Codeforces the Saviour !!

lol the meme has more upvotes than the blog

why did you cut out the part where Codeforces suplexes us

What if it got wrong during system testing?

When you hack your own solution

Its raining contests.

Hope to be a good week for everyone and high rate <3

vovuh is ready with div3

why?

I hope some day people will stop writing stupid comments

All the people who have downvoted my comment, I WILL PROVE THEM WRONG!!

It's probably been said a 100 times before: Hope is a dangerous thing. Hope can drive a man insane.

I want the best for everyone and high rate for them , and they downvoted on my comment !!!!!!!!!!!!!!!!!!

There's no points for hacking in 12 hour hacking phase right? Although the person's solution you hacked would fall on the leaderboard. Am I right?

For the people who have the same number of problems solved the one with lesser penalty is higher up on the ranklist, penalty is the sum of the time at which each question is solved and also 10 minutes are added to the penalty for each wrong submission. Also the penalty for incorrect submissions is only added if you solve the question.

After the contest if you feel that a particular solution from a user has passed the system tests but you think that you can provide a valid test case which that solution might fail than you can try that and if that solution fails that particular user will lose the points gained on that question.

Sacred Games!!!

How do you know tests are poor?

In your level you can try to solve problem 'b' after the contest. And then when your can solve b problem , you can try 'c' ......

See 631 Div2 That was more harder than this one!

I thought it was super hard for a long time before I realized the explosions only hurt the monster in one direction

for D also

Its easy think :: suppose if we start killing from monster i , then cost of killing all monsters will be ai + (a[i+1]-b[i]) +(a[i+2]-b[i+1])+.... etc.;

so to solve it we can maintain an array d[i]= max(0,a[i]-b[i-1]) {dont forget for i==1} , we should also find sum of whole d[i] array as S;

through above info , now we can iterate over our array and check for every position as :

ans= min(ans,S-d[i]+a[i]); in O(N) time;

just above you

You can factorize $$$D$$$ in $$$\mathcal{O}(\sqrt{D})$$$ :)

nvm I misread your comment.

It says so right here in the rules:

It is forbidden to obfuscate the solution code as well as create obstacles for its reading and understanding. That is, it is forbidden to use any special techniques aimed at making the code difficult to read and understand the principle of its work.

(copied from http://codeforces.me/blog/entry/4088)

Obscured by using an algorithm template? :thinking:

Well if you got tle it means your program wasn't O(n)

iamxlr8 A with strong test case?

4 1 3 3 7 8 7 9 10 7 11

3 5 0 -2 5 3 6 7 8 2 4

It's not incredible, it's just obscured.

Each variable/function/class name is replaced with ReplacementFor_ (eg. the struct TREE becomes ReplacementFor_TREE). Numbers are decomposed in HEX+DEC-HEX form (eg. 0 becomes 0x113c+1395-0x16af). String/char are written with ASCII code (eg. "YES" becomes "\x59\x45\x53"). Moreover, every non-essential spacing, line-break or indentation is removed and then random line-breaks are put here and there in unusual places.

He/She has made a simple script that obscures the code before submission so that it's harder to copy/paste but otherwise exactly identical to the original one.

In the last contest, his codes for A, B, C were of quite different styles. I suppose he did this to reduce suspicions this time.

Can you please tell just basics..what he has written. I was unable to understand

see (https://en.wikipedia.org/wiki/Eulerian_path#Hierholzer's_algorithm)

You can start creating a path like

1 2 1 3 1 4...1 n 1

Then, insert after the n, but before the last 1

2 3 2 4 2 5...2 n

and again, like recursive.

The length of the path in every iteration is $$$2*(n-i)$$$ From that you can construct the interval (l,r)

Analyze the cycle for n = 4 and n = 5. Once you do that it's easy to see that the cycle will consist of n blocks, i'th of them will be of size 2 * (n — x) (except for i = n, in that case the block will be just 1) and will look like this : {i,i+1,i,i+2,i,i+3,...}. From that it's easy to restore answer for given l,r.

A general hint for any problem: If you don't know how to solve it, implement backtracking, maybe you notice a pattern.

Wow I was also able to find this but WA on pretest2

Are you sure you always select the optimal start? It's optimal to select the one which has maximum diff[i] — a[i] (or equivalently minimum a[i] — diff[i]).

It is $$$O(n)$$$ 76190446

Haha, I feel ya fam! Just keep going, we shall bleed red :)

same problem occured with me too my solution was of O(N) still got TLE

Try removing those print()s and submit 76193796. It turned to TLE.

1 2 100 50 101 99

• T = 1
• N = 2
• 1st pair : 4 1
• 2nd pair : 5 3

use scanf and printf instead of cin and cout. Or use faster input/output. Same thing happened with me also.

don't use endl, just use "\n"

http://www.cplusplus.com/reference/ostream/endl/

I think you are facing the same issue as I have mentioned here.

Codeforces reply to my query regarding the same issue: Unfortunately, this is the way things are on most programming contests: it is not guaranteed that each problem can be solved in every language

maybe you are using endl instead of \n

I submitted an O(n) solution on Python, it got TLE on test case 3: https://codeforces.me/contest/1334/submission/76161837

I rewrote that solution to C++, it got TLE on test case 3: https://codeforces.me/contest/1334/submission/76185611

I inserted the magic string "struct {(){ios_base::Init i;ios_base::sync_with_stdio(0);cin.tie(0);}}__;", it got accepted :https://codeforces.me/contest/1334/submission/76187533

you can use bitset :D

You should take subsegment from l to r(inclusive) from sequence of the cycle(1,2,1,3,2,3,1)

try fast io .

ios::sync_with_stdio(0) ; cin.tie(0) ; cout.tie(0) ;

+1, WTF? I submitted a solution that failed on $$$n = 1$$$ and spent 12 minutes submitting a fix (because of something unrelated to the test case itself). Now I tried to hack myself and apparently I shouldn't have even bothered, because the test is not allowed?

Unfortunately, this is true. We excluded the tests with $$$n = 1$$$ just before the contest to exclude unnecessary casework, but we forgot to update the statement :(

We are sorry for that.

Great tutorials man. I like your style! I will watch if you have D-F tutorials :D

the question d was easy and i found the logic but I didnt have enough time to write code cz i spend so much time on C, I got frustrated as it was actually the first time I got the logic for c level question in div 2.

(For E) The number of paths is therefore t!/p1!p2!...pn!

I didn't know there was a formula for that. I spent a lot of time write dynamic programming for calculating that, which in the end turned out to be wrong anyway.