Heyo, for those who are reading this, I highly recommend you to participate in Codechef Starters 144. Even though I was never a big fan of CodeChef contests, I watched satyam343 pour his heart, sweat, soul and tears into the contest. Even if you don't usually do CodeChef, you should totally make an exception just this once!

Congrats on doing a problemset for every division!!!

Your contests are great :)

Solid problemset as a tester

:3

As a tester, satyam343 tyr0Whiz :orz:

Hopefully pupil this time!

Congrats sum & cry for becoming cf grandwriters ! ( wrote rounds for all divisions)

Another cry 's contest! I am glad to be a participant!

potato

hoping to have loads fun participating. The first div 3 that is not rated for me)

Wishing to become a specialist this time

As an out-of-competition participant, I wish I was tester.

18o3 Orz.

cry and sum contest are the best!!Always enjoyed their problems.

where is 74TrAkToR ?

Hoping to rise from the trauma of last div2.

I have small question, can we give more rated contest to people below 1899 ?

1. Div-2 is rated for 0 to 2099

2. div-3 is rated for 0 to 1599

3. div-4 is rated for 0 to 1399

technically, purple and violet can take part in same number of contest. I guess there is a small room for improvement in rated contests distributions.

GOOOOOOOOOOOOOOOOOOOOOOD LUUUUUUUUUUUUUUUUCK GYUS

Hoping to reach pupil this contest. Been practicing quite a few amount of hard problems for me. Hoping to not mess up like a dumb ass. Ideone f ed me up and leaked my solutions :(

As a participant, I want +ve delta :/

hope i dont bottle to get +8 :)

why writers of this contest ain't listed

1"

First contest as a pupil let's go :)

hope back to expert:)

This contest is the last stop on our mission to problemset for every division. Does this mean they sum and cry won't be writing more Div3s (Ó⁠╭⁠╮⁠Ò)

Sorry to say that Bangladeshi can't participate due to network issue in Bangladesh.

cowodeforces ?

Finally, a div 3 round :)

Why are Div 3 contests so rare?

cry sum!

i want to be tester

I've always wondered... Will there ever be a Div 3 round where it's not extended ICPC style? Not that I don't like the current format, but it would be very interesting...

Please give some easy and good questions :)

this time my target 50 plus delta

Inshallah going to be expert today

As a tester, I can confirm that the problems are high quality, and this is probably one of the best Div-3s I've ever seen. I highly recommend participation.

Cry orz always the best contests.

All the best!!! -Beijing·China

FINALLY. DIV 3.

GL evbdy !

bro, is the queue infinite?

slow judge

in queue

queueforces

queueforces :(

It's my honor to play the main role in problem A! Tks the problem setters!

Counting is Elation!

Should i move to next problem or wait for WA :(

My post contest discussion stream discussing all problems

Love references to Honkai Star Rail LMAOOO.

Dude the D problem is so tough, the constraint is so tough to deal with

E,F,G all had HSR references wow

countforces!

E was FUN!!!!!

But D was ####

Cheater top 30

queueforces

Is $$$O(n \sqrt n)$$$ not intended to pass in G? Or are my data structure implementation skills just trash? I had to do a ton of optimization to get it to pass.

Well now that the round is over... This was very sad.... TLE everywhere (and one MLE)... I couldn't get C to not TLE on case 6, and F I couldn't get past test case 4... Today felt like SegmentTrees and prefix sums, but clearly I'm just trash...

Anyone else had trouble taking inputs as numbers in B? Had to take each line in the matrix as a string.

I spent so much time to realize that I forgot to take modulo on E

Guys what was it??? a nightmare for me.

Fucking python, my $$$O(n \log n)$$$ with segtree for G fails by TL.

F is an excellent problem, kudos to authors

Hints for F : Bomb

• Solve the problem for $$$k \leq 10^6$$$. You'll realize that the greedy strategy of picking the largest $$$a[i]$$$ is correct. However, this greedy must be applied in a simulation, because the largest element will change at each step.

