MAX

[supple zenith]

Introducing MAX
the fabled framewar tab of legend

The Natey+Slomac+16 guest mappers collab tab of 125 bitesized levels built for speed. With live frame interpolation, a shiny new leaderboard mechanic from Eddy. And preset Dev Times to put your skills to the test.

Live frame interpolation:

When you submit a score while playing MAX, the score will be automatically interpolated to the one-thousandth (0.xxx) of a second. This will show in-game on the leaderboards after submission. This means that competition is never truly over until every last iota of optimization is achieved

In addition, upon submission the “Global Rank” number that usually displays will now display the number of microseconds left until the next frame. This way you can tell if you are improving within a certain frame.

Dev times:

You will notice the levels have no gold to collect. Instead, completion is marked by exceeding all dev times. These times are displayed on the leaderboard for each level, noted as -- DEV TIME --

Author credits:

Slomac
NateyPooPoo
audrey
Wobert
Darkstuff
frankytrees
Skylighter
SuperVolcano
abho
Ellsworth
filip
aloading
ekisacik
Hamwich
kkstrong
lily
paul
w95559w

Huge thanks to all who contributed, and of course to Eddy for working his crazy voodoo magic.

⬇️ Download link

📄 Author Credit Spreadsheet

✅ Completion spreadsheet

[tranquil fiber]

@muted scroll

[oblique bear]

have these been made for non-pros to be able to beat?

[tranquil fiber]

this looks fucked up in so many ways. subframe scoreboards? completion is beating slomac? what even how even why

(yes emm, all dev times should be beatable)

[oblique bear]

i don't mean beating slomac, but beating the levels at all

[tranquil fiber]

and in that case very very very much yes, they're ez

...or so i hear. i've seen a lot of what i assume are rejected slomac maps in userlevels, and they are indeed trivial

[supple zenith]

Yeah the levels are ezpz lemon squeezy

[vestal acorn]

The levels are very easy for completion. Probably the easiest tab for that

[supple zenith]

For the dev times.. this is a really experimental thing, so they are set pretty modestly. The skill range for highscoring is insane in this game, so some will probably think they are way too easy, and some will probably find them difficult

[vestal acorn]

No, none of those are the rejects lol. Will post those later.

[tranquil fiber]

WHOA

[supple zenith]

Definitely easiest tab for standard completion

[tranquil fiber]

i was so sure haha

i guess i have to play this. hm. my internet has been angry so im not sure i can stream, but i would like to

[supple zenith]

Definitely curious to see how someone approaches the dev times 👀

This will be a learning experience for sure

(So plz don’t hate me if it sux)

[tranquil fiber]

u know i love highscoring hahahahaha

[supple zenith]

[frigid summit]

This is interesting, might finally beat my laziness and have me learn how to play cusotm tabs.

[supple zenith]

It’s super easy. Basically a one click install, just have to bypass the windows defender message

[tranquil fiber]

oooooo

[supple zenith]

Lmk if that gives you any trouble

[frigid summit]

Is it as straightforward to go back to my base game save afterwards?

[tranquil fiber]

breakfast time, will tentatively try streaming in ~two hours

[supple zenith]

Yep, one click uninstall. Saves are automatically swapped for install and uninstall

[frigid summit]

Sick, gonna try it out after a nap.

[zenith owl]

wait, does this mean the dev times will disappear if the leaderboard fills?

[supple zenith]

If all scores are over the dev times yeah. Probably unlikely though?

I’m sure we can figure out a solution if it becomes a problem tho

[muted scroll]

[supple zenith]

NateyPooPoo is a dirty liar

[muted scroll]

How dare Natey not give author credits to Dalton

[supple zenith]

Honestly it was a great idea and we should have done it

He’s easily the most prominent framewar mapper

[north barn]

Oooh, bonus points for being closer to the exit when you enter it... I like that

(technically it also factors in velocity, but only on the frame that you make it in...)

[outer swift]

yo @tranquil fiber you should pin the big message

[tranquil fiber]

there's a scroll to top button now. not in here because presumably its short still? idk. can still pin if people prefer that

[outer swift]

i think it's still useful to have pinned

[outer swift]

also, raigan reacted to the announcement post. we should have him set the dev times

[fickle ether]

Genuinely my antivirus just disallowed me from extracting the .zip and then deleted it from my downloads >_>

Which ironically sounds itself like malware

[supple zenith]

If you open windows security from the bottom right of the start bar you can undo the deletion

Or set a protected folder, then download to that

[fickle ether]

Alright got the pack back, but how do I extract it now?

[supple zenith]

Right click and “extract all”

If it disappears you can undo it the same way. Windows security > protection history > the folder should be in that list, iirc

[fickle ether]

Oh wow yea I didn't uh
Test just extracting again

I figured it would stop me

Well it has worked!

[supple zenith]

Nice

[fickle ether]

ty natey

[fallow thorn]

was going to say this. I want Raigan times for all of these levels

[fickle ether]

Mann, trying to find an nprofile that doesn't give invisible tiles but also doesn't have cached Intro times

[supple zenith]

Ohnooo

[fickle ether]

Eh, I'll have to live with some cached intro times from my Metanet campaign.

I remember the profile I used for REDUX tab, which was only a few blushes from the profile Wheaty constructed to begin the tab with everything available to play but no scores, did it, but I can't find that no more.

[fallow thorn]

just installed, so should I copy a blank nprofile in before starting it, you're saying?

[supple zenith]

No need. The installer does that for you

[fallow thorn]

yup just opened and it's fine

[supple zenith]

DS has a weird case where new profiles make all his tiles disappear

[fallow thorn]

you're all making me leave my air conditioned bedroom to play framewars smh

[supple zenith]

[fickle ether]

heeheh, got exactly the dev time on a-00-01

[muted scroll]

Dev times are fairly challenging

Definitely taking a good number of attempts

[supple zenith]

Fairly challenging to kkstrong. So far so good

@late wyvern Max sr score rank

[late wyvern]

Rankings - Speedrun level scores for Max mappack

-# Date: Jun 19, 2025 (17:43) UTC. MIN: 58567.

00: Slomac         - 60946 (125)
01: CeramicBlue    - 62557 (125)
02: NateyPooPoo    - 38316 ( 80)
03: vertexv        - 16255 ( 35)
04: DarkStuff?     -  2397 (  6)
05: Sky!           -  2124 (  5)
06: kkstrong       -  2230 (  5)
07: Nim “Mr. 7280” -  2432 (  5)```

[supple zenith]

Slo and bobio crossing the finishing line for base completion

Probably less than an hour for both

[muted scroll]

Is there an outte command for scores above dev time?

Got the first episode done

[supple zenith]

Hmm, you could probably use the compare function

But no rankings or anything at the moment, that I know of

[muted scroll]

I did some testing in DMs and couldnt get it working

[supple zenith]

I like that the levels are so simple that I can swap through palettes without needing to adjust

[supple zenith]

That’s a lot of milliseconds

You are 10 seconds behind the next frame. Skill issue

[muted scroll]

Im lost

[supple zenith]

I mentioned this in the post but maybe I could have made it clearer. The “global rank” now shows how many milliseconds you have left until the next frame. Useful for seeing how your score improved, even if the timebar didn’t improve (because the timebar only shows 60fps)

I’m guessing this was an episode run?

[fallow thorn]

mine shows giant numbers like that as well

[supple zenith]

Hmm, maybe it’s a smaller denomination than milliseconds

We’ll have to wait for Eddy to pop by and clarify

[old lagoon]

This is a super cool idea with the leaderboards

[supple zenith]

It might be microseconds instead of milliseconds that shows.

Since the interpolation system in outte uses six digits

@late wyvern scores for maxsia00

[late wyvern]

Interpolated highscores for Max (MAX-SI-A-00-00) by Slomac:ansi 00: Nim “Mr. 7280” - 84.204272 01: NateyPooPoo - 84.195410 02: Sky! - 84.187259 03: kkstrong - 84.181676 04: Slomac - 84.088238 05: CeramicBlue - 84.049931 06: DarkStuff? - 84.002587 ++: -- DEV TIME -- - 84.000000 07: vertexv - 83.739624Scores: 8.

[old lagoon]

Does the dev time show in-game as well?

[fallow thorn]

yes

[old lagoon]

Sick

Honestly props to you for getting Raigan on board to set all these times

[muted scroll]

I dont think raigan was involved…

[old lagoon]

Even more impressive they got Mare to play and set the times!

[supple zenith]

Believe it or not she kept wanting to use DI strats in the dev times. She’s a devious one

[fickle ether]

Dev times for all of column 00 done :)

I am cheating slightly, I don't have to play the episodes, I already have all the levels unlocked

[supple zenith]

I’m glad your able to play and submit scores at all

Ok, went through just to make sure every dev time had at least one score above it. Gonna take a break before the real sweating begins

[fickle ether]

Went back and got episode scores just for fairness' sake

@late wyvern max rankings

[late wyvern]

Rankings - Highscore overall 0ths for Max mappack

-# Date: Jun 19, 2025 (18:54) UTC. MAX: 150.

00: Slomac         - 59
01: NateyPooPoo    - 37
02: CeramicBlue    - 26
03: Nim “Mr. 7280” - 12
04: vertexv        - 12
05: DarkStuff?     -  4```

[fickle ether]

Stopped for some 0ths on the way

Now I must depart

[tranquil fiber]

baiiiiiii

[oblique bear]

it can't be that hot where you are, can it?

[fallow thorn]

its 83 the grease

[frigid summit]

Nothing broke chat.

Just one question: is there a way to retain my settings while installing a custom tab?

Like music/sound volume etc.

Also, my score won't submit.

I restarted the game, it works now.

[supple zenith]

You’ll have to redo volume settings. I think key settings stay the same

[tranquil fiber]

i use the gamification profile that sam shared a long time ago, and finally remembered (thanks pralk) to save a copy with volume adjusted. so i just start with that profile every time instead of a fresh one, so i have all the palettes and stuff

[muted scroll]

dev timed the first column as well

This is good stuff

[supple zenith]

Beginning the poll on whether dev times suck or not:

They don’t suck
kkstrong
vertex
DS

They do suck
Audrey

[vestal acorn]

What's the issue with the dev times?

[tranquil fiber]

im bad at game

[supple zenith]

It’s always kind of a guarantee that some will find it easy and some will find it hard. The range of skill levels and the amount of little techniques for highscoring is just crazy for N

[frigid summit]

Damn, just finished playing CS2 with friends, hoped to catch the stream.

[tranquil fiber]

dw, it was not an inspired stream. some pretty fun levels tho, i think you'll enjoy them.

[analog plover]

https://tenor.com/view/just-when-i-thought-i-was-out-they-pull-me-back-in-michael-corleone-al-pacino-the-godfather-gif-19249100

[supple zenith]

@late wyvern scores for maxsic01

[late wyvern]

Interpolated highscores for triangularism (MAX-SI-C-00-01) by Ellsworth:ansi 00: NateyPooPoo - 85.663528 01: Nim “Mr. 7280” - 85.655500 02: Slomac - 85.643626 03: kkstrong - 85.606933 04: DarkStuff? - 85.546476 05: CeramicBlue - 85.540052 ++: -- DEV TIME -- - 85.500000 06: vertexv - 85.140332 07: auds - 78.899329Scores: 8.

