Misuki Profile Banner
Misuki Profile
Misuki

@Misuki998244353

281
Followers
363
Following
92
Media
659
Statuses

陽明交通大学の四年生、日本語を勉強しています。 競プロ(特に数え上げ)が好き。 CodeForces(薄橙) Atcoder(黄) (id: Misuki) ICPC'23 world finalist pfp:

Taiwan
Joined January 2019
Don't wanna be here? Send us removal request.
Pinned Tweet
@Misuki998244353
Misuki
8 months
今年の目標 - CF赤になる - AtC橙になる - これなるべくを解く(ad-hocの練習, 橙~赤)
0
0
7
@Misuki998244353
Misuki
1 month
このチーム名www
Tweet media one
0
33
180
@Misuki998244353
Misuki
4 months
ARC rated 橙以下全部埋めました! うおおお
Tweet media one
6
0
72
@Misuki998244353
Misuki
29 days
5000 AC! (I don't know why it can't fetch data for Codeforces, so I edit it by myself lol)
Tweet media one
3
0
42
@Misuki998244353
Misuki
2 months
次のDiv.2のライターです。対戦よろしく! I'm the writer of next div.2, good luck to all participant!
5
7
39
@Misuki998244353
Misuki
1 month
tourist is the title for rating >= 4000 XD
1
5
37
@Misuki998244353
Misuki
1 month
AGCが生えました👀
Tweet media one
0
1
28
@Misuki998244353
Misuki
2 months
ご参加ありがとうございました!問題を楽しんでいただけたら嬉しいです :) thanks for your participation! hope you like the problems :)
4
0
24
@Misuki998244353
Misuki
2 months
first page!
Tweet media one
2
0
21
@Misuki998244353
Misuki
7 months
first page!!!
Tweet media one
1
0
21
@Misuki998244353
Misuki
1 month
Just feel having a stack of unsolved problem in mind is actually quite good since you can literally thinking about them anytime anywhere. And I've mind solved a few red problems these two day when I can't use my notebook😄
2
0
20
@Misuki998244353
Misuki
2 months
Just start reading a MO book and I found it cover most of the basic combinatoric/number theory knowledge for CP very well. I haven't seen such a complete compilation of such thing, usually these things are scatter around various blogs/problems in my experience :P
3
0
18
@Misuki998244353
Misuki
1 month
Bは面白いけど maroonrkさんの問題はマジ苦手><
Tweet media one
0
0
18
@Misuki998244353
Misuki
2 months
おおおおー 勝ちました😃😃
Tweet media one
3
0
18
@Misuki998244353
Misuki
2 months
btw, I was trying to invent some ARC-style problem initially, but it's far harder than I thought. most of the problems I came up either turns out too impl heavy or the thinking part is too less / non-interesting. seems I still have a long way to go for problem setting.
2
0
17
@Misuki998244353
Misuki
2 months
Everything is finally over, I think I can write sth about the round :) [A, B]: simple and easy to prove ad-hoc [C]: after original C got reject, I can't think of any problem with this diff, they are either too hard or too easy for its position
2
1
15
@Misuki998244353
Misuki
2 months
爆死した ちょっと辛い
Tweet media one
1
0
15
@Misuki998244353
Misuki
11 days
RPS百万を超えた
Tweet media one
0
0
14
@Misuki998244353
Misuki
3 months
Rated橙diffを埋めました 次は赤diff(厳しそう...)
