Server: KQM | Genshin Impact (763583452762734592) Channel: ganyu-ult-frequency-enemy-size (845508390595592202) Messages: 152 Yuehai Ticketing Secretary#4843 (557628352828014614) Yukarix#6534 (112026394325749760) Aluminum | Harbinger of Jank#5462 (306566993711202305) Carrier5by5#8394 (452698567853735937) Greyhound | Cryo Currency Dealer#7836 (264018289359847424) srl#2712 (156096226721398786) xf3#3123 (141941251350986752) jstern25| jstep hutao#1399 (214554770126405632) Artesians, Head of TC#0002 (185216727351558145) fishing with childe#9407 (217820099845947392) KQM | Genshin Impact - ganyu-ult-frequency-enemy-size
KQM | Genshin Impact
ganyu-ult-frequency-enemy-size
152 messages
Yuehai Ticketing SecretaryBOTMay 22, 2021 3:48 AM
@Yukarix - ganyu-ult-frequency-enemy-size 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
YukarixMay 22, 2021 4:00 AM
Hypothesis: Different sized enemies have a different average hit % rate in Ganyu's ult. In other words, Ganyu's ult does not only scale with # of enemies, but also enemy size. Geovishap: 1st try https://streamable.com/h15t6c 25 hits counted(edited)
geo test
2nd try https://streamable.com/mrgwq8 27 hits counted
2021-05-21 23-49-26_Trim
3rd try https://streamable.com/j4tlyk 23 hits counted(edited)
2021-05-21 23-55-35_Trim
YukarixMay 22, 2021 4:24 AM
4th try https://streamable.com/k9mgkk 21 hits counted(edited)
2021-05-22 00-10-04_Trim
5th try https://streamable.com/na98ua 23 hits counted(edited)
geo test
6th try https://streamable.com/szvf8b 28 hits counted(edited)
geo test
Aluminum | Harbinger of JankMay 22, 2021 4:35 AM
Isn't ganyu burst targeted + random? so extra hits is enemy size/ganyu burst size (assuming uniform distribution which might not even be the case with how much variance you are showing).
YukarixMay 22, 2021 4:36 AM
yeah it's targeted + random but i don't think anyone's actually looked at how much different sized enemies get hit
anyways..this ticket will take a while cause i'm lazy and i'm already getting fatigued lmao
Aluminum | Harbinger of JankMay 22, 2021 4:36 AM
my worry is that if the distribution is not uniform within the burst zone then positioning might matter
YukarixMay 22, 2021 4:37 AM
unless someone wants to help out
it could
that's another ticket i want to open up, if enemy position in the burst affects average icicle hits(edited)
also one last ticket is, taunt mechanics
Aluminum | Harbinger of JankMay 22, 2021 4:38 AM
I might've helped but I don't have ganyu and I also have my own silly ticket to be lazy about lol
taunt in general is interesting
I've heard somewhere ganyu's is the best?
YukarixMay 22, 2021 4:38 AM
i think it's the tankiest in scaling but not sure if it taunts differently than others
Aluminum | Harbinger of JankMay 22, 2021 4:39 AM
Well Mona's is invincible lmao
YukarixMay 22, 2021 4:41 AM
true lol
uh i think i need to open up a theory hunt for this
i cant do this mysef
myself*
too much time
LOL
Carrier5by5May 23, 2021 12:36 PM
a Chinese player found how to let Azhdaha get 50 hits
甘雨单q冰锥可命中若陀龙王50次_哔哩哔哩 (゜-゜)つロ 干杯~-bilibili
YukarixMay 23, 2021 2:22 PM
holy
Greyhound | Cryo Currency DealerMay 23, 2021 8:21 PM
Smaller enemies' size doesn't seem to make much of a difference between each other 13/11/16/11/16/14 for small: https://www.youtube.com/watch?v=Q9o8kJpU-5Q 11/12/10/13/10 for average humanoid: https://www.youtube.com/watch?v=OsArJAXUBsI 14/12/12/11/15/14 for fat: https://www.youtube.com/watch?v=sCqHL1BjkBU
Arco
ganyu small enemy
Arco
Ganyu average humanoid
Arco
Ganyu fat enemy
YukarixMay 23, 2021 10:11 PM
appreciate your help
do you think you could also help out with geovishap testing?