[supple zenith]

Sub-framed

[storm burrow]

they don't suck

[supple zenith]

Thank you sir, your vote has been recorded

[fickle ether]

I concur they do not suck

[supple zenith]

It’s really amazing how seamless the interpolation is. I thought there would be a little delay but it’s instantaneous. Even when we were all going at it earlier

[pastel beacon]

Good morning

Completion spreadsheet. Columns for what? Beating all dev times?

Cool idea for a tab. Will have to take a look when I can

[viscid fox]

[supple zenith]

Yeah all dev times for sure. I guess base too? It’s quite easy but should probably have it anyway.

[fallow thorn]

that's nothing, watch this

milliseconded

[viscid fox]

i shall be here waiting for schmole's video series

[tranquil fiber]

shiiiii

[viscid fox]

(which he is probably gonna get it done in like 2 hours)

[pastel beacon]

Done

[pastel beacon]

Might put it to the front of my personal time plans - I'm too many packs behind atm x.x
Will try and play Monday evening

[muted scroll]

Huh, its interesting that the rankings on the leaderboardin game doesnt always match outte's leaderboards

@late wyvern spoiler screenscores for max-si-a-1-4

[late wyvern]

Screenshot for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo v1 in palette vasquez:

Interpolated highscores for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo:ansi 00: CeramicBlue - 84.614369 01: kkstrong - 84.614164 02: Slomac - 84.614039 03: Nim “Mr. 7280” - 84.613891 04: NateyPooPoo - 84.611961 05: vertexv - 84.611726 ++: -- DEV TIME -- - 84.600000 06: auds - 84.598571Scores: 7.

[muted scroll]

Nim is top spot, and not even .614

[fallow thorn]

?

I see Slomac as the top

[muted scroll]

Oh weird we even see different things locally

I reloaded and now mine shows that

Maybe Im just woozy

[supple zenith]

Probably because the game leaderboards can only show to the millisecond, and outte shows to the microsecond

Eddy probably let ties work normally in-game, to be fair

He’s gonna be a busy fella answering all these questions when he hops on today

[north barn]

@digital socket i love you

[digital socket]

WAIT I DIDNT MAKE THAT JOKE

HAHAHAHA

[supple zenith]

Flipped it on em

[tranquil fiber]

glorious

[peak grail]

what nightmare zone shit is this

we've gone beyond framewars now?

we're doing subframe wars?

yall need jesus

[old lagoon]

@peak grail you know you want to

[peak grail]

i straight up do not

no

[old lagoon]

But every level could be the marinara trench

[peak grail]

i could just kill myself

those are both options

that are out there in the universe

[old lagoon]

Or at least go for dev times

[peak grail]

see sec. 2 above re: killing myself

i know the devs

they're psychopaths

i don't want any piece of those times

[muted scroll]

Ive gotten a handful of times better than dev on first completion

Its pretty rare though

[old lagoon]

Important question - were DIs used for any dev times?

[vestal acorn]

no

I thought the dev times were fairly generous. Though there are some noticeably tighter ones

[north barn]

yeah no DIs please LOL

[supple zenith]

No DIs or triples or anything that requires mash strats. Not now not ever

[north barn]

triples are lenient

unless you get <1 block of run-up

[supple zenith]

Triples are BANNED

In my tabs

And tabs I collaborate on

If I’m on the board of directors. NO TRIPLES

[supple zenith]

Next tab I release? Marinara trench. 500 times

[peak grail]

I was made for this life

Put me in coach

I was born ready

[supple zenith]

If you can’t handle me at my framewar tab then you don’t deserve me at my marinara trench five hundred times tab

[fallow thorn]

Smash that like if CCS needs to play the Max tab 👇

used the Apple Genmoji

[oblique bear]

columns or rows?

[supple zenith]

Anything goes

But columns always intended

[north barn]

its gone in a flash

[oblique bear]

ok i'm stuck

[muted scroll]

Bruh.

[supple zenith]

Ngl I think there might be a couple sticky spots for emmoji. What level?

[muted scroll]

Beating dev times for sure, but I havent seen anything that wasnt super straight forward yet

[oblique bear]

si-c-1-4

[muted scroll]

I think I've only encountered like.... 3 drones lol

[oblique bear]

drones are fine

[supple zenith]

Easy route on the right

[zenith owl]

@late wyvern spoiler screenshot for max-si-c-01-04

[late wyvern]

Screenshot for get up there, but don't cheat this time (MAX-SI-C-01-04) by Slomac v1 in palette wyvern:

[oblique bear]

ok got it

[supple zenith]

@vestal acorn @cunning tide there was a typo on si-d-4-3 dev time. It’s been updated if you want to redo your scores for that one

[north barn]

wait this is so pretty multi exit my beloved

[muted scroll]

Did the multi-exit levels come before or after my guest map?

[supple zenith]

Chronologically or in the tab?

I still cannot figure out how to get that c-switch

Spoilers!

That is actually it lol

[north barn]

first place i looked

the instincts never leave B)

[supple zenith]

@mods plz delete

[muted scroll]

chronologically

[tranquil fiber]

was that actually actually it???

[supple zenith]

Yeah haha

[tranquil fiber]

omg im so good

wait i thought u didnt know what it was

[supple zenith]

That was like the fourth map Slomac submitted

[muted scroll]

Ive seen alot of multi-exits and think its funny that mine happened to be so also

[tranquil fiber]

okay but mine was definitely the first

[supple zenith]

I didn’t! But I tested it after you showed it lol

[tranquil fiber]

LOLOLOLOL honestly nice work that was fast, it looked impossible to me haha

[muted scroll]

I appreciate that on my guest map the current demos hit 3/4 exits

[tranquil fiber]

okay could multiple exits cause subframe inaccuracies

cuz i think yes

[muted scroll]

explain

[tranquil fiber]

just read my mind plz

if ninja is going at a certain speed/angle they could end up touching zero exits one frame, and two exits the next. the game chooses which exit is triggered based on the grid, but it's possible u were like closer to the other one. i dont actually know how the interpolation works, so i could be way off base

[supple zenith]

I get what you’re saying

i imagine it uses the exit you touch

[north barn]

After reading Eddy's post on how hitting a switch technically doesn't let you hit the door until the next frame, I thought of the funny scenario where:

• you move extremely fast into two stacked door switches
• the door for one of them is stacked on the switches
• the door for the other is somewhere far off where the ninja goes, next frame
  so what ends up happening is that you end up hitting a different door in order to win despite being on one when the frame started

[tranquil fiber]

LOL

[split patio]

I think that's indeed what would happen, lol

[tranquil fiber]

that, at least, wouldn't cause interpoloation problems! i think!

[muted scroll]

Uh oh he’s here

[split patio]

I'm currently not even sure about the proper answer to this question. The fractional frame gets saved when you touch the open door, but I believe that doesn't stop the simulation of that frame, it just means it'll be the last frame to be simulated. So then you go on to touch the second exit, and it overwrites the fractional frame.

So my gut tells me it'll be the last door you touched on that frame, as opposed to the first one, which is what we'd want. It's very fringe, but I'll note it down, it's an easy fix.

[split patio]

I finished up reading everything and got several questions to answer, lol

[north barn]

It's funny because then you can bring in the much simpler case of clipping through a door and then hitting another door in the same frame

[supple zenith]

this is the relevant one i'm thinking of

[north barn]

I should've asked this first actually, way more interesting question

[split patio]

Pretty confident if you clip through a door the interpolation won't catch it either, because it interpolates the time when you collide with it. It's not spatial interpolation, like the one the game does with tiles.

You'd clip through it in the game as well though, so I think we want this behaviour.

[split patio]

You're right, it could definitely be relevant here. I've thought of an easy fix, which is to only save the fractional frame if it's lower than the current one (if there already is one). That way at the end of the frame you know you have the first door you touched, regardless of the order in which they're checked.

[split patio]

Yeah this is the root cause. I'm actually sending the leaderboards in the same order as in outte, but ties are re-sorted automatically by the game, probably in order of replay ID (so as to be chronological, which is what you'd normally want).

Not sure I can do anything about this, since I need the IDs to be correct so that the game can request the proper replays. However, you can still improve normally and see the real scores and the real order in the boards outte shows on Discord, which will always be accurate, and thus what counts for rankings and whatnot.

[muted scroll]

Eddy’s answer, dont forget you can just git gud

[split patio]

Just get one more frame and problem solved bros

[supple zenith]

One more millisecond even. How hard could it be?

[split patio]

No way that's even relevant in any level!

[muted scroll]

Micro*

[split patio]

Well, to avoid the in-game tie issue, you need a full millisecond, since that's all I'm sending to the game. I'd send the microseconds as well, but I cannot move the decimal period, it's added automatically by the game, so 3 decimals are enforced.

[digital socket]

I probably won't play this tab myself, but I added it to the tierlist
https://tiermaker.com/create/n-custom-tabs-17807640

[split patio]

You're only banned from participating in the project, not playing the tab 😔

[digital socket]

I'm allergic to highscoring

[muted scroll]

Why you no play ascend

[supple zenith]

It takes like an hour to playthrough the levels real quick

Highscoring isn’t mandatory

[muted scroll]

inb4 Natey makes a hardscore tab

[supple zenith]

I mean it isn’t a casual gameplay masterpiece or anything but there’s some entertainment I’d like to think

[split patio]

Yeah, worth it for sure, and playing against the dev times is also a fun incentive

[digital socket]

i did play ascend!

[oblique bear]

on si-e-2-3, ||was it intended to get the c switch? i couldn't figure out how (but i beat the level)||

[oblique bear]

i'm just playing through the levels casually for now

[split patio]

Yeah I think the dev times should be challenging to the average player, otherwise they'd be kinda pointless and not fun. I still have to go through most of the tab, but from what I've read here I feel like they were well adjusted, because I see many scores above them but also several people "complaining" about them 😄

[digital socket]

I might at some point then, my impression is they weren't made for a casual experience

[supple zenith]

Bro there’s literally a level called “the greatest level ever made”. With promises like that who wouldn’t want to play this!

[digital socket]

the greatest level ever made is every level I make 😎

[oblique bear]

most are on the easier side, which imo is a good thing because most mappers tend to fill their map packs with levels that are harder than metanet late ultimate to base complete

[muted scroll]

I think that was once true, but the last 10ish packs that is innaccurate

[split patio]

It is true for Newest content though

[digital socket]

yeah there's been quite a lot of casual tabs, which is great

[split patio]

But packs are of course heavily above-average quality-wise

[digital socket]

though even those can include tricks that the mapper thinks is easy for everyone, but isn't actually

[split patio]

Indeed, but that's pretty much inevitable given the wide range of skills that this game accepts, like Natey said

I don't think base completion of this pack requires any trickery though

But I haven't played yet

[digital socket]

it doesn't, from what I know

[supple zenith]

I don’t do trickery, tomfoolery, or any of the sort

[zenith owl]

i tried to make sure the DIs and triples were concentrated into a single level :)

[split patio]

Should be a very light experience if one is just looking to base complete or even polish their scores a bit

[supple zenith]

Have you ever heard of “emmoji’s law”? It’s states that no matter how hard I try to make every piece of every level approachable, emmoji will find a way to get stuck

(I mean this in the most endearing way possible emm)

[oblique bear]

either this or i end up finding it easier to do 3 minejumps up a wall than whatever was actually intended on the level

[supple zenith]

lol, that too

[oblique bear]

or couldn't figure out what the actual intended method was, so did something weird

[split patio]

Regarding this, I thought about putting it in milliseconds as well, but I figured microseconds would give better granularity, specially for the tighter maps like the aforementioned greatest map ever made.

