Server: KQM | Genshin Impact (763583452762734592) Channel: hu-tao-vape-melt-consistency (835296118795534388) Messages: 79 Yuehai Ticketing Secretary#4843 (557628352828014614) jamie#9248 (712372659312394291) Greyhound | cryo is best crypto#7836 (264018289359847424) jstern25| ₼29s_LECTORS#1399 (214554770126405632) UnicornAlpha#0128 (291444384040484865) Mono$INVEST₼CRYOARCHONKAEYA#8058 (154004693826273280) Ayzel#7399 (276182497145126912) pikachu checkpins surprised#6934 (111978131664928768) xf3#3123 (141941251350986752) Artesians#0002 (185216727351558145) KQM | Genshin Impact - hu-tao-vape-melt-consistency
KQM | Genshin Impact
79 messages
Yuehai Ticketing SecretaryBOTApr 23, 2021 11:28 PM
@jamie, As an author, it is your sole responsibility to complete this ticket. - Any assistance provided by other users should be taken constructively - Disrespectful behavior will not be tolerated - When done, type $close
Theory/Finding/Bug: Title of your submission Evidence: Explanations with calculations and/or Youtube/Imgur proofs Significance: Conclusion
jamieApr 23, 2021 11:45 PM
(EDITED 9th May with n2cj findings) Finding: Tested out the freeze underlying aura and how it pertains to how hu tao + dragon's bane. Ran the standard freeze team hu tao/xingqiu/diona/kaeya and tested different combos (n1c, n1cd, n1cj, n2cj), in order to see if there's a difference which would keep a hydro aura under the enemy as much as possible in order to proc dragon's bane on melt as much as possible. Both n1cd and n1cj were found to produce more than 2x the amount of procs. Whilst testing I also attempted to confirm vape% and melt% with the team as there seemed to be no exact answer in the community and if different combos (n1c, n1cd, n1cj,n2cj), would also produce different %s. General reaction uptime was confirmed to be around 80%, but melt% were much higher than the 1/3 expected, the 3 main combos were found to be around 40-50% melt and 33% vape, n1c,n1cd,n1cj. However, n2cj had an additional pyro attack and therefore suffers at 25% melt and 57% vape uptime. notes about n2cj - sometimes players are forced to run this combo in longer fights to conserve stamina appropriately - tests show this has much different results than the other combos, as there is an extra pyro hit - charge attack melt% plummets down (25%), vape% goes up (57% approx, could be summerised as either 1/2 or 1/3), no reaction% is pretty much the same as the others at 20% - overall despite strong talent% multipliers (m/v), it actually looks pretty scuffed for freeze melt(edited)
Evidence: google sheet with runs, 10+ runs of each combo (hopefully this is enough), and video links of the recordings in the sheet too
dragon's bane vs freeze melt
Sheet1 dragon's bane vs freeze melt,hu tao stats,reference hu tao/diona/kaeya/xingqiu,hp,14459,enemy,lv 85 ruin guard,videos rotation: xingqiu QE > diona EQ > kaeya EQ > hu tao,atk,500,n1 vaporize,1503,n1 no aura,629,crit,no reaction,normal melt vs db melt example,
Significance: Dragon's bane (potentially confirmed?) to have much more power than initially thought. Not everyone has a staff of homa, and many thought it would only work in vape teams. Potentially her best 4 star weapon, would be interesting if someone could compare dps vs deathmatch in melt comps. Also interesting that this melt team is also 3x 4 star supports, and kaeya is f2p with both diona and xingqiu available free from events (as of 1.5). If this is all accepted by an editor then these findings would be good to promote to f2p/low spenders. Also melt comps might have an even higher ceiling than expected, as afaik melt% uptime isn't expected to be around 50%. (EDIT) tldr; n1cd charge attacks vape% 33% melt% 45% no reaction% 22% on melt DB proc% 37% n1cj charge attacks vape% 29% melt% 51% no reaction% 20% on melt DB proc% 33% n1c charge attacks vape% 36% melt% 42% no reaction% 21% on melt DB proc% 14% n2cj charge attacks vape% 57% melt% 25% no reaction% 17% on melt DB proc% 19% all combos n1 no reaction% consistently 60-66% (n2cj n1 I got actually got lower at 56%, but that could be down to sample size variance) melt% and vape% inconsistent but melt% generally higher at 20-30%, vape% under that on melt DB proc% inconsistent and rare, but clearly possible n2cj combo n2 results similar melt% as n1 with around 30%, but vape% approximately halves compared to n1, with its share of the % going to more no reactions, which are around 66%) As for why faster combos give higher db proc chance?, I think due to either (or a combination of) faster xingqiu hydro application which is strong (procs are most often directly after c6 swords), or faster pyro application which helps tip balance to favour the underlining hydro? But I don't know enough about the game to say so I'll leave the conclusion to someone who knows best.(edited)
Greyhound | cryo is best cryptoApr 23, 2021 11:45 PM
I don't know what TC's chosen approach to sub tickets of larger ones is but this looks like data that should be in <#825783454158291004> theorycraft
jamieApr 23, 2021 11:47 PM
I was told it went out of the scope by mono (i think?)