Greyhound | Cryo Currency DealerMay 23, 2021 10:11 PM
I'll be checking/adding more tests I just didn't have much free time today
YukarixMay 23, 2021 10:11 PM
so i think there is a size between the medium size geovishap and humanoid
Greyhound | Cryo Currency DealerMay 23, 2021 10:11 PM
let me know if you want me to check anyhting special otherwise I was going to look into mitachurls next
ThumbsUpElectro1
YukarixMay 23, 2021 10:12 PM
yeah i think mitachurls, and then ruin guards are a little bigger
i'll do some more runs of geovishap today then
Greyhound | Cryo Currency DealerMay 23, 2021 10:12 PM
I also intend to extend the attempts to try and see if I can get any consistant number over 50 or so attempts but that will come some time later, for now I just wanted to see if there's clear differences like how vishap is in the 20-30 range instead of 10-20(edited)
YukarixMay 23, 2021 10:14 PM
Yep that sounds like a good idea
I want to test out the primo geovishap as well, but my shield characters are not well built
so I can't really just afk stand next to them to count clearly
Greyhound | Cryo Currency DealerMay 23, 2021 10:14 PM
leave it to me, time for 55k zhongli to pay his bills
YukarixMay 23, 2021 10:14 PM
euladrunk
YukarixMay 23, 2021 11:53 PM
I did 6 additional tests, this time not as high as last time but the average is still hovering above the 20 mark
https://www.youtube.com/watch?v=TOjnzXt-Thg Counts were 20/19/23/25/22/18
yukarix
Ganyu icicles
Geovishap: 25/27/23/21/23/28/20/19/23/25/22/18
Greyhound | Cryo Currency DealerMay 24, 2021 6:49 PM
counting now
Greyhound | Cryo Currency DealerMay 24, 2021 7:15 PM
25/24/23/24/19/26/22/23/21
very hard to see because of how fat his ass is
YukarixMay 24, 2021 7:19 PM
Interesting we got pretty similar results
Greyhound | Cryo Currency DealerMay 24, 2021 7:25 PM
should I even bother checking azdaha after that bilbili vid? kleekpoint
YukarixMay 24, 2021 9:05 PM
kleeHHAHa
YukarixMay 25, 2021 5:00 PM
@Greyhound | Cryo Currency Dealer What other enemy sizes do we need to test?
Greyhound | Cryo Currency DealerMay 25, 2021 5:01 PM
mitachurl&lawachurl
flower?
slime?
YukarixMay 25, 2021 5:02 PM
oh yeah flower is super small
We can just use that azdaha clip for giant enemies
YukarixMay 27, 2021 4:06 AM
@Greyhound | Cryo Currency Dealer I can test some more over the weekend
what sort of conclusion should we make out of this ticket once we finish?
Greyhound | Cryo Currency DealerMay 27, 2021 8:41 AM
I'm gonna try to see when I can find the time to get more samples to narrow down the RNG the conclusions we're to draw from this ticket are that ganyu's ult be it as a main or sub carry can do either less or more than the initial expectations prior to it
Greyhound | Cryo Currency DealerMay 27, 2021 1:03 PM
https://www.youtube.com/watch?v=k6aLvC-MKKM 14(ran outside)/17/18/16/17 probably hitcount is slightly higher but it's very hard to see anything when he starts hopping around
Arco
Ganyu ult vs lawachurl
https://www.youtube.com/watch?v=GvOIpAR70wE 12(ran outside)/18/17/14/15/19
Arco
Ganyu ult vs mitachurl
srlMay 27, 2021 6:05 PM
@Yukarix this seems to be what governs ult drops
15 second duration, 0.3 second think interval, looks like each icicle takes .5 seconds?