I wouldn't mind changing it if people prefer it or find it more clear. Milliseconds would only go from 0 to 16, whereas microseconds go from 0 to 16666.

[oblique bear]

speaking of which

[supple zenith]

Nope, that’s just a little Easter egg as far as I know

[supple zenith]

I’ve started grasping a feel for it the more I get eyes on it

[oblique bear]

si-d-3-4 level name =

[split patio]

Another reason to use microseconds was to give more information than the in-game leaderboards. If you improve your milliseconds, you'll be able to tell on the boards. But if you improve by less, you won't.

[muted scroll]

You could say the level name is on the money

[supple zenith]

True

[zenith owl]

neon white displays milliseconds but uses microseconds

it leads to some funny tie moments

[split patio]

I guess you could argue our system uses arbitrary precision (probably taps out around the nanosecond range though), but displays microseconds in outte and milliseconds in-game.

Unfortunately, the game re-sorts ties, and since it's only aware of the milliseconds, we get bad orders like in the greatest level ever made.

When in doubt, e.g. ties up to the millisecond, resort to outte to see the proper order.

This order will be correct even if there's a tie up to the microsecond, even if you can't see the actual difference displayed.

Truly equal scores are marked with an = sign. Not sure if any of those has popped up already.

[supple zenith]

Given the limitations it certainly works well enough. At least it stays chronological in game

And outte knows the truth for ranks and such

[supple zenith]

I wonder if we’ll see one 👀

I can think of a potential level

[split patio]

Not yet. The dev scores aren't actually implemented as real scores in the database because that could've interfered with other functions and I didn't want to risk it or have to exclude them explicitly. So the "dev" isn't a real player and you can't compare your scores the way you normally would with the comparison function.

[muted scroll]

Hard coding the command would be an easy workaround

[split patio]

Yeah, I'm sure it'll be simple to implement, I'll think of a good way to do so later so that we can check the requirement automatically

[zenith owl]

do you only get one chance at the interpolated max per millisecond? since i remember for metanet servers it doesn't update your score if you get an identical time

i feel like i'm getting better runs on the greatest level ever and it won't update my score

[split patio]

Nope, it should save any improved score, even if it ties up to the millisecond

Fortunately, the game submits every score, and it's up to the server (me) to decide whether to save it or not

Let's check something

@late wyvern max scores for the greatest level ever

[late wyvern]

Interpolated highscores for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo:ansi 00: CeramicBlue - 84.614369 01: kkstrong - 84.614164 02: Slomac - 84.614039 03: Nim “Mr. 7280” - 84.613891 04: Sky! - 84.613890 05: Eddy - 84.612616 06: NateyPooPoo - 84.611961 07: vertexv - 84.611726 ++: -- DEV TIME -- - 84.600000 08: auds - 84.598571 = 09: Emmoji - 84.598571 =Scores: 10.

[split patio]

I think you're missing 2776 microseconds to the next frame, so you'll know you've improved if the game returns a value under that

[zenith owl]

it's displaying as .614 to me

and everyone above me too

[split patio]

Yeah I know, see this ^

Since I only have 3 decimals to play with in the game, I send the scores rounded to the milliseconds, and even though I do so in the correct order, the game actually re-sorts the boards.

Ties, specifically, in chronological order (i.e. by replay ID)

[supple zenith]

By “next frame” he means the .633. So if you’re improving you should be closing that gap.

Hence the microseconds to next frame would decrease

[split patio]

Indeed

Funny enough, it's the amount of microseconds left to actually get the .617 for real, but at that point the game would start rounding it to the .633

[zenith owl]

does the global rank show the microseconds?

[split patio]

Yep, that's microseconds left till you break the next frame

So the lower the better

It can range between 0 and 16666.

I could use milliseconds left, which is simpler, but then you lose a lot of granularity which, for levels like this one particular, is required.

[muted scroll]

Funny how maxed command doesnt work on MAX pack

[split patio]

Yeah haha, that's a side consequence of storing the real scores, I guess you could argue a level isn't truly maxed (or even maxable) if the ties aren't actually perfect here.

[muted scroll]

That level is still maxable, since the board has tied scores as displayed

[split patio]

Yeah, but internally I store the scores with complete precision, and most importantly the rank field (which is precomputed so that most queries are faster).

So at the very least I'd need to make significant adjustments to how it works. Will note it down though.

There's also another way to temporarily fix the problem:

[north barn]

HAHA

[split patio]

Got this while testing the pack so I knew it was possible

1993 microseconds left, for reference.

Given the limitation of the actual inputs you can have (3 keys, only 60 changes per second), you probably can't reach most microseconds. Meaning that I wouldn't be surprised to see perfect ties in some of these values.

[muted scroll]

https://tenor.com/view/im-feeling-rage-rumi-robinson-imurgency-im-feeling-angry-i-can-feel-my-wrath-gif-1989701579615248681

[oblique bear]

is si-e-4-3 supposed to be ridiculously hard?

[north barn]

@late wyvern screenshot of max-si-e-4-3

[late wyvern]

Screenshot for my little meat tornado (MAX-SI-E-04-03) by Slomac v1 in palette quench:

[north barn]

oh THAT'S funny

looks like a bunch of setup jumps to me though

just touch the wall and fall down without moving and buffer a jump? :raiganthumb:

[supple zenith]

Hahaha, lovely

[supple zenith]

The ‘line up then corner jump strat’ is the easiest imo. Hard to explain probably best to watch some replays

[north barn]

oh, do you have to tap toward the wall a little after falling to get a lowish corner?

-# mfw trying to play a level in my head without opening the game

[oblique bear]

1/255

first map pack i completed in 1 day

and only was stuck to where i needed help once

though this was probably not what i was supposed to do here

i couldn't view replays, i was in an episode

[fickle ether]

You made plenty of those manoeuvres harder on yourself by moving away from the wall instead of just sliding down :P

Congratulations! \o/

[north barn]

gg on the clear!

[muted scroll]

lol

Everyone: use setups to make it easier on yourself

Emmoji: Nah, I do what I want

[frigid summit]

I've done three episodes so far, left a few dev times for later, but it's very, very fun.

@late wyvern interpolated highscores for MAX-SI-D-00-03

[late wyvern]

Interpolated highscores for temu version of slomac's '25 handplaced e blocks' (MAX-SI-D-00-03) by NateyPooPoo:ansi 00: pralkarz - 84.924988 01: Sky! - 84.915191 02: Nim “Mr. 7280” - 84.913646 03: Slomac - 84.908145 04: CeramicBlue - 84.902197 05: NateyPooPoo - 84.895805 06: DarkStuff? - 84.887724 07: vertexv - 84.875897 08: aight - 84.862246 09: kkstrong - 84.853947 10: Eddy - 84.852754 ++: -- DEV TIME -- - 84.850000 11: auds - 84.765516 12: Wobert - 83.645872 13: Emmoji - 82.132306Scores: 14.

[frigid summit]

[split patio]

Btw no need to add "interpolated" for this pack in particular, they will show by default

Also you can use "fractional" or even just "frac"

[frigid summit]

I was about to suggest to make that a feature (for this pack).

But you're three steps ahead already.

[split patio]

[gaunt ingot]

I've cleared all levels now + dev times on my own levels, will try some more dev times later 👍

So the interpolated times feature - is that using the recreated N++ physics to sim on replay submission?

[zenith owl]

@late wyvern scores for max-si-d-0-3

[late wyvern]

Interpolated highscores for temu version of slomac's '25 handplaced e blocks' (MAX-SI-D-00-03) by NateyPooPoo:ansi 00: Sky! - 84.925623 01: pralkarz - 84.924988 02: Nim “Mr. 7280” - 84.913646 03: Slomac - 84.908145 04: CeramicBlue - 84.902197 05: NateyPooPoo - 84.895805 06: DarkStuff? - 84.887724 07: vertexv - 84.875897 08: aight - 84.862246 09: kkstrong - 84.853947 10: Eddy - 84.852754 ++: -- DEV TIME -- - 84.850000 11: auds - 84.765516 12: Wobert - 83.645872 13: Emmoji - 82.132306Scores: 14.

[zenith owl]

i will not stand for this lmao

[frigid summit]

@late wyvern traces for MAX-SI-D-00-03

[late wyvern]

Sorry, I didn't understand your command.

[frigid summit]

@late wyvern highscore traces 0 1 for MAX-SI-D-00-03

[late wyvern]

Sorry, I didn't understand your command.

[frigid summit]

FUCK.

[zenith owl]

@late wyvern anim 0 1 for max-si-d-0-3

[late wyvern]

Replay highscore traces for pralkarz and Sky! in MAX-SI-D-00-03 using palette wyvern:

[frigid summit]

Ok, our routes are the same, time to hunt frames later today.

[gaunt ingot]

@late wyvern scores for max-si-c-1-0

[late wyvern]

Interpolated highscores for garden warrior (MAX-SI-C-01-00) by Wobert:ansi 00: Wobert - 86.832697 01: Eddy - 86.831768 02: Nim “Mr. 7280” - 86.806265 03: Slomac - 86.805790 04: NateyPooPoo - 86.796311 05: kkstrong - 86.793524 06: CeramicBlue - 86.782664 07: aight - 86.760159 ++: -- DEV TIME -- - 86.750000 08: auds - 86.740293 09: vertexv - 86.694284 10: Emmoji - 84.241785Scores: 11.

[gaunt ingot]

lemme just quickly commemorate the only time I'll have a 0th probably 🎉

[frigid summit]

@late wyvern scores for MAX-SI-D-00-03

[late wyvern]

Interpolated highscores for temu version of slomac's '25 handplaced e blocks' (MAX-SI-D-00-03) by NateyPooPoo:ansi 00: pralkarz - 84.925694 01: Sky! - 84.925623 02: Nim “Mr. 7280” - 84.913646 03: Slomac - 84.908145 04: CeramicBlue - 84.902197 05: NateyPooPoo - 84.895805 06: DarkStuff? - 84.887724 07: Kostucha - 84.876877 08: vertexv - 84.875897 09: aight - 84.862246 10: kkstrong - 84.853947 11: Eddy - 84.852754 ++: -- DEV TIME -- - 84.850000 12: auds - 84.765516 13: Wobert - 83.645872 14: Emmoji - 82.132306Scores: 15.

[frigid summit]

I don't believe there's much more to squeeze without another trick.

[frigid summit]

@late wyvern screenshot of MAX-SI-A-01-04

[late wyvern]

Screenshot for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo v1 in palette lava world:

[frigid summit]

What's the trick for this? I got it ||by jumping after gaining max. speed from running||, but how does it work?

[pastel beacon]

Something to do with friction on the ground after some point, iirc?

