We will hold AtCoder Beginner Contest 393.
- Contest URL: https://atcoder.jp/contests/abc393
- Start Time: http://www.timeanddate.com/worldclock/fixedtime.html?iso=20250215T2100&p1=248
- Duration: 100 minutes
- Writer: Nyaan, yuto1115, ynymxiaolongbao
- Tester: MtSaka, sounansya
- Rated range: ~ 1999
- The point values: 100-200-300-425-475-500-675
We are looking forward to your participation!
Why does this have so many downvotes?
Judging by now, this ABC is likely to get a lot of downvotes in the future. (Reason: A~F is quite easy. But now no one passed G.)
Update: A contestant passes G! Such a strong programmer!
Hope this round can be harder and more interesting than last one. Please do not make standard problems any more!
Wish Atcoder Beginner Contest will not be 「AI Beat Contest」.
By the way, it seems AGC disappeared.
Wish AtCoder Beginner Contest won't be "AI Beat Contestant". Also, there's AGC071 on March 30th.
"Please do not make standard problems any more!"
Why? As name of the contest suggests it's the contest for beginners. There is a rule against AI in ABC contests, so one shouldn't care about AI and just enjoy the problems
Actually, if you do want to solve standard problems, you have many choices better than ABC.
Most people, like me, just want to see more interesting and enjoyable problems in AT contests. These boring problems and more and more AI participations truly annoy me.
I can say the same, if you want to solve interesting and enjoyable problems, you have many choices better than ABC. You should understand these "standard" problems aren't standard for beginners and this contest mostly designed for them
I agree with you ! As a complete newbie to this platform, I am grateful such problems exist so that I too can reach the experienced level in the future just like others who find these problems easy or unexciting.
Terrible speedcoder. Took <40min to AK Div 4 in disguise. The last problem is just extra...
If next ABC as shit as this, I will ban you.
Although you are not the real system user, I still give you an up.
maybe we should call this: AI Beginner Contest
it this a AI Beginner Contest?
We will hold AI Beginner Contest 393 for younger AI training.
if they set non-AI solvable contest, it won't be ABC anymore. Why are you complaining? It's not authors fault that "contestants" don't follow the rules
Who can hold a better contest on every Saturday evening?
Our coach orders us to take part in ABC.I want an other contest instead ：(
Oh, maybe G should be used in AHC, not ABC.
ez ac ABCDEF,Im vegeableness
i good vegetable ah
The length of the solution is 30,000 characters!!!It's unbelievable!!!
Update:The problem is G.
What problem?
How G
Looking at the editorial to G, I wanna kill myself.
+1(The solution left me broken and repeatly shouting "F**K!")
Terrible contest :( I would prefer quality over quantity if this is happening due to more frequent ABCs. Such contests are ruining the overall leaderboard as well.
Why can not we use simplex method to solve problem G?
I haven't tried it yet, has anyone tried it?
Because they intentionally made simplex infeasible with uncanny error tolerance (as mentioned in the last sentence of the editorial). This is the best I got.
Thank you for your response. I have carefully re-read the editorial for problem G. In fact, it said if the precision of the simplex solver is sufficiently high, it is possible to accept this problem. There is a publicly available high-precision simplex solver on the Universal Online Judge's simplex template problem. I plan to find some time to try using this solver to attempt to pass the problem.
I believe G is too hard as an ABC G, but since ARC and AGC both focus on thinking, and AHC shouldn't contain problems solvable in deterministic time, I think it should take place as the hardest problem in an ICPC Contest.
IT SHOULDN'T APPEAR HERE!!!!!!
I solved ABCDE lightning fast lol. Weren't they kind of really easy?
how did you solve E? I tried various factoring methods but couldn't figure it out.
my solution is similar to sieve of eratosthenes
I calculated the smallest prime divisor for each number in [1, 1e6] to be able to quickly factor any number in there. This precalculation is done in a sieve-like way. Then I go over all the elements of the array and factor it using the prime factorization computed using the sieve. Now I would have a map “cnt” from any factor in the range [1, 1e6] to how many elements in the array are divisible by that factor. Then, for some current i, we go over its factors and find the biggest one that has cnt >= k. But since this involves factorization, i guess the time is o(N*sqrt(A_i))
since the maximum number is only 1e6, you don't need map to store them, an array with size 1e6 is enough, since each operation for map is O(logn) with a not small constant, and also the TL in this problem is tight.
Breh I don’t understand why this always happens with these factoring problems. I always shoot just above the time limit because of some small BS detail lol.
probably you can check my solution
No. I solved ABCDF. I think E > F.
how to solve F?
Consider to offline. Use sgt to find LIS.
you can also use dp+binary search to find LIS my solution
We can classify ABC problems into two categories: 1. Standard / classical (a.k.a trash speedrun AI-friendly problems) 2. PHD level (you may encounter it once in your whole CP-career, not worth learning) In short, both are trash.
so why do you participate?
I didn't. I stopped participating in ABC ever since they remove the E-x problem. I just have a look at the last problem after the contest was done to see if there's any interesting thing to learn.
how to solve F?
O(nlogn) lis + dealing queries offline
Although I didn't like the contest much, I appreciate the effort on preparing the editorial for Problem G.
On a side note, is it a standard problem if we consider the total cost cannot exceed some non negative integer P (not sure about the bounds) and we can only choose some positive integer x in Problem G.
AI Better Contest, huh?