KQM | Genshin Impact - lumberjack-tier-list

wiremashMay 20, 2021 4:06 PM

100% success rate?

JenjenJLMay 20, 2021 4:06 PM

yes

wiremashMay 20, 2021 4:08 PM

should be, as long as the tree isn't weird and doesn't exist

it's pine btw, not that it matters

I will keep going through the data, it's looking like she's not actually on the icd though(edited)

but still very close

JenjenJLMay 20, 2021 4:09 PM

but she still has the better range

wiremashMay 20, 2021 4:10 PM

yeah, definitely

Mol | anemo fanMay 20, 2021 4:16 PM

some trees just act really weird. i've been writing it off as a particularly demonic tree

wiremashMay 20, 2021 4:18 PM

demon? tree?

azhdaha leave my trees alone

I suppose it was too late for the spoiler tag

wiremashMay 20, 2021 4:38 PM

I adjusted some stuff (added standard deviation to individual sheets + moved stuff around) if I broke the entire front page please tell me.

ah, I did break the entire front page

Mol | anemo fanMay 20, 2021 4:38 PM

the values are gone ya

wiremashMay 20, 2021 4:42 PM

I'll fix it asap

should be fixed now

fingers crossed

Mol | anemo fanMay 20, 2021 4:42 PM

looks fixed to me

some manual timings not in there but who cares imo for manual

wiremashMay 20, 2021 4:43 PM

oh those are intentional

they're the alt variants

Mol | anemo fanMay 20, 2021 4:43 PM

the type 1 or whatevery ou were calling it?

wiremashMay 20, 2021 4:43 PM

the manual timing is already listed on the regular standard use case

no, stuff like xiao combo cancels, aether n3c, etc.

Mol | anemo fanMay 20, 2021 4:43 PM

oh yeah sorry

i misread yep

looks good

wiremashMay 20, 2021 4:44 PM

manual timing's listed for the regular

yeah

I hope lol

not the first time I broke it xD

shouldn't even be working on the live sheet but it's more thrilling this way

Mol | anemo fanMay 20, 2021 4:44 PM

rosaria caveat bit is "check back soon" rn

lol it's all good

wiremashMay 20, 2021 4:54 PM

yups, I'm about 80% there

with a sample size of 32 how the heck did I still get a 0.5 value for median?

I usually try and get more data to even it out but nah, 32 is more than enough

wiremashMay 20, 2021 5:19 PM

Rosaria is now the queen if we learn to cancel (only 3rd place with normal usage)(edited)

just remember to pray to our anemo archon barsibatos for consistency

wiremashMay 20, 2021 5:36 PM

ping me if there are any mistakes!

AyzelMay 22, 2021 1:33 AM

<@!137414081819836416> is there anything to be added

Aluminum | Harbinger of JankMay 22, 2021 1:46 AM

Eula for one, I'd imagine

AyzelMay 22, 2021 1:47 AM

fair

wiremashMay 22, 2021 4:23 AM

I’m actually wondering if it’s okay to close the ticket and add Eula later. I do plan on updating the list as new characters come out in future for as long as I can. As new characters will keep being released I think it’s better I close the ticket for review at some point instead of waiting for every single character and never closing the ticket(edited)

wiremashMay 22, 2021 4:41 AM

by the time I get Eula's data a new character may have been released, for example, like what happened with rosaria(edited)

JalapenoMay 22, 2021 7:16 AM

Discovered something interesting regarding childe as a lumberjack. Probably doesn’t change his viability much, but thought I’d share anyways

Normally, childe’s melee atks don’t ‘break’ trees: https://imgur.com/a/PUykAvd

Childe lumberjack(1)

However, if you hit a tree with his bow NAs before going into melee, they do break trees: https://imgur.com/a/axCiTnH

Childe lumberjack(2)

AyzelMay 22, 2021 7:27 AM

we have this recorded in #marking-non-enemies-for-skill-hits already

wiremashMay 22, 2021 1:43 PM

and the minimum combo is N5

we knew about this, but because we're 99% sure it's a bug we left it out of the main page(edited)

it's somewhere in the pile of mess I call the changelog/methodology section(edited)

wiremashMay 22, 2021 2:48 PM

hmmm the point of using median was so I could eliminate the weird outliers but now I'm thinking mode would be better instead to additionally eliminate the variances in frames due to recordings not being frame-perfect and find the most common instance... tldr should I use mode or median?