I remember excitedly reading convo about it while nsim was being developed, but I've not retained the knowledge

Like, you get decelerated to a capped velocity every frame or something, but jumping helps preserve the slightly higher than max cap? Idunno. I'm shooting shit without knowledge.
Will let the industry experts input on it

[fallow thorn]

It’s airtime, not jumping

here’s how it worked in N https://forum.droni.es/viewtopic.php?f=20&t=10336&p=177221#p177221

[frigid summit]

@late wyvern how many 0ths in MAX

[late wyvern]

pralkarz has 2 out of 150 highscore overall 0th scores for Max mappack.

[frigid summit]

Gaming.

[fallow thorn]

@late wyvern scores for MAX-SI-D-00-03

[late wyvern]

Interpolated highscores for temu version of slomac's '25 handplaced e blocks' (MAX-SI-D-00-03) by NateyPooPoo:ansi 00: pralkarz - 84.925694 = 01: Nim “Mr. 7280” - 84.925694 = 02: Sky! - 84.925623 03: vertexv - 84.925298 04: Slomac - 84.908145 05: CeramicBlue - 84.902197 06: NateyPooPoo - 84.895805 07: DarkStuff? - 84.887724 08: Kostucha - 84.876877 09: aight - 84.862246 10: kkstrong - 84.853947 11: Eddy - 84.852754 ++: -- DEV TIME -- - 84.850000 12: auds - 84.765516 13: Wobert - 83.645872 14: Emmoji - 82.132306Scores: 15.

[fallow thorn]

a perfect tie 👔

[frigid summit]

I wonder whether it's possible to coerce the in-game leaderboards to respect the interpolated time too.

[storm burrow]

@late wyvern anim for maxsie0200

[late wyvern]

Replay highscore trace for vertexv in MAX-SI-E-02-00 using palette vasquez:

[storm burrow]

thought this was pretty cool

[supple zenith]

That’s incredible lol

[split patio]

Haha nice! Was really not expecting that second half

[split patio]

I believe the game re-sorts ties by replay ID, which in the base game is a way of getting chronological order. Thus, I could spoof this by sending different replay IDs, but the problem is that this ID is used by the game to fetch the replay, so it needs to be correct. I could perhaps concoct some convoluted method to work around it, such as using the first 2 digits of the replay ID as the rank, and the rest as the actual ID

[north barn]

yes :)

[split patio]

Bingo

[frigid summit]

As long as outte works as expected, I'm happy. This working in-game would just be a nice bonus.

[split patio]

I think this explanation holds up very well for N++ as well. It's a consequence of the interaction between different factors: acceleration being different on ground vs air, ground friction not being present while in the air, etc. The max speed cap is actually the same, but your average full speed actually fluctuates and on average remains slightly higher while airborne, so one always wants to maximize airtime.

The reverse-engineered values for all the relevant constants can be found on Sim's tool.

[supple zenith]

There is another factor here as well. Sim could probably explain this much better than me, but I’ve observed it with the TAS many times. Sometimes when you land on the ground you get a little bit of extra friction. but say you held that jump for one more frame, then the friction might not apply. So to truly optimize you need to also make sure the jumps are timed to not get this extra friction.

I personally call it “the friction frame” but I don’t know exactly what the reason is. I’m assuming it has something to do with the way acceleration cycles

[split patio]

Interesting, I hadn't heard about this one

Once you get to the microsecond range things get very subtle to optimize, simply maximizing airtime will only suffice for the shabby millisecond range

Also, regarding the greatest level ever made, this also becomes relevant:

i.e. falling onto the middle of the door as opposed to simply running into it

[supple zenith]

Yep. So the perfect TAS would be jumping as soon as starting acceleration is maximized (and allows the next steps to align properly), then maximizing airtime while avoiding the friction frame (or whatever it is), then landing in the door properly

[split patio]

According to the attached calculation from the same N thread it saves almost 0.1 units (more specifically, about 1/265th of a tile), which in N++ would mean saving a frame in 1 out of 37 maps you try this on average (since max speed in 3.333 units per frame here), or 2.7% of the times:
The green player is touching the exit, the blue one does not. As explained, the effect comes from the difference in size - exit radius is 12px, ninja radius is 10px. The difference is extremely small - precisely, the green player is closer than the blue player by sqrt(22^2 + 2^2) - 22 = 0.0907 pixels. It is a very rare occurrence when it can make a difference, since max ground speed is 5 pixels per frame. If you consider the frame boundary alignment to be random across multiple maps, this is useful only in around 2% of all frame wars that include a flat ground, full speed finish.

[cedar stream]

Wdym live frame interpolation

[supple zenith]

The game runs at 60fps, so the timebar only advances at rate of each frame. But we can use the velocity and the XY position of the ninja to calculate the exact score, even if it exists between two frames.

Which it essentially always does

[analog plover]

I think that when you land on the ground, even if you jump immediately, ground acceleration is applied for this frame, so it can alter the acceleration pattern. But if you land on the ground on a frame where you would decelerate anyway, then there's no change.

[supple zenith]

Do both scenarios alternate each frame? Or is there more of a pattern/logic? My gut says the former, since I feel like the adjustment is never more than one frame away

[analog plover]

It would be the same pattern as the accelerate/decelerate in air, so usually every other frame yeah

[supple zenith]

Cool, that’s what I was thinking

[storm burrow]

@late wyvern how many max 0th

[late wyvern]

vertexv has 31 out of 150 highscore overall 0th scores for Max mappack.

[storm burrow]

wonder how many of these will hold up

[vestal acorn]

@late wyvern max rankings

[late wyvern]

Rankings - Highscore overall 0ths for Max mappack

-# Date: Jun 20, 2025 (16:33) UTC. MAX: 150.

00: Eddy           - 47
01: vertexv        - 31
02: Nim “Mr. 7280” - 21
03: Sky!           - 15
04: CeramicBlue    - 13
05: Slomac         - 13
06: vector7        -  4
07: NateyPooPoo    -  3
08: SimVYo         -  2
09: pralkarz       -  1```

[supple zenith]

Damn only 3 left from yesterday. Y’all been workin

[muted scroll]

Only 5 more scores needed to push DEV TIME off the first level

That nub

[supple zenith]

Hmm. Might have to figure something out of it does fall off. At least at that point if you get top20 you know you beat it

Maybe scores around mine would show it if you haven’t. Dont remember if that board works on Eddy’s lbs or not

[muted scroll]

Full rankings in outte or spreadsheet

[supple zenith]

Could add the times to the author sheet

Less convenient of course

[split patio]

Haven't implemented scores around me but it shouldn't be too hard, but also yeah you can ask outte for the full leaderboard

[supple zenith]

@late wyvern anim 0 1 for maxsie21

[late wyvern]

Replay highscore traces for NateyPooPoo and vertexv in MAX-SI-E-02-01 using palette galactic:

[supple zenith]

Not even 24 hours and some of these boards are already crazy tight. Great score vertex

[frigid summit]

My 1 is so sad. 😭

[oblique bear]

still better than my 0

[fickle ether]

Looks like all my 0ths were eaten

A shame

[muted scroll]

This pack will be harder to keep 0ths than any before

Everyone has something they are good at

[zenith owl]

wondering who between bobio/natey/I will hit all the DIs in that one level first

[late wyvern]

Replay highscore traces for vector7 and Eddy in MAX-SI-A-01-00 using palette vasquez:

[zenith owl]

yeah less than 5 mins ago lol

[frigid summit]

My specialty is low input levels, clearly.

[old lagoon]

I have downloaded MAX now

[supple zenith]

Not even framewar tab is safe from reroute snipes 👌

[fickle ether]

As well it shouldn't be

[zenith owl]

well what better way to save frames

i've got a little bit before work if people wanna watch me highscore in real time

[muted scroll]

@late wyvern spoiler anim 0 1 2 for maxsid23

[late wyvern]

Replay highscore traces for kkstrong, Slomac, and Eddy in MAX-SI-D-02-03 using palette vasquez:

[muted scroll]

I wonder which route is actually best

[supple zenith]

I really want to believe in the full go around, because it’s seems the least painful

@late wyvern anim 0 1 for maxsid12

[late wyvern]

Replay highscore traces for NateyPooPoo and Nim “Mr. 7280” in MAX-SI-D-01-02 using palette galactic:

[supple zenith]

Struggling to even match Nim’s score then out of nowhere four frames up

[frigid summit]

I really, really like this tab so far, the concept is amazing. Even if I'm nowhere near good enough to compete for 0ths, it's still fun to squeeze as much as possible from certain levels.

[muted scroll]

I appreciate Natey competing on his own maps, usually he refrains. But the playing field is already so level with how easy each map is

[supple zenith]

I spiritually cannot sit out from framewar tab

[fickle ether]

Alright, coming back, dev times in column-01 lesgo

[analog plover]

Damn the competition is already looking quite fierce on this tab
Getting the top spot on the rankings after a few days is gonna be quite an acomplishment, compared to other tabs

[supple zenith]

I’m still picking up 0ths relatively quickly, but I’m losing them just as fast lol

[storm burrow]

@late wyvern how many max 0th

[late wyvern]

vertexv has 11 out of 150 highscore overall 0th scores for Max mappack.

[storm burrow]

not many

[supple zenith]

@late wyvern max rank

[late wyvern]

Rankings - Highscore overall 0ths for Max mappack

-# Date: Jun 20, 2025 (19:32) UTC. MAX: 150.

00: Eddy           - 43
01: Nim “Mr. 7280” - 37
02: Sky!           - 22
03: NateyPooPoo    - 12
04: CeramicBlue    - 10
05: vertexv        - 10
06: Slomac         -  7
07: vector7        -  5
08: SimVYo         -  2
09: jp27ace        -  1
10: kkstrong       -  1```

[supple zenith]

Nim pushing hard

[muted scroll]

snake zula screams sky

checks author list

Yuuuuup

[zenith owl]

tbf snake zula is where all the hell tech is concentrated to

[muted scroll]

It just embodies "This is fun." -sky, probably

[zenith owl]

imo it wouldn't have been the framewar tab without the basics 2 somewhere in there

the warrest of frames

[viscid fox]

ok this begs a question

[muted scroll]

Yeah, what only you would think, exactly what I am saying

[viscid fox]

is there any DI in this tab that is mandatory for 0th

[frigid summit]

For 0th – quite likely that there's going to be some DIs found.

[zenith owl]

tmk only for snake zula so far

a lot of levels have DI spots patched out

[frigid summit]

For dev times, they're set so that DIs are not required.

[supple zenith]

It’s nearly impossible to make a tab where this is not the case unfortunately. But I did try pretty hard to minimize the potential for them

So yeah there will be a few I’d imagine

[frigid summit]

I recently got an idea for a tab, but it's unlikely I'll ever make it. Willing to sell it for $10.

[fickle ether]

Alright!

All dev times on -01 acquired. Did grind a few levels, just a bit, none any more than 20 minutes, and yea, didn't score any 0ths

People are already quite on top of things

[supple zenith]

10 gold is the best I can do

[muted scroll]

Call. Raise -2 gold.

[supple zenith]

Someone’s got deep pockets eh?

