Meta Hacker Cup 2024

5 months ago, # ^ |

Nice catch, thanks!

→ Reply

ankancool

5 months ago, # |

+12

It's happening!! Super excited!

→ Reply

SpeedyAlgorithmist

5 months ago, # |

yay , super excited !!!!

→ Reply

tickbird

5 months ago, # |

how hard to be ranked in 2000 in Meta Hacker Cup compare to Div1+2 round in cf?

I'm genuinely curious

→ Reply

_inferno_

4 months ago, # ^ |

same question here lol

→ Reply

Raj_Tiwari

4 months ago, # ^ |

I guess around 4-5 question in div 1+2

→ Reply

Murad134

5 months ago, # |

Yayy.... it's happening!!! very excited to participate for the first time.

→ Reply

formidablechief27

5 months ago, # |

why doesnt hacker cup use simple testing like codeforces does. i had a really hard time last season

→ Reply

Avik_AB17

5 months ago, # |

+23

Counting down the days to Meta Hacker Cup 2024. Long live Hacker Cup!

→ Reply

Swistakk

5 months ago, # |

-18

Too bad Codeforces doesn't have "haha" react

→ Reply

5 months ago, # ^ |

+13

Because the ai submissions will fall on their face?

→ Reply

papa-ka-para

3 months ago, # ^ |

Good luck to you @Swistakk . Last time you missed final round by just 44 seconds. This time, I am rooting for you to reach Final <3 .

→ Reply

Swistakk

3 months ago, # ^ |

Thank you! Recalling that I was so close last year actually gives me some hope xD

→ Reply

_harveyspecter

5 months ago, # |

nice, i am excited! i hope ill get t-shirt :)

→ Reply

5 months ago, # |

-6

Same thing every year. Make promises to change submission format at the end of the year, and keep it same the next year :clown:

→ Reply

https://codeforces.net/blog/entry/119880?#comment-1063469

5 months ago, # ^ |

-54

Real sad Google stopped organizing. They were the only ones who did annual programming contests right.

→ Reply

Dominater069

5 months ago, # ^ |

Where did they promise?

→ Reply

4 months ago, # ^ |

-10

Maybe 'promise' wasn't the right word, but I've been here long enough to recall SecondThread talking about updating the format. Even the comment you linked says they are looking into it. It used to be a fine format when there were so many annual contests, and it diversified the submission formats a bit. Now, I think it may be time to consider updating it for real. The only reason I can think of for not doing so is that they actually haven't put in any effort. Anyway, good luck everyone, may the stack limit be with you.

→ Reply

SuperJ6

4 months ago, # ^ |

+11

Perhaps they looked into it, decided it wasn't right for whatever reason, and moved on. Why do people want to make contests as same and boring as possible?

→ Reply

4 months ago, # ^ |

By boring you're talking about a submission format that reduces stress for participants? I'm sure it even reduces stress for them cos they won't have to answer hundreds of clarifications.

→ Reply

Red0

4 months ago, # ^ |

-10

Or maybe get an actual IDE? Cause I bet that all of the top coders have their own IDEs which can compile the large testcase in literally ≤ 1 min.

→ Reply

4 months ago, # ^ |

I fail to see how an IDE helps here

→ Reply

Red0

4 months ago, # ^ |

well the only reason it's stressful would be if you test and your compiler can't process the testcase. And that can be solved using an IDE with the proper configurations. Otherwise the time limit is definitely sufficient.

→ Reply

4 months ago, # ^ |

How does a compiler process a testcase? You should read about IDEs, text editors, and compilers to understand their functions. You're mixing stuff up. Even if you're using the fastest IDE (whatever that is) how fast the program executes is still dependent on your computer specs.

→ Reply

caustique

6 weeks ago, # ^ |

Alright, then what's your problem with the current format? If you are able to compile a program on your own computer, download an arbitrarily large file, run your program with that input file and then upload the output — then you're all set to compete in the tournament.

→ Reply

SpeedyAlgorithmist

5 months ago, # |

← Rev. 3 →

-20

nvm

→ Reply

cpp_enthusiast

5 months ago, # |

+46

I wanted to point out that the date of the finals coincides with the date of the Putnam collegiate math competition. It is not uncommon for people who are interested in competitive programming to also take part in such math competitions. Perhaps this could be looked at.

→ Reply

Yugandhar_Master

5 months ago, # |

Sir, is the submission system same like last time? If yes, please change it to like some online judges or like codeforces.