I'd prefer not to switch over because it takes more effort if median is good enough but I'd rather not present incorrect results if I can

wiremashMay 23, 2021 12:21 PM

OKAY SO I GOT EULA

she's the first claymore user to require N4 when chopping trees

(edited)

will gather more data Soon™(edited)

I genuinely wanted qiqi

wiremashMay 23, 2021 12:55 PM

once I finish eula data I will close the ticket

Mol | anemo fanMay 23, 2021 2:10 PM

there is something freeing about finding something to like in standard banner characters so you don't have to feel bad about hitting soft pity. enjoy eula! (but not for chopping trees haha) also, I think you changed the top pick stars on the graph, and they look really good /clear now

wiremashMay 23, 2021 4:12 PM

I did?

I didn't change the stars

but yay

Mol | anemo fanMay 23, 2021 4:20 PM

oh they must be rendered differently in google sheets if you are on ios or windows pc. odd. anyway good to know

wiremashMay 23, 2021 5:08 PM

yeah it's just some unicode thing it's up to the software / device manufacturer to decide what image they're going to provide for it

does it look better or worse on ios?(edited)

wiremashMay 23, 2021 5:36 PM

the worst part about eula is that you can't spam click and get somewhat consistent results like xingqiu

because her slower attacks will queue more easily, meaning you're going to have to click so slowly you might as well manual delay

your spam click would have to be literally so slow it fits the category 2 method:

but worse

so yeah... N4 or manual, no in-between like xingqiu slow spam clicks(edited)

Mol | anemo fanMay 23, 2021 5:40 PM

it's just a little larger / smoother, but both are good

test her CA?

is it the same animation as diluc?

wiremashMay 23, 2021 5:42 PM

as beidou (and diluc)(edited)

might even be the same frames

I'll test it anyway but I won't do much

charged attacks take way too much stamina to be useful

especially if they're continuous like with claymore characters

Mol | anemo fanMay 23, 2021 5:44 PM

oh yeah don't really worry about it i was just curious if there was variation in claymore CAs in terms of anything

like i think albedo and kq share a charged attack

wiremashMay 23, 2021 5:45 PM

I've just briefly tested it

beidou, diluc and eula need 4 CAs

Mol | anemo fanMay 23, 2021 5:45 PM

ok yeah no need to really test her then

wiremashMay 23, 2021 5:47 PM

animation looks exactly the same to me

but I'm just eyeballing it

yeah, I've tested charged 4 beidou and it was 158 frames all the way at the bottom below zhongli

Mol | anemo fanMay 23, 2021 5:47 PM

yeah since it's not any more use even on the generally slow claymores might as well just forget it

wiremashMay 23, 2021 5:48 PM

I don't think, given their very identical animations, eula will be much better

though

based on my initial impressions

I think eula might really be somewhere near the bottom

and she's worse than xinyan when it comes to consistently hitting trees(edited)

literally dances past trees

will have to count frames to know for sure

Mol | anemo fanMay 23, 2021 5:49 PM

well in a way it's nice that you get to close out testing with a totally new unit you actually own haha

wiremashMay 23, 2021 5:49 PM

yeah

it's good I guess

till the new character comes out

Mol | anemo fanMay 23, 2021 5:50 PM

i was surprised doing this how much variation there was in actual usefulness, and i think the list is going to be helpful to people because i generally don't cut trees until i need like 200+ wood

wiremashMay 23, 2021 5:52 PM

I'll just open a new ticket "update to lumberjack data" or something

like what steph did with the elemental particles

tbh I don't really like calling it a tier list

raw speed doesn't make or break any character

all it does is present their raw speed

like. xingqiu is 6th on the list but I consider him possibly #1 when all factors are considered

or xiao who's 8th

they're definitely top 3

Mol | anemo fanMay 23, 2021 5:53 PM

you could present the editor's choice as the front sheet with "raw speed, manual speed" referenced to this sheet

wiremashMay 23, 2021 5:53 PM

make the spreadsheet fancy?