Tweet media one
1
0
12
@Misuki998244353
Misuki
20 days
WF really don't like counting problem
2
1
12
@Misuki998244353
Misuki
3 months
CF #958 [A]: ceil((N - 1) / (K - 1)) [B]: after compress 0...0 into 0, check [cnt1 > cnt0] [C]: k <= popcount(n) + 1, which is achievable except k = 1 [D]: # turns is O(lgn) since after a round, the rest vertices should connect to one chosen node, so max |CC| halve [E]: casework
3
1
12
@Misuki998244353
Misuki
2 months
AGC004-E Salvage Robots AC! なんか古いAGCの銅diffが解けそうな気がする やってみようか👀
0
0
11
@Misuki998244353
Misuki
2 months
solve ABCD, the problems are pretty good and the difficulty is similar to ARC : )
Tweet media one
3
0
11
@Misuki998244353
Misuki
1 month
Just heard there are a lot of Indian teams register TOPC (preliminaries of Taiwan regional ICPC), was wondering why, I thought only teams in Asia Pacific can participate👀
1
0
11
@Misuki998244353
Misuki
2 months
virtual last div.1+2 and brick on D hard 🥲
Tweet media one
3
0
11
@Misuki998244353
Misuki
17 days
First round of new ARC!
Tweet media one
0
0
10
@Misuki998244353
Misuki
2 months
AGCぞい!
Tweet media one
0
0
10
@Misuki998244353
Misuki
6 months
ok, we gonna stuck in Cairo (again) 💀
Tweet media one
1
0
10
@Misuki998244353
Misuki
14 days
finally an div.1 that is not too late🙌
Tweet media one
0
1
10
@Misuki998244353
Misuki
6 months
have fun with various people :)
Tweet media one
0
0
9
@Misuki998244353
Misuki
18 days
implementation of G is 💀💀💀
1
0
9
@Misuki998244353
Misuki
6 months
secret place?
Tweet media one
0
0
9
@Misuki998244353
Misuki
3 months
Just start to have some motivation to solve ABC because I found I can prepare some template while solving them, but maybe I'm overpreparing the template when something like this appear in my default template lol
Tweet media one
1
0
9
@Misuki998244353
Misuki
2 months
ABC 366 [A, B]: [C]: map maintain frequency of each value [D]: 3D prefix sum, 2D is also acceptable [E]: calculate contribution of each dimension independently, enumerate x then binary search possible y [F]: find optimal order then sort then dp
1
0
9
@Misuki998244353
Misuki
3 months
finally solve this one after two years XD
Tweet media one
0
0
9
@Misuki998244353
Misuki
2 months
ARCやる
Tweet media one
0
0
9
@Misuki998244353
Misuki
2 months
ABC 363 G - Dynamic Scheduling これを解けた 実装の細部をバグ過ぎて2時間かかりました むずい
1
0
8
@Misuki998244353
Misuki
2 months
ABC
0
0
8
@Misuki998244353
Misuki
2 months
[C+]: thanksfully feecle6418 comes to the rescue with a really interesting problem :) [D]: a standard problem for div.1, but seems not for div.2 [E]: was thinking about creating a counting problem make use of symmetry a lot, but turns out there are other solution exist?
0
0
8
@Misuki998244353
Misuki
11 days
ABC373 [D]: potential DSU, or you can just DFS [E]: I use two segtree + bs... there should have easier solution [F]: (min, +) convolution [G]:
0
2
8
@Misuki998244353
Misuki
2 months
Just learnt static toptree, and seems it's tree version of segment tree as you can do path/subtree modification/query all in O(lgn) and point update tree dp and even CD on it while the implementation is not hard. Maybe it would become next generations' segtree?
1
0
8
@Misuki998244353
Misuki
1 month
TOPC(Taiwan ICPC regional preliminary) is starting
0
0
8
@Misuki998244353
Misuki
2 months
ABC 368 [C]: total damage deal in (t1, t2] is (t2 - t1) + 2 * (t2 / 3 - t1 / 3), then just binary search [D]: (the number of vertex whose subtree have [1, K) vertices in V) + 1 [E]: process conditions in increasing order of time, for each vertex v, maintain max time a bus arrive
1
0
8
@Misuki998244353
Misuki
29 days
@akshansh_777 Thanks you! It's time when I create my CF account, so about 5 year? I start to practice much only after studying college though, which is about 3 years ago.
