it was not hard

just needed to read it properly

adventofrealizingicantread.com

Under 10m

Quiet today...

2m

1.25m

1m, glhf!

^

.5m

.1666666m

1s

i was not prepared to think

phisicis

Two wrong answers on p2 danf

f

Make it three

Okay how tf did my p1 work

I did it all wrong

But I still somehow got the right answer

nice

Yay, top 100 for both parts!

embarassingly enough i literally have no clue how the fuck to solve this

This was a fun one

i literally don't even know where to start actually programming something for this

the leaderboard has completely filled and i don't actually have an abstract idea yet

genuinely busting out the pen and paper

Brute fuckin force

besides "brute force it lol?" except that... isn't actually working

where the fuck do you even brute force from

I brute forced it

just go in a spiral?

idfk

i guess

Worked for me

No, I just did every initial velocity up to like 1k for both

oh

I just brute forced all values between -500 and 500 for both

well at least i have no pressure and can just take however much time i want to think about this now

I picked too low of bounds at first, which cost me a lot of placement

true

i just barely kept my rank too

@UnrelatedString At some x value you'll overshoot first tick

but for the y

but ksun is now only 14 points below me

and my #3 gap is up to 148

which is technically still clearable but it's getting close to the end

I'm in 99th now, that's not bad

like i'm trying to obtain an upper bound because it seems like there should be one but there's nothing obvious

nice

I just kept increasing it by a few hundred untul my answer didn't change

crazy idea

there is probably a way to take any point and determine the required initial dx,dy to hit it

although that isn't unique so

maybe not

I liked today, it felt very open ended unlike yesterday

today is a good problem i'm just not smart enough for it

this is making me suicidal

^

but ironically

^

don't escalate me to cms

you know, it kind of feels liberating

to be able to just like

take it slow

think about this for as long as i want, go through different approaches and ideas, and just enjoy the problem

Oh yeah, the stress at your level trying to do it in like...top ten must be insane

since i never had a shot at the leaderboard in the first place i'm still freaking the fuck out because i like to still be able to pride myself on three digit ranks

holding my rank should in theory be fine because i am faster than ksun although as the problems get harder his CP experience might just overtake mine

@RedwolfPrograms yeah, suffering from success :P once i'm up that high already there's no "good enough" until i can get 1st some year hopefully, then i'll be fully satisfied lol

the thing is like

But then you're pressured to do it again, or else you'll feel like you've gotten worse :p

how do we know the answer is necessarily positive? for all we know we might have to shoot the spear down

@RedwolfPrograms yep...

you can probably get a pretty small lower bound because otherwise it'll just go under

the thing i had trouble with is like

how do you determine if the spear went under, over, too far left, too far right, etc

I did equally many negative as positive, but yeah the lower bound is going to be like -100 at most

and under only = too far left because the target is on the right side but what if it isn't

i feel like the way this is that's not even worth distinguishing

You only have to deal with the target being on the right though, right?

yes

well you can just scan over the input anyway

ok useful fact

it seems like it should just be an issue of your y velocity not sending it straight through

for part 1 at least

if you throw the spear up, it will always hit x=0 exactly

and the number of steps it takes to do that is 2dy+1

yeah

Oh, that's clever

and then it becomes equivalent to throwing the spear down at -dy-1

oh wait it

and since the x and y values are independent because this fluid only applies drag along one axis somehow

mmm

your upper y bound is just the lowest edge negated plus like one or whatever

i already got the x and y independent bit but i couldn't think of anything usable about y lmao

because otherwise when it hits x=0 it will just fall down too much at once and completely miss

okay, so we can easily get the maximum bounds for y

...you know, it would be nice if there was a worked out example that actually showed a maximal y velocity

yeah

and then given a particular dy, we can also then just run the simulation and determine during which steps it is in the correct y range

i think this might be the first time they've literally not given us an actual sample input-output

Didn't they

they do give us the output, 45

there's a bunch of stuff illustrating the physics of the probe but none of the samples have anything about maximal y

Also, it's definitely not the first time, they didn't do it on 2019 Day 25 (or for that matter, any of the more complicated intcode challenges)

oh wait nvm

literally cannot read

Anyway, done 1286/736

I used a inefficient brute-force approach and got the wrong answer once for each part by not having a wide enough search space

f

(It still took only a few seconds, though)