as it went more in a hu tao team direction
Greyhound | cryo is best cryptoApr 23, 2021 11:49 PM
I see, in that case nevermind me GanyuSip I thought it would be part of it since the ticket discussed dbane before and also requested melt to vape %s of underlying aura reactions as the hunt
jstern25| ₼29s_LECTORSApr 23, 2021 11:50 PM
video example of dbane melt procs
jamieApr 23, 2021 11:54 PM
all good! my first time doing something like this anyway, this is what no resin and a few days off from work does to a man. Plus being homaless and needing to cope with a budget weapon
jstern25| ₼29s_LECTORSApr 23, 2021 11:56 PM
keking it led to a great conversation at least
jamieApr 23, 2021 11:58 PM
yes mate thanks for all the advice!
@UnicornAlpha if this is all approved you could maybe adjust some of your values in your hu tao calc?
UnicornAlphaApr 24, 2021 12:00 AM
yep! this looks like some great work KleePog
Greyhound | cryo is best cryptoApr 24, 2021 12:02 AM
further hutao vape melt buffs GanyuSad where my ganyu reverse melt buffs at
jamieApr 24, 2021 12:05 AM
also just for clarity i tried to get the title to be: ht dragon's bane freeze procs% + freeze team melt%
jstern25| ₼29s_LECTORSApr 24, 2021 12:36 AM
melt % potentially can go even higher with faster cryo applying supports than diona (sucrose, rosaria, ganyu)
jamieApr 24, 2021 1:07 AM
most likely, also it would be a good idea to test in the other direction, with single cryo instead (zhongli+kaeya) and seeing if it achieves a similar proc rate + melt% rate, but i figured that was for a follow up submission.
yeah I said not to put it there
because we want to crunch down on tickets
so I advised to constrain it to this one idea
@jamie I would suggest putting the percentages from your data in the significance and wrapping up the ticket if you have nothing left to add
from a surface glance this looks good
jamieApr 24, 2021 6:59 AM
cheers mate, added the percentages in a tldr edit. I'll close this now
Yuehai Ticketing SecretaryBOTApr 24, 2021 7:00 AM
Ticket Closed by @jamie
Save transcript Reopen Ticket Delete Ticket
AyzelMay 3, 2021 6:08 AM
by DB proc melt%, do you mean the percentage of melts where DB passive proc'ed?
