My solution was based on that, but specially dealt with the case when $$$M=1$$$ (in this case we have an obvious winning strategy). I don't know whether that is correct.

Fell down from 2800+ rating in the last ARC++ :)

Well the previous AGC was about 5 months ago and the next one has not appeared yet

I submitted G at 2:28:44 then got Pretest passed with 2000 / 2000 ms. I think I'm going to get TLE in system test :(

Upd: I was so lucky that the submission passed in 1953 ms! :)

I learned from zero how to process an image in python during the contest, just to solve problem H...

C is very similar to https://qoj.ac/problem/8527.

Why can D1B2 be solved in $$$\mathcal O(n^2)$$$ but has $$$n\le 100$$$? Why can D1C be solved in $$$2n+\mathcal O(\log n)$$$ operations but has $$$3n$$$ limit?

I also (always) use MOD for modulus.

Hope the judges work well.

My code got rejudged again and TLE after the announcement of "The system testing and rejudges are done."????

Please give me a reason for this chromate00

How I solved D...

I got stucked on C and went to D.

I remembered that in the 1D version we only need to look at the (at most) $$$3$$$ characters in the middle.

Then I guessed that we also can just look at the central $$$3\times 3$$$ area in the original problem.

I submitted that algorithm but got WA.

I was disappointed. I thought my algorithm was completely wrong and hopelessly changed the area to $$$4\times 4$$$ then submitted it again.

...Then I got Accepted!

The distance between Takahashi and Aoki (as a function of time) is convex unimodal when both of them is moving OR one has stopped, but not when you put the cases TOGETHER.

So you need to do ternary search separately for both of the cases.

Edit: it's not "convex", should be "unimodal".

Hope to get IGM. Just 23 points away.

Upd: Failed to debug E and is getting -ve delta. :(

Your solution are actually correct.

Because when you rotate $$$v_{\lfloor\frac n2\rfloor}$$$ to be perpendicular to $$$v_0$$$, consider the acute angle between them. There are at least $$$\lfloor\frac n2\rfloor-1$$$ lines in this angle (specifically, $$$v_1,v_2,\ldots,v_{\lfloor\frac n2\rfloor-1}$$$ or $$$v_{\lfloor\frac n2\rfloor+1},v_{\lfloor\frac n2\rfloor+2},\ldots,v_{n-1}$$$). Obviously they, as well as $$$v_0$$$, will be always getting further from $$$v_{\lfloor\frac n2\rfloor}$$$ during the first rotation. That are at least $$$\lfloor\frac n2\rfloor$$$ lines, so the efficiency always increases during the first rotation.

Now $$$v_0$$$ and $$$v_{\lfloor\frac n2\rfloor}$$$ are already perpendicular. Obviously, the sum of distance to the two lines of any other line is $$$25000$$$, which is a constant. So we can just "delete" the two lines, remaining a subproblem of $$$n-2$$$ lines.

By induction, the correctness of the solution can be shown.

OMG round $$$(1111111111)_2$$$ as a tester!

The third conclusion is completely wrong... for example $$$(1\times 3)\oplus(3\times 3)=3\oplus 9=10\ne 6=3\times 2=3\times(1\oplus 3)$$$.

Note that this also teaches us a lesson about solving an unknown problem: Don't completely believe conclusions by just try some most simple cases to check them!

As a tester, the problems are challenging and, of course, Turtle is sooooooo orz.

Wish you all have fun with this contest!

Yet another speedforces round. Failed to get to IGM :(

Ok I became the 'Grandmasters' in your words... /sad /sad

That's because the contest type has too much randomness. Sometimes we call it 'lottery contest'.

Be careful of ARC AGC-ization!

Why can't I vp the contest :(

It showed You are not authorized to fill out this form when I click the Start virtual contest button.

MikeMirzayanov please fix it

Why E and G have (basically) same statements but different constraints? I solved G first, then got RE on E because of keeping the size of arrays $$$2\times10^5$$$...

But in the picture here the last two characters of the string is ba.

(And it says that the rightmost position is invalid, so I think the picture here is wrong.)

There is a typo in the first picture of G's solution: the string is $$$\tt{abaabaabaabaaba\color{green}{ba}}$$$, instead of $$$\tt{abaabaabaabaaba\color{red}{ab}}$$$.

Why my stupid $$$\Theta(n\log^2 v)$$$ solution for F ran for just 827ms, can sb hack it or did it just run too fast because of constant factor? 297330202

Does the div system means there will be fewer ARCs (~20 $$$\to$$$ ~8) for oranges? :(

(And there are too few AGCs.)

But if the first submitter intentionally leaked the code, they also violated the rules and should be punished.

I want to get full score in NOIP :)

the difficulty is close to that of ARC184

I think that contest already had a difficulty of an AGC...

Ok now it's even harder :(

Well, in fact there is a Show rated only button in Customize.

Well the condition of we div.1 are even worse...

Passed the example of D 5min after contest ends...

As a tester, I am zltzlt and Zunpet fan.

On "七夕节" (a Chinese festival about love), I'm going to participate this contest with only myself in the computer room.

Origin & difficulty & data aside, I think F and G are good.

Yes, there is a similar (template) problem on luogu. Someone passed it in 2 min and they probably copied their code.

ACAM in ABC362, then Steiner tree in this round?????? Will ABC365 have a poly multipoint eval as G?????????