ooh

Mol | anemo fanMay 23, 2021 5:53 PM

in the partial experience i have i agree that xq is just so consistent and nice

wiremashMay 23, 2021 5:53 PM

with their character portraits and stuff

Mol | anemo fanMay 23, 2021 5:54 PM

like just take out your choices, put them on a page "best lumberjacks" then sheet 2 be raw table

but putting their numbers in sheet one

wiremashMay 23, 2021 5:54 PM

the raw table was supposed to be the summary

Mol | anemo fanMay 23, 2021 5:54 PM

yeah

wiremashMay 23, 2021 5:54 PM

I guess we can do a summary of the summary

Mol | anemo fanMay 23, 2021 5:54 PM

i think it's good as it is

but

if you wanted to recommend xq for example i think that's the cleanest way

wiremashMay 23, 2021 5:55 PM

I'll have the top 4 in a grid then, complete with their character pictures

maybe

I still need to learn how to use google sheets

idk how others make their sheets look so fancy

Mol | anemo fanMay 23, 2021 5:55 PM

me either lol i saw some of a stream like oh wow cool, excel doesn't have to look quite so businesslike haha

wiremashMay 23, 2021 5:56 PM

yeah same

but having pictures and just four large boxes would definitely make things clearer

it's gonna take a while

my mind is blank

any ideas would be great

Mol | anemo fanMay 23, 2021 5:56 PM

lol or you could just close it and be free

wiremashMay 23, 2021 5:56 PM

yeah that's true

Mol | anemo fanMay 23, 2021 5:56 PM

you've done the tough stuff and it's good

wiremashMay 23, 2021 5:58 PM

I've done the tedious stuff

the tough part is actually presenting the data

in a way that people can actually understand

something like that

idk

I guess it's okay as it is as long as nobody thinks it's a tier list

but given how the eula release went I'm starting to doubt it

Mol | anemo fanMay 23, 2021 6:01 PM

yeah i think something like a "Good Picks" page on the front would prevent people from defaulting to "highest = best" like we thought was funny about a manually timed ZL

like yeah ZL is technically the fastest possible lumberjack if you hate yourself

wiremashMay 23, 2021 6:01 PM

haha definitely

though keqing's only 1 frame behind

they're within error

Mol | anemo fanMay 23, 2021 6:02 PM

also i appreciate that you dropped kaeya, i have started swapping to xq to cut wood when he is in my team

kaeya's swings are just not as comfortable angle wise

from recommended

wiremashMay 23, 2021 6:02 PM

yeah the only reason I had him was because of his movement speed

but lots of other characters can also move fast

Mol | anemo fanMay 23, 2021 6:02 PM

i was a big fan of him because i was doing something normal people will not do probably, running really skewed loops where i was hitting 8/11 trees one time only to dash between them

wiremashMay 23, 2021 6:03 PM

and you could just switch to kaeya while running anyway

well, maybe this isn't very practical either

Mol | anemo fanMay 23, 2021 6:03 PM

kaeya is the only dash-spam character that also had 20% stam consumption reduction; i found it was impossible to do the 11 tree loop without stopping to regen stamina using anyone else

wiremashMay 23, 2021 6:03 PM

yup, but the buff is party-wide

Mol | anemo fanMay 23, 2021 6:03 PM

but like i said i think this is a really weird use case

wiremashMay 23, 2021 6:03 PM

so all you need is him in the party to get that buff

Mol | anemo fanMay 23, 2021 6:03 PM

so dropping the recommendation is fine

wiremashMay 23, 2021 6:04 PM

xiao, kaeya and some other anemo is probably good

add a xingqiu for the smaller trees and you have a great chopping team

Mol | anemo fanMay 23, 2021 6:06 PM

yeah, just for clarity on why kaeya, i think the fastest movement from tree to tree overall is slow dash spamming with the tallest male character model, so that's why i was using kaeya in testing. it's posted in tips and tricks library https://www.youtube.com/watch?v=d3sqmqSkw44

uNEven.