[old lagoon]

I think shallow pockets

Raising negative gold means he's only offering 8

[split patio]

Found a cool hacky way to fix the improper sorting of the millisecond-ties in the game, the order should now match the real one shown by outte:

@late wyvern max scores for greatest level

[late wyvern]

Interpolated highscores for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo:ansi 00: CeramicBlue - 84.615157 01: Eddy - 84.615077 02: Slomac - 84.615052 03: pralkarz - 84.615035 04: Nim “Mr. 7280” - 84.614836 = 05: vector7 - 84.614836 = 06: kkstrong - 84.614164 07: Sky! - 84.614050 08: vertexv - 84.614047 09: NateyPooPoo - 84.611961 10: jp27ace - 84.611483 11: DarkStuff? - 84.611079 12: aight - 84.610852 ++: -- DEV TIME -- - 84.600000 13: auds - 84.598571 = 14: Emmoji - 84.598571 = 15: Wobert - 84.496868Scores: 16.

[split patio]

Which means now you'll see yourself rising up the ranks if you improve the microseconds even if it's still tied to the millisecond

[frigid summit]

Oh, so the sorting is happening server-side, not client-side?

Since no DLL patching is neeeded?

[zenith owl]

@late wyvern anim for max-si-b-03-01

[late wyvern]

Replay highscore trace for Sky! in MAX-SI-B-03-01 using palette wyvern:

[zenith owl]

got the 3 di run

[supple zenith]

Heck yeah Eddy. That’s awesome

Silly run sky

[fickle ether]

@late wyvern scores for si-a-01-01

[split patio]

The game re-sorts the ties by replay ID, so it's client side. But I'm crafting some custom replay IDs that will be sorted properly in-game but still resolve to the correct runs when queried.

[fickle ether]

Oh outte hasn't been informed

@late wyvern scores for max-si-a-01-01

[late wyvern]

Interpolated highscores for robert is your uncle (MAX-SI-A-01-01) by NateyPooPoo:ansi 00: DarkStuff? - 77.926127 01: Eddy - 77.907532 02: vector7 - 77.818249 03: NateyPooPoo - 77.791234 04: vertexv - 77.757015 05: Nim “Mr. 7280” - 77.585531 06: Slomac - 77.421557 07: CeramicBlue - 77.403904 08: kkstrong - 77.199644 09: aight - 77.113383 ++: -- DEV TIME -- - 77.000000 10: pralkarz - 76.568588 11: jp27ace - 75.924540 12: Sky! - 75.777385 13: auds - 73.938612 14: Emmoji - 73.563456 15: Wobert - 71.634615Scores: 16.

[fickle ether]

Got one :)

I innovated the beginning, though I didn't know I had because I only just now watched Eddy's (or anyone's) replay

Animation seems to agree that it's faster. Eddy meets me again after the turnaround on the right, then I pull ahead with a better ending.

@late wyvern screenshot max-si-a-01-01

[late wyvern]

Screenshot for robert is your uncle (MAX-SI-A-01-01) by NateyPooPoo v1 in palette plus:

[fickle ether]

Realize I didn't do "screenscores" my b

[fickle ether]

but you need a better turnaround, tsk tsk tsk

[zenith owl]

look man, the 3 dis took me like 900 atts to hit a single time

[supple zenith]

Screenscores are so 2022-2024. Now we do anims

Me clicking this screenshot and wondering why nothing is moving

[split patio]

I did the same lmao

[supple zenith]

Ooh I just sub-milliseconded a 0th. Already putting this new implementation to work Eddy

[split patio]

Sweet

[fallow thorn]

@late wyvern max rankings

[late wyvern]

Rankings - Highscore overall 0ths for Max mappack

-# Date: Jun 20, 2025 (20:06) UTC. MAX: 150.

00: Nim “Mr. 7280” - 40
01: Eddy           - 37
02: Sky!           - 23
03: NateyPooPoo    - 16
04: CeramicBlue    -  9
05: vertexv        -  8
06: Slomac         -  7
07: vector7        -  4
08: jp27ace        -  2
09: SimVYo         -  2
10: DarkStuff?     -  1
11: kkstrong       -  1```

[fallow thorn]

finally

my arms hurt so bad

that was my goal for today tho

really fun pack

[split patio]

Glad this mappack has incentivized multiple comebacks

I see jp in there as well

[fickle ether]

@late wyvern anim max-si-b-01-02

[split patio]

Let's see if nw pops by now that he's actually played N++ again, this pack is right up his alley

[late wyvern]

Replay highscore trace for DarkStuff? in MAX-SI-B-01-02 using palette plus:

[fickle ether]

Got another one :P

[zenith owl]

.720

[supple zenith]

That would be crazy tbh

[fickle ether]

Wow that happened fast

.729!

[zenith owl]

734

[fickle ether]

.768 O_O

[old lagoon]

These numbers just don't feel right to see

[zenith owl]

.809

[fickle ether]

.829!!

[supple zenith]

I’m actually starting to adjust quite nicely to them. Same goes for the microseconds remaining variable on submission

All seemed gibberish at first but it’s quickly feeling more normal

[zenith owl]

.871

holy shit

[fickle ether]

Whaaaaat

[frigid summit]

Did Nightwish ever interact with this server?

[zenith owl]

all I have time for for now, gotta head to work

[fickle ether]

Wel you've left me a good deal of work

Lemme see if I can beat this

[split patio]

Oh yeah, he was very active in the community during the years he was also active highscoring, except at the start iirc.

Unfortunately, since he left the server, you can no longer track his posts (another shitty functionality of Discord), but they'd be in the thousands.

[fickle ether]

Eh honestly I don't wanna do that right now

May come back

[zenith owl]

Lmao

[fickle ether]

Good fight tho

Good back and forth

[zenith owl]

Ye, these kinds of battles are so much more interesting since every minor improvement matters

[frigid summit]

Ooh, I wasn't aware, I always thought he was this silent goat who beat everyone, but never said a thing.

[olive nova]

you can search for messages from people who left the server with the user id or username, just won't show up in the prompt when typing from:

[split patio]

It just searches the term for me

[olive nova]

are you sure that's their actual discord username

[split patio]

Oh, you're right, it was just the display name

It's been 5 years (mid-2020) since nw stopped playing

https://tenor.com/b0CbE.gif

[frigid summit]

What's wild to me is that I used to play the flash version of N in elementary school, and now, 20 years later, there's not only this amazing successor that's probably my favorite platformer ever, but a brilliant community that reverse-engineered the physics engine, created a bot that does pretty much everything you could think about related to N++, and keeps producing content that's fun and fresh despite the franchise being relatively old.

[fallow thorn]

not to mention I've been playing 1.4 on my iPad, and for a 20 year old game it still totally holds up

just as fun as when I first played it in the 2000s

[old lagoon]

Not to mention you still have people in the community now that were in the community then

Not many, but a few 👋

[frigid summit]

I wish I was in the community back then, but I only learned English as a teenager.

[old lagoon]

and how old are you now? xD

[frigid summit]

I'm 25.

[old lagoon]

ok yeah you were not that old 20 years ago haha

[winged anvil]

Played through a handful of levels, enjoying it so far. Will probably aim for all dev times

[supple zenith]

@late wyvern Scores for maxsia30

[late wyvern]

Interpolated highscores for crank (MAX-SI-A-03-00) by Darkstuff:ansi 00: Nim “Mr. 7280” - 85.559941 = 01: NateyPooPoo - 85.559941 = 02: Eddy - 85.559935 03: Sky! - 85.555952 04: Slomac - 85.548123 05: CeramicBlue - 85.542346 ++: -- DEV TIME -- - 85.533333 06: vertexv - 85.515249 07: kkstrong - 85.503463 08: Emmoji - 85.324114 09: Wobert - 85.216846 10: auds - 84.988583Scores: 11.

[supple zenith]

Tried my damnedest

I’m reluctant to use the word max here, but we might have one

[split patio]

Tfw when you're 6 microseconds away 😔

[supple zenith]

lol

I got plenty of both your score and the max

[old lagoon]

How do you know down to the microsecond what you get?

[supple zenith]

It’s the “global rank” number after submission

[old lagoon]

ah right I do remember that being said

[supple zenith]

So I was getting “6725” which is the max. And “6731” which is Eddy’s score

Oops I said the word

maybe the max

[old lagoon]

He said the title!

[analog plover]

https://tenor.com/view/breaking-bad-chemistry-experiment-scientific-walter-white-gif-17820266

Me trying to mathematically solve the greatest level ever made

[supple zenith]

Was wondering when you’d join the convo

[split patio]

Next millisecond in the works 👀

[analog plover]

Didn't quite start yet but I set up the Excel spreadsheet we bout to crunch some numbers

[north barn]

Jesus christ sky

[late wyvern]

Sorry, I didn't understand your command.

[north barn]

It's fascinating how much modular arithmetic is baked into the high level of this game. Completely humanly inaccessible due to being centered around 1fs, or so one may believe

-# I mean, humans have hit several hundred in a row

[analog plover]

If there are many tied true exact maxes, is it still ranked based on who did it first?

[olive nova]

Most sane timing window based difficulty comparison

[split patio]

Yeah, true ties fall back to chronological order again

And speaking of true ties, I've adapted the spreads function to work with frac scores:

@late wyvern smallest max spread

[late wyvern]

All levels with the smallest interpolated highscore spread between 0th and 1st for Max mappack:

MAX-SI-A-03-00 - 0.000000 - Nim “Mr. 7280”
MAX-SI-A-02-02 - 0.000018 - NateyPooPoo
MAX-SI-A-01-04 - 0.000080 - CeramicBlue
MAX-SI-D-03-02 - 0.000102 - Slomac
MAX-SI-C-00-01 - 0.000464 - Nim “Mr. 7280”
MAX-SI-D-00-03 - 0.000669 - Slomac
MAX-SI-B-00-03 - 0.001098 - Sky!
MAX-SI-B-03-00 - 0.001931 - Nim “Mr. 7280”
MAX-SI-D-00-02 - 0.002375 - Sky!
MAX-SI-A-01-01 - 0.002501 - Eddy
MAX-SI-C-02-03 - 0.002840 - vertexv
MAX-SI-C-03-04 - 0.003095 - NateyPooPoo
MAX-SI-B-02-01 - 0.003435 - Eddy
MAX-SI-B-00-01 - 0.003474 - CeramicBlue
MAX-SI-C-01-00 - 0.003699 - Eddy
MAX-SI-A-00-00 - 0.004118 - CeramicBlue
MAX-SI-D-00-01 - 0.004542 - Sky!
MAX-SI-E-00-02 - 0.004834 - Sky!
MAX-SI-E-02-01 - 0.005215 - NateyPooPoo
MAX-SI-D-00-00 - 0.006935 - Sky!```

[split patio]

@late wyvern max maxable

[late wyvern]

Level highscores with the most ties for 0th for Max mappack on Friday June 20 at 23:33:23 (+0000):

MAX-SI-A-03-00 - 3 - Nim “Mr. 7280” ```

[split patio]

First maxable level unlocked

[frigid summit]

@late wyvern screenshot of MAX-SI-A-03-00

[late wyvern]

Screenshot for crank (MAX-SI-A-03-00) by Darkstuff v1 in palette lava world:

[fallow thorn]

if I did somehow literally max it, that would be Extremely Neat, I was just tryna get the "zeroth" (0th)

[tranquil fiber]

max maxable is such a badass command

[fallow thorn]

the actual one that took a while was the temu one

[north barn]

it may perhaps be confusing that "max" refers to a mappack

[fallow thorn]

@late wyvern anim trace slow 0 1 2 3 MAX-SI-D-00-03

[late wyvern]

Replay highscore traces for Slomac, Sky!, CeramicBlue, and pralkarz in MAX-SI-D-00-03 using palette vasquez:

[fallow thorn]

@late wyvern scores for MAX-SI-D-00-03

[late wyvern]

Interpolated highscores for temu version of slomac's '25 handplaced e blocks' (MAX-SI-D-00-03) by NateyPooPoo:ansi 00: Slomac - 84.935784 01: Sky! - 84.935115 = 02: CeramicBlue - 84.935115 = 03: pralkarz - 84.925694 = 04: Nim “Mr. 7280” - 84.925694 = 05: vertexv - 84.925298 06: NateyPooPoo - 84.895805 07: DarkStuff? - 84.887724 08: vector7 - 84.880217 09: Kostucha - 84.876877 10: aight - 84.862246 11: kkstrong - 84.853947 12: Eddy - 84.852754 ++: -- DEV TIME -- - 84.850000 13: jp27ace - 84.834038 14: auds - 84.765516 15: Wobert - 83.645872 16: Emmoji - 82.132306Scores: 17.

