We will hold AtCoder Beginner Contest 428.
- Contest URL: https://atcoder.jp/contests/abc428
- Start Time: http://www.timeanddate.com/worldclock/fixedtime.html?iso=20251018T2100&p1=248
- Duration: 100 minutes
- Writer: sheyasutaka, Nyaan
- Tester: sounansya, physics0523
- Rated range: ~ 1999
- The point values: 150-200-300-400-475-525-625
We are looking forward to your participation!









Is this new judging system an update to the judge machine’s configuration, or does it simply add support for more programming languages? If they’ve actually changed the judge machine, maybe programs could run a bit faster? (I have a bit of a TLE phobia…)
No C++ 17 or 20 is wild
Actually there is C++ 20 (precisely, GNU++20) but IOI version (No AtCoder Library, boost etc)
language brainfuck is among availible languages, wtf
++++++++++[>++++++++++<-]>--.++++++++++++++++.-----------------.++++++++.+++++.--------.+++++++++++++++.------------------.++++++++.
Why 30min later?
Strange start time.
Did you update the contest start time?
No he didn't
Is the contest delayed by 30min? Technical issues of the new judge?
upd:according to contest page it is delayed and new judges won't be used
It seems to be slower and slower for us to visit AtCoder... Even submitting a problem takes a few minutes :(
Yeah, there was a time that I even couldn't enter it in half an hour
Is it rated?
Classic question.(((
Hope it will be high quality.
why did E<<<D???
same feeling
The comparision of tuple in E gives TLE under correct time complexity.
Don't know why E was even included in this contest. D >>> E imo.
I found E harder than D, it's a matter of taste. One could easily take a long to code approach in E and spend more time solving E than D even if the idea is conceptually easier.
E is such a classic problem.Bad E.
Yes.Just need to know diameter of the tree and facts about it
but for beginners is a good chance to learn about the property of the diameter of a tree
that's why the contest call"abc" i guess
Alright, I agree with u. I'm a beginner as well.
(The favs only)
This is too bad.How the difficulty gap between D,E and F,G so big?
And E is too classic that it is a shit here.
Though I spent too much time on D,I think it's a good problem here.
I agree with you. I think D is good, but E is rubbish(it's too classical and boring). What's more, the difficulty gap between E and F is too big.
What is wrong with my code for c
i think E > D :(
E solution is that take the diameter of the tree and do the bfs statistics to the dist of these two nodes.
But question D seems to be able to be written using binary search. Please give me some advice.
Wow! Coder ME competed in Atcoder Beginner Contest 427&428 and gained -97 rating points! Share it!
Rubbish Round
swap(D,E) plz
I think D>>>E.
i think so
another speedcoder?? btw is F an ad-hoc? I simply didn't understand how the official editorial works
Could the new judging system be opened for testing before the next competition starts? I don't want to face a completely new and unfamiliar system when I get to the ABC contest.
By the way, long live chez scheme!
There actually is a language test, https://atcoder.jp/contests/language-test-202505, but it's only shown in Japanese mode (Japanese home page -> 言語更新ADT先行運用のお知らせ(Announcement of early operation of language update ADT) -> https://atcoder.jp/posts/1566 -> Language Test 202505
thanks for reply
My Java code strangely TLE and has an O(n) complexity. I tested it locally, timing the read from n to the output, and it only took 1300+ms. I can't figure out why it times out. I've never encountered such a Java timeout before. Does this mean Java won't be able to use DFS for 5e5 data in the future? Here's the submission link. I think it's due to the OJ system changes. Could atcoder officials please check it out?
https://atcoder.jp/contests/abc428/submissions/70268852
I haven't inspected your code sry, but FYI AtCoder has not changed their system for this contest (they planned but then rolled back due to a technical issue).
Well, there must be something wrong with my code. I found that Java's DFS is 5 times slower than BFS, while C++ has no such difference.
swap(D,E) please
D is terrible for me and I got rating-9 at last
E is too simple for this place.
How to approach problem C can anyone please help
my approach for D (This could help to understand better)
Could someone call me what's wrong with my code in problem F QAQ. I found that when migrating operations 1 and 2, there are actually intervals smaller than v that may move, so I tried using a segtree to solve this problem. At the same time, the line segment tree can also be used to split operation 3. But I successfully passed the sample and 10 points, but still had 6 WA points QAQ. Here is my code, because I am not very familiar with it. The lazy markers are relw_ay and rel_num, and lazy_ay and lazy_num correspond to whether the interval is offset to the left or right and the corresponding offset position. (I am a Chinese, please forgive me for not reading correctly. qwq)
ok,I konw, for query 3 ,x is possible to be larger than Wn ,I didn't consider this QAQ, it is a bit stupid (). if someone need ,here is my last AC code
I didn't even have time to solve E as I spent a lot of time of D! E is too classic!
My issue
Why is my own code for D wrong? In the editorial, it says that we can use
floor(sqrt(x))instead of the functionf(x)to compute $$$\lfloor \sqrt{k} \rfloor$$$. But when I usefloor(sqrt(x)), the judge result is this. However, when I finish my code according to the editorial, my judge result is this. Could anyone explain the reason please?I feel the same way—something seems off, but I haven't figured out what it is yet.
Update: Might be the issue relates to floating-point precision