) https://youtu.be/2vikaPLhSPo

(XQ)
87/100 r3 in abyss (theoretical 84%)

(Sucrose)
42/50 r2 5 enemies hit in abyss from pai (theoretical 96.875%) https://youtu.be/dbYao5ego1Q

(Ayaka)
97/100 r3 5 enemies hit in abyss (theoretical 98.976%)

(EMC)
47/50 r3 in abyss (theoretical 93.6%) https://youtu.be/f0T14Pxa7uA

(Sara)
49/50 r3 4-5 enemies hit in abyss (theoretical between 97.44% and 98/976%) https://youtu.be/tHPgAaspY1U

Significance: To find out why diona doesn't proc sac bow correctly, both sac and fav bow were tested on diona to ensure that it was simply not diona's hold E being weird in general, sac weapons were tested on hitting multiple enemies to see if it triggered correctly, a sac weapon (sword) was tested on hitting one enemy multiple times with the same skill (emc e) to see if it triggered correctly, and sac bow specifically was tested on hitting multiple enemies to see if it triggered correctly. All in all, everything works pretty much as expected except diona's hold E on sac bow, so I don't know what's causing this. Since diona as a battery sometimes uses sac bow, knowing she doesn't proc it as well as math dictates means energy calculations with sac bow should change to reflect that she doesn't follow standard math. (edited)

Shouga

50 r3 hold E sacbow procs

Shouga

Diona r3 sac bow 280/300 procs

Proc fails on 16, 18, 26, 52, 60, 61, 83, 87, 88, 119, 132, 145, 164, 192, 198, 201, 247, 256, 266, 277

Shouga

Diona 89.4% Crit Rate, r1 fav bow, 49/50

EmilysPaimon

Sucrose R2 Sacrificial Fragments on 5 enemies - 50 tries

Shouga

EMC 47 50 r3

P.S. most of the no videos are disk space issues on my part, y'all could test/record it yourselves and I'm pretty confident you'll also hit under

Ticket Tool BOT Feb 7, 2022 2:14 AM

Ticket Closed by @302641102450720779

NZPIEFACE Feb 7, 2022 7:04 AM

you need

way more tests

@Dooners how many tests, what crit rate, and what refinement would you suggest?

Dooners Feb 7, 2022 2:43 PM

Crit rate? I don't think there's a problem with fav, right? It's the sac bow that seems like it's having issues. All the numbers besides sac bow diona seem reasonable. But I agree that there are not nearly enough trials to be conclusive, though it's really interesting that every sac diona trial has lower procs than expected. I haven't watched any footage, but likely explanations include a) some of her paws completely whiffing by either flying past the target or hitting the ground while flying b) the wait between skill uses weren't long enough to refresh the passive. Maybe both were accounted for in the video evidence though, haven't checked.

Number of trials would depend obviously on what confidence you want and how much power you're looking for, there's calculators for that like https://www.stat.ubc.ca/~rollin/stats/ssize/b1.html

NZPIEFACE Feb 7, 2022 2:46 PM

wait

fuck

i read it wrong

Dooners Feb 7, 2022 2:46 PM

Quickly plugging in the numbers for the trial that does have a video, you need around 280 attempts for 0.8 power

Flashbang

Dooners Feb 7, 2022 3:05 PM

ok and then actually doing the hypothesis test:

    Exact binomial test

data:  94 and 100
number of successes = 94, number of trials = 100,
p-value = 0.09378
alternative hypothesis: true probability of success is less than 0.96875
95 percent confidence interval:
 0.0000000 0.9735503
sample estimates:
probability of success 
                  0.94

so if we only take into account the sac bow trial with video evidence, it is not conclusive yet (at 0.05 alpha)

D.....K Feb 7, 2022 5:26 PM

isnt 5 percent enough to reject null hypothesis?

Dooners Feb 7, 2022 5:31 PM

yea and we're at 0.09

tho if we use the unrecorded data for R2 sac we're at p-value = 0.0001917 that the trial probability is lower than expected probability

however since there's no footage we don't know if any of the paws are missing the target, which would skew the results

D.....K Feb 7, 2022 5:33 PM

Ticket Tool BOT Feb 7, 2022 5:56 PM

Ticket Opened by @844955255819993178

pai Feb 7, 2022 5:57 PM

@Homura you can edit your write up to include the new video now

Homura Feb 7, 2022 5:57 PM

aight

pai Feb 7, 2022 5:57 PM

Homura Feb 7, 2022 5:59 PM

there

Dooners Feb 7, 2022 7:23 PM

alright I've gone through your new 300 trial video, these are the timestamps of failed procs i got

so 20

If you have nothing else to add you can close it and I will verify that sac bow procs on diona's hold E is inconsistent with the expected rates. Figuring out exactly why is out of scope of the ticket, so that's fine

Homura Feb 7, 2022 7:24 PM

aight ty

Dooners Feb 7, 2022 7:25 PM

ty for the evidence and ty youtube for the fast forward 10 seconds button

Ticket Tool BOT Feb 7, 2022 7:25 PM

Ticket Closed by @302641102450720779

Support team ticket controls

Dooners Feb 7, 2022 7:25 PM

Verified.

Min Feb 8, 2022 6:40 PM

jesus christ i just read the whole channel and the length you went through for this is @Homura

Rito Feb 8, 2022 8:05 PM

Yeah he's a beast doing all those tests

CrazyWeke123 Feb 14, 2022 7:55 PM

Verified with the help of Dooner's timestamps

Ticket Tool BOT Feb 14, 2022 9:32 PM

Transcript Saving