targetting is a bit confusing
looks like every interval, it checks if a global flag called _ABILITY_Ganyu_UnbreakableIce_UnavailableMark is set to 0 or not
if set to 0, then - set _ABILITY_Ganyu_UnbreakableIce_UnavailableMark to 1, fur 1.45s - looks like set position to the target found
so i'm reading this correctly, it just generates random from center, looks for a target every 1.45s, if ok then drop it at that target otherwise keep w/e random position it generated
the random range is 0.5 to 9.5, i guess it's just uniformly random across this circle somehow? idk my geometry is bad
Greyhound | Cryo Currency DealerMay 27, 2021 6:06 PM
the circle is 10 meters so that checks out
srlMay 27, 2021 6:06 PM
i guess if you could work out the size of the mob, you can math out the expected number of hits based on size
YukarixMay 27, 2021 6:07 PM
I was wondering if there was something laying around somewhere that had sizes of mobs
is the aoe of the icicle also in the json?
Greyhound | Cryo Currency DealerMay 27, 2021 6:07 PM
I expect icycle range to be ~2.5-2.8m
srlMay 27, 2021 6:07 PM
"defendArea": { "defendRange": 50.0 },
there's monster leash range? lol
Greyhound | Cryo Currency DealerMay 27, 2021 6:08 PM
yeah, can't have a mob chase you the entire map
srlMay 27, 2021 6:10 PM
can't find anything on mob size
i suspect that's built into the model itself?
there's model size scale factor
this is prob ur icicle size
radius 2.5
Greyhound | Cryo Currency DealerMay 27, 2021 6:11 PM
checks out with my experience
srlMay 27, 2021 6:11 PM
hmmm radius? seems kinda big?
isn't circle radius only 9.5?
YukarixMay 27, 2021 6:11 PM
it is pretty big
srlMay 27, 2021 6:11 PM
ah
never paid attention
Greyhound | Cryo Currency DealerMay 27, 2021 6:11 PM
they're both chonky
srlMay 27, 2021 6:12 PM
well ok i think if u visually measure a monster size
you can roughly calculate an expected value from that?
maybe use the monster height as a reference point
cause that's listed in the files
YukarixMay 27, 2021 6:12 PM
i have no clue how to do that well
i don't have a good photo editor
srlMay 27, 2021 6:12 PM
take a screenshot
measure how many pixels
loll
mspaint
Greyhound | Cryo Currency DealerMay 27, 2021 6:13 PM
I don't think height matters as much as wideness
srlMay 27, 2021 6:14 PM
well you have to measure widness
and use something as a scale
so if you have have height, and the monster is let's say 4:1 height to width in pixels
Greyhound | Cryo Currency DealerMay 27, 2021 6:14 PM
abyss tiles to the rescue as usual
srlMay 27, 2021 6:17 PM
then u have the width, then just draw a circle
well at least this one only got one random circles
i was actually looking at amber.. and that has 3 circles? O_O
srlMay 27, 2021 7:09 PM
hmmm what's that massiveAttackRatio thing
is that like a dmg reduction for hitting big targets?
Aluminum | Harbinger of JankMay 28, 2021 11:29 PM
Probability of untargeted icicle hit on an enemy in the center: Uniform distribution: (radius_of_icicle + radius_of_enemy)^2/(radius_of_zone)^2 Uniform radius and Uniform angle: (radius_of_icicle + radius_of_enemy)/(radius_of_zone) Maybe validate against these two first Maybe someone can add something for Gaussian (or a formula to reverse engineer sigma assuming gaussian) Correct me if my math is incorrect.(edited)
Aluminum | Harbinger of JankMay 29, 2021 12:32 AM
If you think the distributions for selecting an arbitrary point in a circle are annoying. Consider how one could possibly select an arbitrary chord in a circle uniformly and whatever that even means.
srlMay 29, 2021 2:31 AM
how noticeable would it be if Q is biased to falling in center?
Aluminum | Harbinger of JankMay 29, 2021 2:34 AM
I think it depends on enemy size but the difference might be a few icicles.
since there's a lot of untargeted icicles right?
srlMay 29, 2021 2:35 AM
ya, 40 untargetted
hmm but if we repeat the test of let's say on a fatui for like 100 times and the avg agrees then it's prob fine right
Aluminum | Harbinger of JankMay 29, 2021 2:40 AM
a stats person could tell you how confident you can be after that... well in some sense anyways.