[fallow thorn]

yeah now that you mention it, that could become a problem

check out my new travel planning app, it's called Explorer.exe

[analog plover]

numbers: crunched

[split patio]

I just saw it live as I played the level myself, lol

[fallow thorn]

@late wyvern max scores for greatest level

[late wyvern]

Interpolated highscores for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo:ansi 00: SimVYo - 84.615162 01: CeramicBlue - 84.615157 = 02: Eddy - 84.615157 = 03: Slomac - 84.615052 04: pralkarz - 84.615035 05: Nim “Mr. 7280” - 84.614836 = 06: vector7 - 84.614836 = 07: jp27ace - 84.614528 08: kkstrong - 84.614164 09: Sky! - 84.614050 10: vertexv - 84.614047 11: NateyPooPoo - 84.611961 12: DarkStuff? - 84.611079 13: aight - 84.610852 ++: -- DEV TIME -- - 84.600000 14: auds - 84.598571 = 15: Emmoji - 84.598571 = 16: Wobert - 84.496868Scores: 17.

[fallow thorn]

Broe Moment

[analog plover]

Lol according to my first upper bound on the max I new there was at most 5 millionths to clean and it was indeed possible to clean all of them

[split patio]

Lol, how did you even prove an upper bound here

[north barn]

-# And does that mean it's truly maxed?

[analog plover]

[fallow thorn]

exhaustion

[analog plover]

Should be true max yeah

[analog plover]

The key is to manipulate the last jump so your height at the end (taking into acount interpolation) is as close as possible to the height of the door. Upper bound is assuming your height is exactly the same as the door

At the end I touch the door being 0.01291322 units off-centered, which is the smallest possible deviation across all possible jumps, also assuming the perfect start x velocity wise.

[split patio]

https://tenor.com/bXqIM.gif

[fallow thorn]

it’s simple, you simply compute 6 (input states per frame) to the power of t states per frame (t) until one of them touches the exit

[split patio]

Thanks for spilling the beans, now everyone knows the strat

[fallow thorn]

Now that we have the physics engine solved this is small potatoes

[split patio]

What other factors come into play? I assume for sure jumping at the exact frame max speed is reached, and maximizing air time, right?

[analog plover]

The setup is actually easy. The first jump has frame perfect timing, but then you do a smallest jump then a biggest jump and voilà

The rest is almost disapointingly easy. You run on the ground until the very frame you reach max speed, then it's a matter of always being in the air and avoiding the unlucky frames where you can get a speed loss upon touching the ground.

[split patio]

Natey mentioned those friction frames, what's all that about?

[analog plover]

I'm pretty much 100% sure that once you reach max speed, there are no better alternative than always being in the air. There aren't any way to manipulate positively your speed by running on the ground a little bit or adding neutral frames or whatever

[analog plover]

You have a roughly 50% chance of getting a small slow down when you touch the ground. This is not due to friction but the difference between ground accel and air accel. Basically you want to touch the ground on one of the frame where you decelerate (remember that the pattern is decelerate, accelerate, decelerate, accelerate...). Otherwise you get a penalty.

[split patio]

Huh, interesting, I didn't realize that asymmetry at first so all frames looked the same and thus I couldn't possibly fathom where that strange effect was coming from

But that makes sense

[north barn]

I remember Eddy or someone else posting an image showing the ninja's velocity oscillating over time, and since it's not a perfect modulo there are occasional frames where it decelerates (was it accelerates?) twice in a row I think?

There's probably a way to calculate the ratio

[analog plover]

So the difference between my run, bobio's and eddy's all come from the height at the end. It's pretty easy to maxout the speed otherwise

[north barn]

Basically you're aiming for the ninja's velocity to be in a certain modulo range when touching the ground

[analog plover]

I'm not sure what modulo has to do with this

[north barn]

The ninja's speed is "wrapping" in a sense around the maximum speed, and the ninja's air velocity being higher is because it simply hits higher values on average before being wrapped, right

[analog plover]

Yes

[fallow thorn]

I have been jp'd! great to see you again my friend

[north barn]

So you more or less are aiming for that velocity to be in a state where switching to the "ground" wrapping gives the case where you end up with slightly more total velocity

(If this is correct, it might be possible that for specific running lengths, the difference doesn't matter at all...?)

[split patio]

The image I posted was probably this one from the N days, pretty sure Sim generated a couple ones for N++ as well:

[north barn]

I remember the one for N++ looking like it was very aggressively shifting back and forth (like, almost every other frame, hence the 50% figure)

[north barn]

Bingo

It was double decel I'm blind

[split patio]

These are both when running on the ground, at different scales

[analog plover]

ground is double decel, air is single decel

[north barn]

I'm moreso referring to how it handles when the total value accumulates to a certain threshold, normally it alternates accel-decel-decel on the ground but it decelerates three times since within the threshold it is gradually increasing

Nice to know that it seems to be backwards in the air case, though?

Unless I misunderstood

[north barn]

So, the connection from this that I was making is that this is actually just a kind of modulo arithmetic, and so we are attempting to control a chaotic function, in a sense xP

[analog plover]

The most important takeaway I think is that in the ground case, the speed fluctuates between 3.266666 and 3.333333, while in the air it fluctuates between 3.288888 and 3.333333. So you can (sort of) say that the average max speed on the ground is 3.300000 and the average max speed in the air is 3.311111

Also both functions indeed have a slow increase, with a shrap decrease every few seconds. But there is (un)fortunately no way to manipulate that in order to gain time

[north barn]

But that's exactly what we've been doing!

All those little "unmaxes" where a random frame of no motion were put in somewhere, that was controlling this exact behavior

[analog plover]

No, as I said previously there's no way (at max speed) to sneak in neutral frames or other shenanigans in order to gain time

Sneaky neutral frames are indeed op, just not at max speed

[north barn]

Wait, so what happened about the zenny 0th or another one that I'm forgetting the name of right now

[split patio]

I think zenny was running a little bit before the first jump

[zenith owl]

That intro unmax would be the relevant level for this one

[north barn]

I swear there have been multiple incidents of max speed unmaxes

[analog plover]

It's not an hold right level. You have to release at the start to not hit the drone

[split patio]

@late wyvern max maxable

[late wyvern]

Level highscores with the most ties for 0th for Max mappack on Saturday June 21 at 01:01:40 (+0000):

MAX-SI-A-03-00 - 4 - Nim “Mr. 7280” 
MAX-SI-A-01-04 - 3 - SimVYo         ```

[north barn]

Sorry I had to think about this for a second - is touching the ground not a sort of equivalent in nature to intentionally placing a neutral frame optimally?

[analog plover]

Yeah sort of. Or at least it's an other way to attempt manipulation. But again I'm pretty convinced you can't manipulate speed in your favor by running on the ground at key moments

It's hard to fully rigorously prove it's impossible to manipulate speed that way... but I'm pretty sure

like 99% sure

[zenith owl]

I mean, don't you accelerate faster on the ground?

Surely you could run for a single frame at a time

[analog plover]

Yes, but since the speed in the air never goes below 3.288888, in practice it's never possible to actually accelerate the frame you touch the ground. That's the real reason why you can't manipulate it that way. In order to accelerate on the ground, your speed has to dip below 3.266666 at some point, sorry it's hard to explain.

[north barn]

There is something in my intuition that tells me that this should be possible for specific lengths

I don't think a ground/neutral frame does that kind of bumping, but the idea is to cut a frame where the ninja would accelerate above the threshold and instead move it just a little closer to gain a larger sum of velocities

I can trust your experience in neutral frames/ground touches not cutting it for this, perhaps, but there must be some principle where this can happen

[analog plover]

I have thought about that eki, if I think I understand what you mean. Sadly (fortunately) doesn't seem possible

[north barn]

Regardless, I think this would sufficiently explain the "bad landing frames", at least

So if it can't speed you up... at least it must be possible to manipulate it to slow you down less

[north barn]

I'd like to see the cycle length/approximate ratio for air movement, because that might give an insight into how often the "bad frames" switch between "parities" (.x00, .x33, .x66 vs .x17, .x50, .x83)

If it's a game of modulo arithmetic, maybe there's a way to make more of it obvious to the player

[analog plover]

The cycle length in the air is quite long, roughly 240 frames

On the ground I don't remember but eyeballing the chart it's roughly 120 frames

[north barn]

Huh, that's like, 3/4 of the map, right?

[analog plover]

Yeah that's most of it

[north barn]

So probably beyond the first frame perfect jump, the consistent jump heights (full jump and 1f buffer) being a specific length is useful in keeping parity

Wow, that's a remarkable connection

(If only it weren't important to touch the ground less...)

(...but then none of this would make sense obviously :p)

[split patio]

Bruh, take a gander at this, lol

I think those are both the maxes

[analog plover]

skill issue 😔

[split patio]

Should've started grinding immediately rather than reading up the convo

[north barn]

the grindset

I think this convo was worthwhile though because it gives insight on how optimizing future stretches looks like

[split patio]

For sure

[tranquil fiber]

this tab has made for terrifying

[north barn]

If running in the air for a long time has a consistent parity on when it's best to land for 3/4 of the level, and you have consistent heights of jumps that will land you with a matching or opposite parity, then we know how to lab out a level ahead of time

[analog plover]

Yeah air optimization is a lot easier than I initially thought

[split patio]

For completion, here's the picture for nv2, you can see the figures match the N++ ones, which means the physics constants do (and so does the framerate).

[north barn]

adding to my notepad tasklist to note the lengths of common jump setups in order to track their parity values

[analog plover]

I think the smallest jump and the biggest jump have different "parity". And then there are other setups if you can hit the ceiling, as in the basics. So it's pretty hard to prevent any setup

Also it's pretty easy to just get lucky and not get the slow down, it's just a 50% chance after all

[north barn]

more like 120/241 or whatever

(50% will do.)