1
0
8
@Misuki998244353
Misuki
1 month
Register abstract algebra lecture this semester, hope I can understand what the professor is teaching 🙏
3
0
8
@Misuki998244353
Misuki
2 months
ABC 364 [A, B]: [C]: greedily pick from large A/B to small A/B [D]: binary search min x where there are >= k points in [b - x, b + x] [E]: dp[# dishes eat][sum of A]: min sum of B [F]: use idea of Kruskal to find MST and maintain set of segments forms a CC [G]: Steiner tree dp
0
0
7
@Misuki998244353
Misuki
5 months
trying to solve some class problem, seems problems start to become challenging from class 9!
0
2
7
@Misuki998244353
Misuki
4 months
ARC 104 F - Visibility Sequence 昨晩から考え続けて、「解けた! -> 間違いました -> 解けた!」のループに入り、今やっと解けました
0
0
7
@Misuki998244353
Misuki
3 months
Just learnt it's possible to precompute modulo inverse of a_1, a_2, ..., a_n for any a in O(n + lgp), always thought it can only be done in O(nlgp) :O
2
0
7
@Misuki998244353
Misuki
25 days
Gにやられた😣
Tweet media one
0
0
7
@Misuki998244353
Misuki
6 months
finally back to home :D
0
0
7
@Misuki998244353
Misuki
2 months
gl&hf to all participant!
1
0
7
@Misuki998244353
Misuki
4 months
Can't we have shorter Div.1 contest? 🫠
Tweet media one
2
0
7
@Misuki998244353
Misuki
6 months
下手クソな日本語でtatyamさんと写真を撮ってもらいました 嬉しい!
0
0
7
@Misuki998244353
Misuki
3 months
ABC 363 [A]: 100 - R % 100 [B]: max(t - sort(a)[n - 1 - p], 0) [C]: bruteforce all permutation [D]: kind of like binary search on tree [E]: Dijkstra but edge weight of (u, v) is max(dis_u, a_v) [F]: O(N^{2/3}) divisor dp [G]: ??
1
0
7
@Misuki998244353
Misuki
3 months
virtual yesterday's div.1+2 and the result is pretty good!
Tweet media one
0
0
7
@Misuki998244353
Misuki
1 month
#ARC 生えろ
0
0
7
@Misuki998244353
Misuki
2 months
@FLzY8 sorry, it's kind of hard to cutoff all O(n^4) while keep all O(n^3) pass since the atcoder modint template is insanely fast compare to normal way to deal with mod : (
1
0
7
@Misuki998244353
Misuki
1 month
ABC369 [C]: a subarray is arithmetic sequence <-> its difference array only have one kind of value [D]: dp [E]: precompute all pair shortest path, then for each query enumerate all possible order of visiting K edges [F]: come up with O(nm) naive dp then use segtree to optimize
0
0
7
@Misuki998244353
Misuki
5 months
ABC 354 [A]: [B]: sorting [C]: monotone stack [D]: find the period of rectangle then do some math to find occurrence of each grid [E]: bitDP [F]: run LIS for prefix/suffix and check whether a_i can be insert between them [G]: Dilworth's theorem
0
1
6
@Misuki998244353
Misuki
3 months
E - Rearrange and Adjacent XOR Didn't know anything when trying to solve it last time, but this time, I solve it without any guessing, it's amazing!
0
0
6
@Misuki998244353
Misuki
5 months
2287->2370(+83) IMに戻った
0
0
6
@Misuki998244353
Misuki
3 months
Just investigate a bit on how can we find lca faster than <O(nlgn), O(1)> or <O(n), O(lgn)>, then I found this! by doing this, we can find lca in <O(n), O(1)> online, and the idea behind that blog is pretty clean
0
0
6
@Misuki998244353
Misuki
17 days
難し過ぎ!
Tweet media one
0
0
6
@Misuki998244353
Misuki
17 days
Multiplication of Big Integers AC!(394ms) I was trying to make every entry store value upto 10^9 but it's quite annoying to deal with overflow so I give up and store upto 10^1 and use 1 NTT instead of 3, and turns out it's still quite fast!