Genshin Impact - 7 Comparisons of Sprinting

From Left to Right: - Dash spamming with correct timing - 2 Dash spams in a row - 2 Dash spams, alternating through each party member, then swapping to the 1st party member and dash spamming with correct timing - Teenager bunny hopping - Mona sprint - Adult bunny hopping - Adult sprinting The 2 dash spam alternating party technique could pot...

wiremashMay 23, 2021 6:06 PM

oo interesting

Mol | anemo fanMay 23, 2021 6:06 PM

but i just wanted to say i think the 3 of the final 4 i can use are fantastic and good choices

wiremashMay 23, 2021 6:07 PM

it's flexible

Mol | anemo fanMay 23, 2021 6:07 PM

and ultimately i support dropping kaeya as a recommendation from the list

wiremashMay 23, 2021 6:07 PM

xingqiu could be swapped for keqing

rosaria and xiao could be swapped

or some other character not on the recommendations

Mol | anemo fanMay 23, 2021 6:07 PM

yeah they're all different and not everyone will have them

wiremashMay 23, 2021 6:08 PM

they're not all bad

yup, the numbers are just to help you decide, not to tell you what to use

some people will prefer using a character solely for the fact they're in their current party

Mol | anemo fanMay 23, 2021 6:09 PM

yeah i think the full list is amazing because you might be running a much better character already in the party and just be causing pain for yourself chopping trees with a slow or weird AA series that doesn't work for it

wiremashMay 23, 2021 6:14 PM

imagine not using manual timing for zl

actually that one's probably really obvious

maybe more kaeya or jean

I can't believe I'm about to put a [Manual timing recommended] on a claymore user

wiremashMay 23, 2021 7:16 PM

I'm wondering if there's a mob that shares the same ~~icd~~ HITLAG as trees(edited)

I imagine it'd be pretty useful for future theorycrafting and frame counting if they can just hit a defenseless tree instead of finding an unhappy enemy who fights back

wiremashMay 23, 2021 8:23 PM

wow. Eula is worse than charged attack beidou

also don't know why I have an insane 104 samples

wiremashMay 23, 2021 8:32 PM

manual timing is definitely not worth the effort either as it still puts her slower than xiangling doing her full N4

wiremashMay 23, 2021 8:47 PM

I still need to finish doing a massive cleanup of the changelog/evidence/methodology sheet so I can't close it yet (or can I?), but I have all the data fully complete(edited)

wiremashMay 23, 2021 8:56 PM

remember this?

Mol | anemo fanMay 23, 2021 9:58 PM

hahaha

wiremashMay 24, 2021 8:45 AM

https://docs.google.com/spreadsheets/d/1Q4HKzkaw7YFNZyIJSjRINZJonFe1kiDwz1dEbpcZiNk/edit?usp=sharing Just gonna link it again I've cleaned up the methodology/evidence/changelog/credits section and put them into separate sheets. Recordings are now on the characters' individual data sheets complete with timestamps. Hopefully, it's good enough now for anyone to clearly understand how the data was gathered (and point out flaws in my methodology)

Lumberjack

Introduction Lumberjack Editor's Personal Recommendations Manual Xingqiu,Frames,Rosaria,Frames 83,91 Seconds,Seconds 1.383,1.517 Attacks,Attacks 3,3 Overview,Overview Xingqiu's wide swings and relatively stationary attack animations,One of the fastest characters. Her long reach helps her hit tre...