String, heavy DS, heavy math

You are right, but G is an easier version of this problem (that this problem is on a tree). I think many Chinese participants have seen this problem. (Though few people solved it) Atcoder, it's a good job and don't set problems like this again.

Atcoder, if you aren't able to set Gs then just remove them.

In the picture on the top of the website, why the username is "maroonk"?

Support!

And I wonder how he (she?) generated them...

Stucked on C for $$$45$$$ minutes and finally got 1 place lower than sevlll777, who solved 1 problem less than me. The slowest episode ever.

No because $$$3500-2500\ne2500-2000$$$ :)

Good luck & high mark in Zhongkao!

I am rui_er fan!

Unfortunately I can't participate, but GL&HF to you!

I overthought C instead...

No it's super involution :)

Problem F is almost same with https://www.luogu.com.cn/problem/P3625

I commented so I must drop to CM by September?

Winner of Div. 2 is called WanYuanShenWanDe??? Why Genshin Impact ("YuanShen") is everywhere, even in Codeforces?

ABCD is so easy (for me) that I solved them in 1 hour.

No jiangly will win

Upd: oh he really won! orz

Hope solve 1 problem in platnium :)

Doesn't Codeforces have these things? Example: CF version of luogu discuss 61884, xlpg0317's rubbish blogs & comments

Luogu usually punishes some contests and problemsetters if the contest is bad, for example preventing them from problemsettings. Luogu punished 8 contests in 2023, but Codeforces did not even publish a clarification after Goodbye 2023 with the blog having 4500+ downvotes!

Off-topic: I got rank 2 in Caféforces round 2. (In fact, there are also some other codeforces style contests on luogu.)

It's just the Goodbye 2023 on Atcoder. Imbalanced difficulty, existing problem, letting wrong solutions pass...

It also disrupts some tools like CCH. Cloudflare will be checking the tool forever... :(

The higher, the better :)

from itertools import chain, combinations, permutations
from collections import defaultdict, Counter, deque
from math import ceil, floor, gcd, inf, log2, sqrt
from bisect import bisect_right as upper_bound
from bisect import bisect_left as lower_bound
from random import choices, randint, sample
from functools import lru_cache, reduce
from heapq import heappop, heappush
from sys import stdin, stdout
from gzip import decompress
import builtins

and then some useless algorithm templates,

and a line which is the only line that solves the problem:

exec(decompress("something I can't understand"))

finally, some useless operations on a variable.

He must tried very hard to cheat. LMAO

Now I can see Tutorial Good Bye 2023 (en) on the contest page. But when I click on it, I see

Can't find such topic

What the f*cking sh*t is it???

So where is the tutorial?

Why so many Gs (and Ex when there was Ex) are math?

Delayed agaaaaaaaaaaaain... Okay I'll have to go to sleep 15 min later...

Wasn't able to solve D during the contest :(

https://atcoder.jp/contests/arc168/submissions/47766688

Seems the Div. 3 is BEFORE NOIp...

Did you mean the Div. 2 after NOIp?

But neither of them is Div. 1...

No it will get MLE :(

And the memory limit is intentionally set to $$$512\operatorname{MB}$$$, since you need long long it will cost $$$22\times2^{22}\times8\operatorname{B}=704\operatorname{MB}$$$.

UPD: Ohhhhhhhhhhhh I optimized the space (halved it) and got AC

Codeforces rounds >>>>>>>> school exams

Div. 1 B and C1 are great! (Maybe also C2 but I didn't solve it) I like the way to optimize the algorithm.

Hope IM.

Upd: Success!

Div. 1 A can be solved with multiset in $$$\Theta(n\log n)$$$: submission

Actually A-F (maybe even G) was easy to think, but the implementation...

Why it is downvoted?

UPD: it'is upvoted now! :)

It seems the ratings are coming back! :)

My rating rose to 2042 from 2028 but today I found it returned to 2028. Is my rating rolled back? I found no evidence that the contest will be unrated.

Tomorrow I will fall 100 rating in CF first and then fall to blue in ARC.

The given polyominoes are connected. In other words, the squares that make up a polyomino can be reached from each other by following only the squares up, down, left, and right.

It seems that the first polyomino of the second hack violates the constraints.

D is fkin implementation

E is fkin DP

F is fkin segtree template

Look at the score distribution, so weird...

In that case you know how much points you will get for resubmitting now.

When you realize your code is wrong, won't you correct your code in any case to avoid FSTing?

Maybe also hide solved problems?

1e11 ($$$10^{11}$$$) is not large enough for infinity. The answer can be up to $$$2\times10^{14}$$$.

You are wasting time typing these words.

It works well for me now.

Number of Tasks: 8

from this ABC, we will use 7 problems per contest.

I just meant the Finals NOT THE MIRROR.

So I registered unrated.

Considering that the special rules will affect strategies of participants, the Finals itself is unrated.

But now the contest page is still showing

Rated Range: 2000 — inf

Where to enter CF handle? I can't find it.

Keep orange please :)

You are not allowed to view the requested page

SkyWave2022 will AK IOI

But here says AtCoder World Tour Finals 2023 will take place on Sep 8th and 9th. So I think they should be 2023.

Why their names are World Tour Finals 2022? They should be World Tour Finals 2023.

OMG I will have to sleep one hour later...

Wish to become Master. Just a bit below it :)

Candidate Master needs 1900 rating, not 1800.