We will hold AtCoder Beginner Contest 469.
- Contest URL: https://atcoder.jp/contests/abc469
- Start Time: http://www.timeanddate.com/worldclock/fixedtime.html?iso=20260801T2100&p1=248
- Duration: 100 minutes
- Writer: vwxyz0, cn449
- Tester: sounansya, sheyasutaka, harurun4635
- Rated range: ~ 1999
- The point values: 100-200-300-400-475-500-575
We are looking forward to your participation!








I can't participate in this contest. I registered for CF Round 1113, which, after all, was the key competition for the debut to increase the rating.
Note that the time of abc and Round 1113 doesn't conflict.
I think it'll not because the difference between the timings of ABC contest and Round 1113 CF is smth like ~2.5hrs (i think so)
Oh, the start time of Round 1113 has been adjusted from 19:35 to 22:35. I didn't see this news.
No 600+ G in ABCs! I believe it is good to keep the high quality problem like 2300+ rated problem G for ARCs and AGCs and make ABCs problemset within the upper bound of 2100-2200.
To avoid the contest being a typing speed competition, we can have problems C-F on a harder side.
Does Problems D and E ought to be this easy, especially compared to problem F?
A great contest. The only caveat is to reduce the jump between D and E; but even that is acceptable since it's subjective to have such gaps between any two problems.
This was just a constructive feedback from my end to the AtCoder officials. I still find them to be high quality rounds, but this is one key area that can be worked upon based on recent rounds. Earlier, even this part was well handled.
Rated register is the worst choice I've ever made.
I think so. I chose Unrated register.
WTF, I lost rank but gained rating.
Can anyone explain your approach for the task C and D. Thanks!
for D: https://pastebin.com/eiPZMTX3
F is ACM ECPC 2017 D
Solution Code (available before start): https://github.com/AhmedOS/Problem-Solving/blob/master/ACM-ECPC-2017/Solutions/D-Dream-Team.cpp
you sent this the moment the contest ended huh
LMAO, bro was waiting for contest to end so he can send it
F = https://codeforces.me/problemset/gymProblem/101856/D
vwxyz0 cn449
Problem statements were really bad, but good problems.
F is a DSU. G is Matrix Power
Hello Atcoder! I can hack my own solution of D using the following testcase:
3 21 22 3Whose answer should be 3 but my output was 2. Would you rejudge problem D?
how to solve D ?
I was thinking I need 2 loops 1. {x,minn} 2. {x,maxx}
where minn and maxx are from last M'th round.
but then for these pairs i need to know if the pair is getting partially overlapped with all M pairs. (this I couldn't think of how to do fast) with lesser complexity
Let $$$f_1[i]$$$ denote the total number of finals played by player $$$i$$$, which is initialized by counting the occurrences of each player in the input pairs. For a pair of players $$$(x, y)$$$, the number of distinct finals covered by them is given by:
where
common(x,y)represents the number of times playersx and y have played together in a final. A pair of players is valid if and only if they cover at leastmdistinct finals:where
mis the total number of finals.Approach
We process players in increasing order of their IDs from $1$ to $$$n$$$, maintaining an ordered set (
PBDS) of pairs(f2[player], player)for all previously processed players.For each player $$$y$$$:
Adjust for shared finals: If player $$$y$$$ has played a final together with player $$$x$$$ (where $$$x \gt y$$$, since we store the larger index as the first component and push $$$y$$$ to $$$G[x]$$$, or vice-versa based on
if (x < y) swap(x, y)), their combined count needs to subtract the shared finals. Since we maintain $$$f_2$$$ arrays, we temporarily decrement $$$f_2[x]$$$ in the ordered set for all $$$x \in G[y]$$$.Count valid pairs: We then query the ordered set to find how many previously processed players satisfy the condition:
using:
int C = ((int)(os.size())) - os.order_of_key({m - f1[y], -1});thank you! for explanation. a bit ugly — but this doesn't use pbds.
https://atcoder.jp/contests/abc469/submissions/78059490
I will try with pbds as well :D
you can use treaps as well
(https://atcoder.jp/contests/abc469/submissions/78039175)
Problem statements are really horrible.
After the contest I submitted a linear time solution to E that gets AC but the code is wrong, it gives wrong answer for a test case I found with brute force. I think there are people that submitted code with the same wrong idea in the contest. Will the solutions be rejudged?
How to solve problem E?
Binary search on the answer.
If you are checking whether v is able,you can set $$$o=1$$$ and $$$x=\frac{-v}{1-v}$$$,then check there's a non-negative subarray with k 'o's or not.
will always choosing the smallest range that has exactly k wins result in the highest win rate?
ok never mind i got it
why the fuck we need to ./Main.cpp: In function 'int main()': ./Main.cpp:25:13: warning: comparison of integer expressions of different signedness: 'int' and 'std::vector::size_type' {aka 'long unsigned int'} [-Wsign-compare] 25 | if(i<ans.size()){cout<<ans[i]<<endl;}else{
man this much is too much like first think logic then this shit i did love this they should remove this things
maybe try ssize()?
I can't participate in this contest because of CloudFlare.
how can i optimise problem c ?