I did the same thing yeah

alright, so if we let my be the bottom y edge

And I was changing x instead of dx for drag, and somehow it still worked for p1

Interestingly, that wrong answer for part 1 (I think it was 4950) was the right answer for someone else

oh specifically

@RedwolfPrograms Which is either the world's most insane luck, or an obvious math thing that I missed

if you throw the spear up, its return down backtracks its y positions up

whoops

yeah i should've known this from like

physics class

@RedwolfPrograms Probably because in the winning part 1 answer what matters is the y value, and for every y value there is an x value that gets it in the x part of the target for both the intended drag calculation and your screwed up one

momentum on impact is same throwing something down or up

you know what i'm just gonna not make assumptions about where the target area is and go for completeness

so we literally just need the velocity that + 1 lands it straight in the box backwards

or not

that might not necessarily work

because it might be impossible to find a dx that lands it in the right horizontal range at the right step

yeahhhh

if miny is negative then we can lower bound dy at miny-1

otherwise just lower bound at 0 even though it could be higher but that takes math

It's interesting that in every single day except Day 14 I did better at Part 2 than Part 1

And that the same phenomenon wasn't true last year

actually easier to think about this in three cases

if the target area sits below the horizon, we lower bound dy at miny - 1 and upper bound it at -miny

if the target area sits above the horizon we lower bound dy at 0 and upper bound it at maxy

although the spec becomes ambiguous if the target area is above the horizon, actually

It was true in 2019 as well, but that year I binged the entire thing in a few days, and still saved Christmas with hours to spare

so i think we might actually want to just assume it's below

yeah

and then for each dy we can just keep running the simulation and check on which steps it's in range

wait fuck we don't need a lower bound

just keep decrementing until you find a valid dy

true

and then just search for if there's a valid dx that hits the range within those steps

you could bsearch but just like, brute force it lol

oh i'm fucking dumb

minx, maxx, miny, maxy = map(int, re.findall(r"\d+", data))

istg my original solution might've actually been valid if not for this

f

mfw 1off p2

f

my attempted brute force is just not giving me any results

rip

and now it's giving me too many

and now it's giving me a single wrong one

looks like a bounds issue

my problem is when the hell do i stop iterating

wait no i literally printed the first one that missed

      -------Part 1--------   -------Part 2--------
Day       Time  Rank  Score       Time  Rank  Score
 17   00:39:28  2712      0   00:48:06  2308      0

mfw

hey, at least i have not submitted a single wrong answer yet

true

copium

I'd be interested to see just how different our approaches ended up being lol

okay mine just gave me nothing again after like 30 seconds because i made the bounds huge

This was mine, full vanilla JS for the first time

github.com/hyper-neutrino/advent-of-code

or rather, p2 link probably more helpful: github.com/hyper-neutrino/advent-of-code/blob/main/2021/…

@RedwolfPrograms Full respect to Eric Wastl that is :P

lol

you know, i've only not placed 5 times total (d1p1 because internet, d6, d17), so that isn't really something to be all that disappointed with i suppose

...i may have identified a very dumb logic error

fixed it, gave me 139 despite that i already tried that, tried it again and unsurprisingly wrong

So wait why didn't you do a brute force? Was it just since you weren't going to get top 100 anyway, so y'all were going for something more elegant?

i know i'm trying to do a brute force

after trying to think of bounds then promptly giving up

@RedwolfPrograms i tried to go for a somewhat more elegant solution but it's still not fully smart cuz i could use math to make it better

oh wait i literally forgot to calculate the y position from the velocity

:/

or just record it since it's brute force anyways

...

4002/3163

how did you end up doing it

i literally just brute forced with the bounds we thought through

and somefuckinghow it took me half an hour to realize i was forgetting to output the maximum y coordinate for part 1

:/

i am typing this on my phone in bed and manually lowercasing my is

lmao

@hyper-neutrino in my case it literally worked but i put the velocity in instead of the corresponding triangular number

:/

i should've mentioned that

i thought of that like very early on even before i gave up lol

figured that might be a potential mistake

lol

I find it absolutely insane you can go into a store and buy a pocket calculator for a dollar. Think about that. It was most likely built in China, and shipped all the way around the world. It fits in a pocket, and does computation in an instant that ordinary people would only be able to start fully understanding in like the last two centuries. And at minimum wage, you can buy fifty of them with a day's wages.

this took much longer than it should've

mostly due to fiddling with the possible velocities, then realizing i could just optimize for the input anyway

lol