YukarixMay 29, 2021 9:27 PM
i'll do some testing on slimes today
what else we need @Greyhound | Cryo Currency Dealer
flowers?
feel like those will be a pain to test
Greyhound | Cryo Currency DealerMay 29, 2021 11:40 PM
flowers are just hilichurls but when they do their reposition underground thing they escape 2 guaranteed + whatever random icycles during that time, what is still left on the table is ruin guards and graders, we should start getting some larger sample sizes to try and find an average then run them by the formulas suggested and try to figure out what type of random ganyu's ult uses, unless you think that goes beyond this ticket and we do that in a different one at a later date, I'll probably be slow on getting footage because my finals start this monday
YukarixMay 30, 2021 12:17 AM
Hmm
The part of what sort of random ganyu's ult uses should be a different ticket
I can test ruin guard tomorrow then
I killed all of mine
today
YukarixMay 30, 2021 10:08 PM
Ruin Guard:
18/19/21/21/23/28
YukarixJun 1, 2021 6:20 AM
todo: ruin grader
xf3Jun 15, 2021 8:34 PM
@Yukarix whats the progress on this
Greyhound | Cryo Currency DealerJun 15, 2021 8:39 PM
missing ruin grader testing&formulating a conclusion from the data
YukarixJun 15, 2021 9:23 PM
need help finishing this up, i've been real busy and burnt out lately
Greyhound | Cryo Currency DealerJun 15, 2021 9:31 PM
if you can give me the conclusion you want to draw from it I can get the rest done and you can post the writeup&close(edited)
YukarixJun 15, 2021 11:26 PM
Hmm
i think we need a follow up ticket to this with the math behind it but a general conclusion that Ganyu's single target burst damage looks to generally scale with enemy size is good
Greyhound | Cryo Currency DealerJun 15, 2021 11:27 PM
in that case I can gather the videos in one message and you can close after? KeqingThink
and we work on finding the rng formula in a new ticket at a later date?
YukarixJun 16, 2021 1:46 AM
sure ThumbsUpElectro
that would be helpful
tyty
Greyhound | Cryo Currency DealerJun 16, 2021 7:16 PM
Finding: Ganyu's rng ult icycles are more likely to hit a target(beyond the base 10) if they are larger in size (who'd have thought keqewide ) up to all 50 for enemies large enough such as Azhdaha Evidence: Azhdaha 50 hits: https://www.bilibili.com/video/BV13N411Z7JM#reply4609628931 Hilichurl 13/11/16/11/16/14 hits: https://www.youtube.com/watch?v=Q9o8kJpU-5Q Humanoid 11/12/10/13/10 hits: https://www.youtube.com/watch?v=OsArJAXUBsI Fat fatui 14/12/12/11/15/14 hits: https://www.youtube.com/watch?v=sCqHL1BjkBU Geo Vishap 20/19/23/25/22/18 hits: https://www.youtube.com/watch?v=TOjnzXt-Thg Primo Vishap 25/24/23/24/19/26/22/23/21 hits: https://www.youtube.com/watch?v=CLAbZ-wppuE Geo Lawachurl 14(ran outside)/17/18/16/17 hits: https://www.youtube.com/watch?v=k6aLvC-MKKM Mitachurl 12(ran outside)/18/17/14/15/19 hits: https://www.youtube.com/watch?v=GvOIpAR70wE Ruin guard 18/19/21/21/23/28 hits: https://youtu.be/SLCT8oQaDX4 Significance: Ganyu's ult is more crack cocaine than expected(edited)
icycle RNG formula to be determined in a future ticket
Yuehai Ticketing SecretaryBOTJun 16, 2021 9:42 PM
Ticket Closed by @Yukarix
Save transcript Reopen Ticket Delete Ticket
jstern25| jstep hutaoJun 23, 2021 11:35 PM
Verified
Artesians, Head of TCJun 23, 2021 11:36 PM
verified
fishing with childeJun 24, 2021 8:19 PM
working on pr
Yuehai Ticketing SecretaryBOTJun 24, 2021 8:31 PM
Transcript Saving