[split patio]

False alarm, bobio had done it 5 mins after Sim got the max, lol

But he's still grinding the level, guess he doesn't trust the simulator author

[analog plover]

Lol it also took me like 5 minutes, very easy max once you know the setup

[fallow thorn]

i cant tell what im asposta do, so medium jump and then small and large

[analog plover]

smallest and largest

so there's just the first jump to worry about

[fallow thorn]

yeah you just said do the smallest and then largest which is why i'm confused

oh

so variable -> frame -> hold

[analog plover]

yep

[fallow thorn]

gotcha

[north barn]

Have to minimize ground contact does unfortunately limit the choice of setups, but at least it accurately predicts what will happen

[split patio]

It's essentially only 2 timings that matter, so it's indeed not bad, specially since you have visual cues

[analog plover]

So there's only two frame perfect parts, when to initiate the first jump and how long to hold it, the rest is no brainer

[split patio]

Didn't even need my trusty monitor marks this time

[analog plover]

And yeah there are other ways to attain the max, as long as the last jump is a max jump and that it starts at the same time. But this setup that I use is surely easiest

It's really convenient that the perfect jump at the end is the max one lol

[zenith owl]

Exactly the kinds of shenanigans this tab was meant for

[analog plover]

the 10 frame jump also gives you an alignment almost as good. Maybe it's like a millionth slower lol

this truly is the gretest level of all time

[split patio]

Interestingly the demo analysis shows the replays are not identical, excluding obvious buffer artifacts

Notably my first jump is one frame earlier (but ends on the same frame, so it's 1 frame longer)

[analog plover]

Oh yeah that's true thare is one frame of leeway on the first jump I forgot. Cause after reaching max speed you decelerate the next frame, regardless of if you are in the air or on the ground

(And the deceleration is the same in the air and on the ground, it's just your speed multiplied by drag)

[north barn]

Wait, is it the same drag?

[split patio]

Oh interesting, so it's a consistent behaviour

[zenith owl]

Pivotal information lmfao

[analog plover]

always has been

[north barn]

Interesting, so the drag only changes when you're slow

[analog plover]

lol I thought you knew, sorry for not pointing it out earlier. A lot of what I said doesn't make sense if you don't have that in mind woops

[north barn]

Ah, but the accel is different, so that's where the higher average comes from

[analog plover]

Now you get it

[north barn]

This still makes sense at least, because it assumes the same threshold on ground/air (which I knew) and I suppose that just means that the decrementation after a velocity above the threshold is always the same

It is always better to have more velocities above the threshold, but the hypothetical here is to have exactly one frame dedicated to raising the position within the oscillation without bringing it over the threshold, and then having an even/odd (depending on the framing) number of total running frames to technically have a greater sum of velocities in that duration

[fallow thorn]

@late wyvern scores for MAX-SI-A-01-04

[late wyvern]

Interpolated highscores for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo:ansi 00: SimVYo - 84.615162 = 01: CeramicBlue - 84.615162 = 02: Eddy - 84.615162 = 03: Nim “Mr. 7280” - 84.615162 = 04: Slomac - 84.615052 05: pralkarz - 84.615035 06: vector7 - 84.614836 07: jp27ace - 84.614528 08: kkstrong - 84.614164 09: Sky! - 84.614050 10: vertexv - 84.614047 11: NateyPooPoo - 84.611961 12: DarkStuff? - 84.611079 13: aight - 84.610852 ++: -- DEV TIME -- - 84.600000 14: auds - 84.598571 = 15: Emmoji - 84.598571 = 16: Wobert - 84.496868Scores: 17.

[north barn]

Buuut I suppose that sudden small accel would need to come from a very strange source indeed.

[analog plover]

I suggest you look at the equations and play with the numbers yourself at this point lol

[north barn]

Perhaps I will :)

As for the relevance to maxing though, there's one other factor in the quick parity calculation, and it's that landing should shift the cycle, actually

The ground has a much shorter cycle, so landing for one frame shifts the time forward by the equivalent of a few frames... I think.

[split patio]

Natey missing all this maxing action

[analog plover]

If you land with bad parity, you decelerate for one frame more, that's why it's bad. I suppose it does alter the cycle, but then there's nothing to do to recover the impact of the deceleration

[fallow thorn]

maxing to 6 decimal places feels so satisfying

[split patio]

It's maxed to full double-precision floating point in fact

[analog plover]

Ok I'll stop rambling. Maybe I'll play some other levels too

[split patio]

Or tied, at least. Maxed only if you believe Sim

[analog plover]

||I've been attempting my secret max speed manipulation this whole time||

[split patio]

https://tenor.com/bgYyH.gif

[north barn]

I meant landing with good parity also shifts the cycle, but I assume only to shorten the total time spent on one parity by a small amount. I still don't know the exact distance equivalent of the cycle, so it doesn't matter too much, I think...

The main goal is to find a player-visible shorthand for hitting the correct parity :)

[analog plover]

No it doesn't

Ok I really wanna stop talking about it lol

[north barn]

Oh, because you're trying to decel on the ground!!!

Sorry I just got it

Yeah that makes perfect sense, you make sure the frame(s) where you accel is not on the floor, but instead surrounds it

And that's why all this matters, determining the decel frames and intentionally manipulating jumps so you land on them. Wait, that means there's more frames where it's good to land, not less :)

So it's more like 121/241...

ok, I'll let you guys move on from this lol

(At least, until I have a table of jump parities ;p)

[analog plover]

It took me a while to wrap my head around this function. It's an interesting and unintuitive math problem

[north barn]

ok LAST LAST thing, the 2f window where you can jump optimally is super interesting because it means you can actually do a jump of either parity depending on which of the frames you jump on

If you jump on the second frame (decel frame), you want an even parity (landing on the decel frame), and vice versa for the first one

So perhaps there's more choice than is first obvious :)

Oh, and this means you don't actually need to maximize airtime!!! Just land correctly every time :)

(Note: this is sarcasm, because I am asking to hit more 1fs.)

First case to test is 1-tall hallways, because maybe holding max jumps vs spamming minjumps has different parities so one counter-intuitively could be faster than the other

OK I'LL SHUT UP THIS IS SUPER INTERESTING

[north barn]

Corrected, I'm blind

[storm burrow]

@late wyvern how many max 0th

[late wyvern]

vertexv has 6 out of 150 highscore overall 0th scores for Max mappack.

[north barn]

max%: tie all interpolated 0ths, obviously

(I wonder if we can figure out which level is the most difficult to achieve this on)

Okay, enough talk, I'm logging on

[supple zenith]

When I made this level I knew two absolute and inarguable truths.

1. You would TAS this level.
2. Nobody would get the true max before you had the chance to

[analog plover]

All I knew is that I would have some fun time with this one. It did not disappoint

[zenith owl]

@late wyvern scores for max-si-a-01-04

[late wyvern]

Interpolated highscores for the greatest level ever made (MAX-SI-A-01-04) by NateyPooPoo:ansi 00: SimVYo - 84.615162 = 01: CeramicBlue - 84.615162 = 02: Eddy - 84.615162 = 03: Nim “Mr. 7280” - 84.615162 = 04: Sky! - 84.615162 = 05: Slomac - 84.615052 06: pralkarz - 84.615035 07: vector7 - 84.614836 08: jp27ace - 84.614528 09: kkstrong - 84.614164 10: vertexv - 84.614047 11: NateyPooPoo - 84.611961 12: DarkStuff? - 84.611079 13: aight - 84.610852 ++: -- DEV TIME -- - 84.600000 14: auds - 84.598571 = 15: Emmoji - 84.598571 = 16: Wobert - 84.496868 17: ekisacik - 84.494514Scores: 18.

[zenith owl]

ez

[supple zenith]

I have made it home for the evening and can now read through 500 messages about my favorite subject.. Analysis of how ninjas move in a super niche video game at the level of microseconds.

And yeah I should probably go ahead and snag this score while I’m at it

[north barn]

@late wyvern anim 10 for max-si-c-03-04

[late wyvern]

Failed to render animation

[north barn]

What

@late wyvern anim 10 for max-si-c-03-04

[late wyvern]

Failed to render animation

[north barn]

Oh okay so it consistently does that

Huh

thank you natey

[supple zenith]

I think we had a similar convo long ago. And yeah.. thank god that acceleration cycle manipulation isn’t a highscoring strat. Sounds truly awful

[supple zenith]

I hate accidental oneway clips 👍

[spark quiver]

done with all dev times. really cool idea for a tab and even though im not really the target audience, had a great time with it nonetheless

[supple zenith]

Heck yeah. GGs

[supple zenith]

Absolutely love when things like this happen in normal play, and comparatively there’s a lot more room there to manipulate for these variables. Crazy awesome that it’s usable for this precision max.

Got the score. Not bad at all when you know what to do

[tranquil fiber]

i love how we basically have two TAS channels now

[supple zenith]

I see you picked up 4 0ths as well. Nothing to scoff at there with how tight everything already is

[zenith owl]

honestly this tab would go crazy if @jaunty marsh was here

[supple zenith]

Oh you mean that dude who showed up out of nowhere 100%ed the game turned out to be a super cracked highscorer steamrolled his way to third place on the rankings then went to go play monster hunters one day and vanished entirely?

Yeah he should pop by, would be nice

[north barn]

Finished the pack! I think I will finally concede that highscoring N++ is a pretty strictly 2-handed game. I could get to 100% and pull off a lot just fine one-handed, but god damn some of these inputs suck on one hand

[supple zenith]

Honestly 100% with one hand is impressive to me, and kind of crazy. What the heck are you doing with your other hand?

[north barn]

Drinking water : )

(And in co-op, playing the second player!)

[supple zenith]

Pro strats tbh

[tranquil fiber]

i reserve my left hand for suicide

[north barn]

I just checked, and yeah, it's failing on my replay specifically... it's failing on replays below a certain rank, hmmm

Oh shit, I figured out the bug. Outte fails to render replays slower than the dev time LOL

Optimization or just a bug in it not knowing what to refer to?

@late wyvern scores for max-si-e-04-03

[late wyvern]

Interpolated highscores for my little meat tornado (MAX-SI-E-04-03) by Slomac:ansi 00: Eddy - 82.701591 01: ekisacik - 82.701516 02: Slomac - 82.430786 03: CeramicBlue - 82.296601 04: auds - 82.116715 05: aight - 82.023918 ++: -- DEV TIME -- - 82.000000 06: NateyPooPoo - 81.741983 07: vertexv - 81.124640 08: Wobert - 78.136756 09: Emmoji - 73.739607Scores: 10.

[north barn]

What the fuck

why is it that close

I'm guessing identical setups but I let go a frame early or something?

actually i wanna see this

@late wyvern anim 0 1 for max-si-e-04-03

[late wyvern]

Replay highscore traces for Eddy and ekisacik in MAX-SI-E-04-03 using palette quench:

[zenith owl]

probably door alignment shenanigans

[north barn]

Yeah okay I just needed to hold that jump

What the hell is going on with our relative position during these runs LOL

...Huh, this reveals a lot about the routing, actually

[north barn]

@late wyvern scores for max-si-e-04-03

[late wyvern]

Interpolated highscores for my little meat tornado (MAX-SI-E-04-03) by Slomac:ansi 00: ekisacik - 82.714915 01: Eddy - 82.701591 02: Slomac - 82.430786 03: CeramicBlue - 82.296601 04: auds - 82.116715 05: aight - 82.023918 ++: -- DEV TIME -- - 82.000000 06: NateyPooPoo - 81.741983 07: vertexv - 81.124640 08: Wobert - 78.136756 09: Emmoji - 73.739607Scores: 10.