Yuehai Ticketing SecretaryBOTMay 3, 2021 6:08 AM
Ticket Opened by @Ayzel
jamieMay 3, 2021 6:41 AM
AyzelMay 3, 2021 6:42 AM
AyzelMay 4, 2021 1:32 AM
What cons is ur Xingqiu
pikachu checkpins surprisedMay 4, 2021 5:20 AM
any n2cj tests? might be relevant considering a low-spender scenario generally means hutao c0
i can record footage if it helps, my xq is at c5(edited)
not sure what to expect from the additional hydro apps that we get from xq c6 and from n2cj(edited)
but yeah this is definitely very nice work, ty for sharing PrayKeqing
oh shit
pikachu checkpins surprisedMay 4, 2021 5:20 AM
should be relevant for whales as well
might need to check this again
pikachu checkpins surprisedMay 4, 2021 5:21 AM
o h
low ping
its fine
pikachu checkpins surprisedMay 4, 2021 5:21 AM
oh yeah theyre safe
34 ping on n1cd tests
AyzelMay 4, 2021 6:26 AM
shit do we have to re-check every test to make sure people didn't have high ping on Xingqiu LMAO
xf3May 4, 2021 6:32 AM
jamieMay 4, 2021 2:26 PM
yeah i asked around after and it seems like alot of c0 players are forced to do n2cj in longer fights, so i was interested in getting around to testing that too. Just wanted to get this initial testing thumbs upped so i didn't add 33% more work on a bad submission. since this is my first one
also plan on testing variations of this comp like diona+rosaria, zhongli+kaeya in order to understand more exactly what double cryo is bringing to the table. but again i focused on this comp initially as it seems to be the freeze melt standard
jamieMay 4, 2021 2:36 PM
Yeah i suppose it could help for whales who have high refine db. Even in the 80% general reaction uptime, a lot of the no reactions were discovered to be pyro aura hits, which was nice to learn only by using dragons bane. Makes dragon's bane a fraction more stronger in calcs. Probably doesn't affect homa users outside of more accurately understanding the no reaction% vs melt% vs vape% ratios. Still it is pretty nice, i believe we should be able to bump up freeze melt hu taos dps "power levels" too, since from what i've asked around people usually low balled it generally at 1/3. What i find most interesting is that on the chinese server barely anyone is actually running melt, geo res and zhongli bennett comps seem to dominate significantly. So it would also be interesting in the future to compare exactly which comps are her best, though i'm predicting western freeze melt to be her best, just puzzled as to why it hasn't caught on at all in china. Would be good for a future analysis for sure
pikachu checkpins surprisedMay 4, 2021 5:08 PM
i somehow forgot about homa existing entirely LOL
well, weap banners are scuffed anyway, and personally i happened to have a db lying around(edited)
also, if your xq is c6 then that might be increasing db uptime, which is something to consider
either way this is by no means a bad ticket or a bad first submission
definitely a lot of material for further analysis
as for why china goes for geo res+zhong so often, i cant say i know... maybe zhongli is just more popular there, and/or theres a lot of trending people who use it without consideration for alternate setups
similar to why many think whiteblind is bis for noelle (yt/reddit overrepresentation)
jamieMay 4, 2021 7:38 PM
Alright guys, that leaves 2 questions then 1: do you want n2cj data or are the 3 n1c n1cd n1cj combos fine? 2: do we want someone with <c6 xingqiu to give their data or is this something to be noted as assumptions/grains of salt/considerations etc? Let me know what you need in order to close for review
jamieMay 6, 2021 2:31 PM
oh @Mono$INVEST₼CRYOARCHONKAEYA apologies on the ping but I forgot to @ you on this
n2cj data would be helpful for THIS ticket
lower con XQ can be verified somewhere else
jamieMay 6, 2021 3:13 PM
ok no problem i'll work on this later on
jamieMay 9, 2021 3:13 PM
hey, i've managed to add n2cj data into this submission. Added video proof and data into the sheet and now edited ticket with its tldr data and some some of my thoughts on it's results
did a healthy sample size of 14 runs too
tldr; its noticeably worse than the other combos for freeze melt in general, despite solid base talent% values. dragons bane proc% is low aswell, similar to n1c(edited)
will take a look after I get back from work
jamieMay 19, 2021 7:15 PM
Just circling back on this, can i close this for review?
Yeah go ahead
Yuehai Ticketing SecretaryBOTMay 19, 2021 9:17 PM
Ticket Closed by @jamie
Save transcript Reopen Ticket Delete Ticket
ArtesiansMay 22, 2021 1:15 AM
Sheet checks out, approved.
Need a +1
Yuehai Ticketing SecretaryBOTMay 25, 2021 10:25 AM
Transcript Saving