• Split each $$$a[i]$$$ into multiple elements, with values $$$a[i]$$$, $$$a[i] - b[i]$$$, $$$a[i] - 2b[i]$$$, etc. This is an Arithmetic Progression, so we won't store these values explicitly.

If you consider a number line as $$$[0, 10^9]$$$ (where each point denotes a term of the arithmetic progression , it's clear that you will always take a suffix). Therefore, you can do a Binary Search on Answer.

Assume that the smallest element that you picked is $$$> mid$$$. Then, for each $$$a[i]$$$, you can compute how many terms will it contribute to the answer, and what is the sum that it would contribute to the answer. If the count is $$$< k$$$, you update the result, and look in $$$[low, mid - 1]$$$. Else, you look in $$$[mid + 1, high]$$$. Keep in the mind that you don't necessarily need to take all terms == mid.

In problem D , can we apply binary search on all three dimensions simultaneously ??

Hey everyone, how did you approach problem 1996C - Sort? I initially had an O(nq) solution which was too slow, and ended up using prefix sums for an O(q) solution. I'm wondering what other approaches I could have taken, as my code was quite long so I wanted to know if there was an easier way of approaching the problem.

I desperately submit problem D in last 5 min thought it was gonna TLE but accepted. Feels very lucky tbh

Easiest solution for E — 272838782

doesn't judge ignore '\n' symbols between answers for every test in test case?

Here's my "wrong" solution for B — https://mirror.codeforces.com/contest/1996/submission/272741233 Here's fixed version, but I lost 10 minutes because of the slow queue — https://mirror.codeforces.com/contest/1996/submission/272756934

Also I tried that for C and it worked — https://mirror.codeforces.com/contest/1996/submission/272781207

Was e based on a prefix array with maps and time complexity O(n)?

How to E?

Nice contest, thanks to the authors. Can't believe I couldn't AK today :\

I was unable to submit it went into a continuous loop of cloudfare verifying you as a human

How to solve problem G?

.

I hate when testers lie

a lot of cheating going on in D and E i believe

Hello, for problem 1996D - Fun, how did you guys manage to optimize it? I only had time to find a mathematical formulation of the problem, resulting in the following python code:

import math

n = int(input())
x = int(input())

total = 0
for a in range(1, min(x-2, math.floor((n-1)/2))+1):
    for b in range(1, min(x-a-1, math.floor((n-a)/(a+1)))+1):
        total += min(x-a-b, math.floor((n-a*b)/(a+b)))
        
print(total)

I can mathematically deduce the result of min(x-a-b, math.floor((n-a*b)/(a+b))), the issue is that if the result is the latter i don't know how to optimize summing those values.

272848365 why my code wasn't working for problem c?

How to solve E

No hacking phase? edit : Oh, it just started.

is this the solution for Question E

https://pastebin.com/CYYQNq7h

please open upsolving

one day the earth will be not suitable for living, because temperature is increasing and people are greedily using resources. So why people fight for short term materialistic contest?

how this is TLE on problem E

    string s; cin >> s;
    int n = s.size();
 
    map<int, vector<int>> freq;
    freq[0].push_back(-1);
    
    int count = 0;
    modnum ans = 0;
    for (int i = 0; i < n; ++i) {
        count += (s[i] == '1');
        int diff = count - (i + 1 - count);
        for (auto x: freq[diff]) {
            ans += countValidPairs(x + 2, x + 1 + (i - x), n);
        }
        freq[diff].push_back(i);
    }
 
    cout << ans << "\n";

In queue

upsolving this one is gonna me mad, one of the best contests i stumbled across

Top $$$5$$$ in the official standings are all Vietnamese LOL.

Countingforces

First contest in almost 3 years. And the quality of problems is as good as ever!

So many cheaters lmfao, for problem D(maybe more), I can see lots of similar 3 liner submissions !!! Doesn't Codeforces have plagiarism check?