[north barn]

FUCK YES

hand destroyer

shake it off

[north barn]

@late wyvern screenscores for max-si-b-03-04

[late wyvern]

Screenshot for imposition (MAX-SI-B-03-04) by ekisacik v1 in palette quench:

Interpolated highscores for imposition (MAX-SI-B-03-04) by ekisacik:ansi 00: ekisacik - 86.016667 01: NateyPooPoo - 84.600000 02: Eddy - 84.583333 03: Slomac - 84.150000 04: CeramicBlue - 84.116667 05: aight - 84.033333 ++: -- DEV TIME -- - 84.000000 06: Wobert - 83.566667 07: vertexv - 81.916667 08: kkstrong - 80.916667 09: Nim “Mr. 7280” - 80.733333 10: Emmoji - 79.883333 11: auds - 79.766667Scores: 12.

[north barn]

I tried on this one. I don't normally for my own levels, but this one's different :)

(This level has oneways on it, by the way! Does not render in the preview.)

And that's where I'll end it off for the night. How'd I do?

@late wyvern max how many

[late wyvern]

ekisacik has 2 out of 150 highscore overall 0th scores for Max mappack.

[north barn]

Fuck.

goodnight

[pastel beacon]

Looking at this, do you know if you've beaten a dev time in-game? Or do you need to ask outte?

[gaunt ingot]

no dev times for episodes, the dev times for levels do appear in-game

[north barn]

yeah it's only on the level scores but it shows up as a time

[pastel beacon]

Ahh, episode view ofc

[pastel beacon]

trying to view level LBs of levels where I have a time crashes my game

(just played through the first episode quickly)

tried CTP and APEX tabs. no issues. just an issue with MAX

[pastel beacon]

@split patio - in case the message gets buried by highscoring convo

did a quick run of A-04. can view my level times without issue. Seems to be just an issue in A-00

tried with a blank nprofile; crashes in A-00 level times too, can see others fine.
done with debugging for now

[frigid summit]

I assume you've tried restarting the game, but if not, that's what helped me when I had some leaderboard issues with A-00 – everything else worked, but I couldn't submit any times for that episode, but it got fixed after quitting the game, reinstalling the tab, and opening it again.

Leaving the ping on in case it gets buried.

[pastel beacon]

Yeh tried a few things incl closing, restarting, uninstalling and reinstalling. Restarting comp as well.
I think I submitted my score ok the first time, but I went back in to see the dev times and yeh that's when it caused issues

@late wyvern full scores for max-si-a-0-0

[late wyvern]

Full interpolated highscores for Max (MAX-SI-A-00-00) by Slomac:
Scores: 20.

[pastel beacon]

Oh, didn't need full scores.
But yeh it submitted without issue

[frigid summit]

Hope you get it fixed, with or without Eddy's help. The tab is fantastic.

[fallow thorn]

These are not interpolated

[late wyvern]

Sorry, I didn't understand your command.

[vestal acorn]

all the 0th scores show as this in that first episode. no crashing though

nvm. I reloaded the game and it now crashes for me

[supple zenith]

Confirming that this happens for me as well. Just on the first episode’s level LBs

[north barn]

Oh, is it because of the dev time?

Cause don't those happen to be the full boards

[supple zenith]

I’m sure it is, but there are also full level boards on the second episode

[split patio]

I'm here, let's debug this

Is it only happening on full boards then?

[supple zenith]

It’s happened on boards with over 20 scores if you include the dev

So the first episode boards have 20+1 scores now

The rest of the column has full boards with the dev, but there are only 19 scores from players on those

[split patio]

Sounds reasonable, I'm probably sending 21 scores for those boards

Of course, that would completely blow up the game, cause why not

[north barn]

Not sure what the "20" could be from...

[north barn]

(referring to this)

[supple zenith]

The speedrun boards don’t crash for me, but they do have this same nonsense score at the top

[split patio]

They work differently, in particular, the global boards try to insert your personal score (stored in your savefile) into the leaderboard when it should be there based on the score. This might be causing the crash.

Alright, should be fixed now

[split patio]

I was getting garbage instead, I feel like it's the text starting to garble but stopping mid-way

[north barn]

Didn't need a game patch? Nice

[split patio]

Nope, it was purely server-side, I just had to truncate the boards to 20 scores

[frigid summit]

Eddy the 🐐.

[north barn]

I'm guessing that this prevents Scores Around Mine from working though unfortunately

This does bring up a really good question of how the game handles when the score on your save doesn't match the one on the servers (though, the pack does properly modify save data to match the displayed times). Having experienced the mac bug a few times I'm familiar it being missing from each

[split patio]

There's a wide set of edge cases to consider here, I tested a lot of that when I was first developing the custom leaderboards a couple years ago. Some things were not even doable because of the game actively messing with the leaderboards instead of simply displaying what the server sends.

But regarding what you mention specifically, I believe that it will always display the higher of the two scores. If your local score is lower then it will simply leave the leaderboards alone, but if your local score is higher then you can run into problems, since it will insert your local score into the leaderboard. You might think this insertion is done properly, but it can actually overwrite another player's score if your rank happens to change.

This happens in the global boards only, and maybe on the around mine boards (although they don't work exactly the same, none of the 3 do). It doesn't on the friends board, which made them ideal to use for speedrun mode without any corruption issues due to your highscore being different.

[north barn]

Fascinating, thank you

[north barn]

Oh, I just understood, it was because of the dev time being an extra time

While you're here, please take a peek at the bot failing to render anims that are below the dev time : )

[supple zenith]

Boards working here

[north barn]

-# I totally didn't do something funny on one of the levels that I wanted to post

[split patio]

Oh yeah lemme check that as well

[split patio]

Yeah because the dev score is not a real score in the db, I "inject" it artificially once the boards are almost ready, so I forgot to truncate to 20

To be fair, does a score below the dev time even deserve to be anim'd?

[supple zenith]

TFW you pick up five 0ths, check the rankings again and your count is exactly the same

[split patio]

lmao

[north barn]

-# I promise this is worth the bit

[jaunty marsh]

MAX 🕵️

[zenith owl]

get in on this yuqu

it's dope af

[tranquil fiber]

MAX

[frigid summit]

MAX

[winged anvil]

MAX?

[jaunty marsh]

even if i suffer from skill issues?

its probably dope if you like framewars i guess

[winged anvil]

I don't generally like frame wars, but I've been enjoying it so far (maybe because I've got in early enough that the boards aren't crazy yet)

[jaunty marsh]

that ping of Sky made me actually take a quick peek into the custom tabs channel, crazy how many tabs people created so far. props! i've been long enough off of N that i might be able to enjoy the game again, but i kinda want to improve my rythm gaming for a bit more i think

@late wyvern MAX rankings

[late wyvern]

Rankings - Highscore overall 0ths for Max mappack

-# Date: Jun 21, 2025 (16:47) UTC. MAX: 150.

00: Nim “Mr. 7280” - 29
01: jp27ace        - 25
02: Sky!           - 21
03: Eddy           - 19
04: SimVYo         - 14
05: NateyPooPoo    - 13
06: Slomac         - 11
07: CeramicBlue    -  7
08: vertexv        -  4
09: aight          -  3
10: ekisacik       -  2
11: vector7        -  2```

[jaunty marsh]

dang yeah that is still very fresh

[supple zenith]

I’m getting steamrolled by JP today

Sim too I think

[muted scroll]

jp27MAX

SimVMAX

[jaunty marsh]

what retirement does to a man

[oblique bear]

i suffer from skill issues, and it is fun as a casual pack too

[supple zenith]

@late wyvern scores for maxsia32

[late wyvern]

Interpolated highscores for 10, 9, 8... (MAX-SI-A-03-02) by filip:ansi 00: NateyPooPoo - 81.249963 01: jp27ace - 81.235732 02: Eddy - 81.217264 03: vertexv - 81.155012 04: kkstrong - 81.084988 05: Slomac - 81.072287 06: Nim “Mr. 7280” - 81.052150 07: aight - 80.910360 08: CeramicBlue - 80.786895 09: Sky! - 80.757477 ++: -- DEV TIME -- - 80.750000 10: ekisacik - 80.377345 11: auds - 80.143041 12: pralkarz - 80.061065 13: Emmoji - 79.531089 14: Wobert - 79.191111Scores: 15.

[supple zenith]

37 microseconds off of the next frame

[split patio]

Lowest I recall so far in this pack

But is it actually...

[supple zenith]

First two digit number I’ve seen on submission. Kind of took me by surprise lol

[split patio]

"Has Eddy changed this to milliseconds after all?"

+----------------+----------------+---------+----------+
| level          | player         | rank_hs | micro    |
+----------------+----------------+---------+----------+
| MAX-SI-D-02-00 | NateyPooPoo    |       7 | 0.000003 |
| MAX-SI-B-00-04 | EddyMataGallos |       8 | 0.000012 |
| MAX-SI-D-01-02 | vector7        |       3 | 0.000016 |
| MAX-SI-B-03-03 | vertexv        |    NULL | 0.000017 |
| MAX-SI-A-04-03 | vertexv        |    NULL | 0.000020 |
| MAX-SI-B-04-01 | Untamed Nim    |    NULL | 0.000021 |
| MAX-SI-B-01-03 | Untamed Nim    |    NULL | 0.000022 |
| MAX-SI-A-03-03 | pralkarz       |      10 | 0.000023 |
| MAX-SI-D-04-01 | ekisacik       |       9 | 0.000025 |
| MAX-SI-D-03-03 | Slomac         |       6 | 0.000026 |
| MAX-SI-D-01-00 | vertexv        |    NULL | 0.000032 |
| MAX-SI-A-03-02 | NateyPooPoo    |       0 | 0.000037 |
| MAX-SI-A-00-04 | Slomac         |    NULL | 0.000040 |
| MAX-SI-A-00-01 | Slomac         |       3 | 0.000040 |
| MAX-SI-D-04-04 | Untamed Nim    |    NULL | 0.000040 |
| MAX-SI-C-00-03 | DarkStuff      |    NULL | 0.000040 |
| MAX-SI-A-01-00 | kkstrong       |    NULL | 0.000041 |
| MAX-SI-D-00-00 | DarkStuff      |    NULL | 0.000042 |
| MAX-SI-B-02-00 | vertexv        |       8 | 0.000048 |
| MAX-SI-B-02-02 | pralkarz       |    NULL | 0.000049 |
+----------------+----------------+---------+----------+```

3 microseconds left only in that top run

Although your 37 is the lowest one on a 0th run

This is all excluding MAX-SI-B-03-04 of course (where everyone has 0 microseconds left)

(NULL rank = obsolete score)

[split patio]

Alright it should be fixed now, led me to tweak several other score formatting things, which was the root cause here

[supple zenith]

Man I wonder how many times in my multi-thousand attempt Metanet 0ths that I was only a handful of microseconds off the 0th.

Crazy to think you could be so close, but so far

[split patio]

Indeed, I've always wished obsolete runs were kept in the server so that we could scrape them all

And have completely detailed leaderboard history, but alas 😔

We could check all these things

[supple zenith]

Would be cool for sure

[north barn]

Thank you Eddy :>

@late wyvern anim 11 for max-si-c-03-04

[late wyvern]

Replay highscore trace for ekisacik in MAX-SI-C-03-04 using palette quench:

[north barn]

I hope you think it was worth it.

[split patio]

Damn, that looks so tight in the anim

[north barn]

I saw it and was like "...that has to be possible"