0
1
6
@Misuki998244353
Misuki
3 months
ABC [A]: vector::insert [B]: test if two segment intersect for all dimension [C]: binary search [D]: BFS [E]: 2 * sum(c) - diameter [F]: bruteforce a in [1, 10^6], check a > 10^6, b = 2 by taking sqrt(n), you can store all a^b for a in [1, 10^6] to avoid overcount [G] impl hell
0
0
6
@Misuki998244353
Misuki
1 month
Just rethink about shortest path algorithms a bit, actually none of them would try to avoid using same edge twice, so they are actually computing shortest walk instead of path, but it just turns out 1st-shortest walk would always be a path. An example to verify this is that, if
1
0
6
@Misuki998244353
Misuki
1 month
@ORANG3_CRUSH for CP oriented geometry, this one cover it pretty well
1
0
6
@Misuki998244353
Misuki
5 months
GL&HF
Tweet media one
1
0
6
@Misuki998244353
Misuki
2 months
AGC 067 oxxxx [A]: G must have clique with size n / 2, so n can't be too large, then X is clique on G <-> X is independent set on complement of G, and G is good <-> comp(G) is bipartite.
1
0
6
@Misuki998244353
Misuki
5 months
class 9👀
Tweet media one
2
1
6
@Misuki998244353
Misuki
3 months
ABC 362 [A]: [B]: [C]: check sum l <= 0 <= sum r [D]: Dijkstra [E]: for all O(n^2) d for arithmetic sequence do dp [F]: for each edge, upper bound of # match cross it is achievable by letting every (x, y), x, y belong to different subtree when rooting centroid [G]: AhoCorasick
0
0
6
@Misuki998244353
Misuki
7 months
Too slow :(
Tweet media one
0
0
6
@Misuki998244353
Misuki
6 months
didn't perform well this time, but I guess that's it. Maybe try harder next time if I have another chance
0
0
6
@Misuki998244353
Misuki
3 months
Let's goooooooo
Tweet media one
2
0
6
@Misuki998244353
Misuki
25 days
mod p^kの制約を管理すればいいのか なるほど
0
0
5
@Misuki998244353
Misuki
1 month
@FLzY8 imagine use this team name and got 1st at regional to rickroll all people 💀
1
0
5
@Misuki998244353
Misuki
5 months
tried to find a random icpc problemset on baekjoon and sorted the problem by difficulty then virtual it as if I'm doing some cf round. It was quite fun!
0
0
5
@Misuki998244353
Misuki
5 months
今日のARC unratedで参加する 寝不足から
0
0
5
@Misuki998244353
Misuki
6 months
fifty... thousand?
Tweet media one
0
0
5
@Misuki998244353
Misuki
6 months
participating open ceremony
Tweet media one
Tweet media two
0
0
5
@Misuki998244353
Misuki
4 months
[F]: d is valid degree sequence for tree <-> d_i > 0 for all i and sum of d_i = 2(n - 1), then greedily increase the degree of vertex with min cost [G]: use hld-like dp to maintain frequency of each color under subtree
0
0
4
@Misuki998244353
Misuki
5 months
勝った!
Tweet media one
0
0
5
@Misuki998244353
Misuki
2 months
first time participate in TOKI!
0
0
5
@Misuki998244353
Misuki
1 month
ABC
0
0
5
@Misuki998244353
Misuki
4 months
ABC 359 [A] [B]: if a_i == a_{i + 2}, ans++ [C]: we can go to (±1, ±1) or (±2, 0) in one step, use the first to make s_y == t_y then use the second to make s_x = t_x [D]: bit mask dp [E]: monotone stack
2
0
5
@Misuki998244353
Misuki
3 months
yuki 435 [A] pollard rho [B] greedy [D] pollard rho + dp seems KACTL's pollard rho have bug :/
0
0
5
@Misuki998244353
Misuki
3 months
yet another 3 hour round
Tweet media one
2
0
5
@Misuki998244353
Misuki
6 months
excursion, the wether is sooo hot
Tweet media one
Tweet media two
Tweet media three
0
0
5
@Misuki998244353
Misuki
3 months
@akshansh_777 I think all of them is one idea ad-hoc, so it's quite easy to have non-sorted diff for a contestant
1
0
5