→ Reply

5 months ago, # |

In it, instead of solving problems manually, contestants will create an autonomous code generation system before the start of the contest.

and

Each contestant can compete either in the human track or the AI track, but not both.

Why are we not allowing a contestant participate in both the human track and the AI track?

→ Reply

bicsi

4 months ago, # ^ |

+21

Does it even matter? Nobody is going to choose the AI path anyways (I have only 3 candidates in mind and they are all companies)

→ Reply

4 months ago, # ^ |

Maybe because the problem statements are same?

→ Reply

4 months ago, # ^ |

Yep, this is why

→ Reply

Sreedhar3210

5 months ago, # |

+13

How will the T-Shirt be looking this year

→ Reply

formidablechief27

4 months ago, # |

-10

why dont u post youtube streams nowadays

love your channel

→ Reply

accidentallygivenfuck

4 months ago, # |

Instead of having a timer for test data download, will AI participants have a timer for problem statement download? :pepesmug:

→ Reply

4 months ago, # ^ |

+49

Yep. And your timer will start when the contest starts. So it’ll be like a 6 minute contest basically.

→ Reply

notdevd

4 months ago, # |

Finally!!

→ Reply

chuka231

4 months ago, # |

Is the judge system same as before like locally running and submit results?

→ Reply

4 months ago, # ^ |

Yes

→ Reply

chakred_namor

4 months ago, # |

+11

Hi everyone. I have a question: Will they deliver T-shirts to Ukraine?

→ Reply

I_love_Leyeli_Meredova

4 months ago, # ^ |

Hi! The T-shirts are delivered by FedEx

→ Reply

4 months ago, # ^ |

If our delivery partners ship to your region, then we will. I recall last year Ukraine was one of the regions that FedEx didn't ship to. I'm not sure if this has changed since then.

→ Reply

lperovskaya

4 months ago, # |

+16

Yes yes yes yes let's goooooooo

→ Reply

jrsak_vrmr

4 months ago, # |

letss gooo

→ Reply

leonlian

4 months ago, # |

LETS GO HACKER CUP IS BACK!!!

→ Reply

yoco

4 months ago, # |

is it happening?

→ Reply

maya.ah.afifi

4 months ago, # |

Is the only on-site round is the last? I also wanna know what is the difference between the AI and human versions?

→ Reply

4 months ago, # ^ |

The difference is written in the blog

→ Reply

14_15_16

4 months ago, # |

Excited

→ Reply

Alamin2000

4 months ago, # |

What is the last date of registration?

→ Reply

divyalaldinani

4 months ago, # |

as its written, today is the day when registration starts, but its showing page is unavailable.

→ Reply

4 months ago, # ^ |

If you weren't signed into facebook before, we were accidentally preventing you from seeing the page before because your age verification was failing. It's fixed now.

→ Reply

KrIsN123

4 months ago, # |

Why does the page show unavailable? The registration starts today right?

→ Reply

4 months ago, # ^ |

It should be available now; please try again.

→ Reply

pathetique

4 months ago, # |

Was there always an age limit?

→ Reply

4 months ago, # ^ |

Yes, it is and has been in our terms of service

→ Reply

Muhammad_Aneeq

4 months ago, # ^ |

didn't seemed to be a problem last year.

→ Reply

nor

4 months ago, # |

+21

How comprehensive will the test cases be (if there will be any)? From what I have seen, test cases tend to make or break the solution search strategy in attempts at autonomous code generation like AlphaCode, for example.

→ Reply

4 months ago, # ^ |

Do you mean in the full data? We'll endeavor to make them comprehensive as we usually do, but there's always the general possibility that there's some weird edge case that a solution we didn't consider fails. You can see examples of the kinds of test cases you can expect in the previous years' data.

→ Reply

Aniket19o7

4 months ago, # |

← Rev. 2 →

Is anyone facing problem in signing up for Facebook?

→ Reply

divyalaldinani

4 months ago, # ^ |

yeah me

→ Reply

Nigar92

4 months ago, # |

+10

Will the finals be online?

→ Reply

4 months ago, # ^ |

Yes

→ Reply

4 months ago, # |

In the human track rules, unless I missed something, it is not prohibited to use directly AI, or something like in the AtCoder wording ("It is prohibited to directly input all or part of the information issued as problems into software"). Not sure it would be useful (yet...), but anyway does that mean it might be allowed?

→ Reply

4 months ago, # ^ |

Oh TIL https://atcoder.jp/posts/1247

→ Reply

NinjaSenpai

4 months ago, # |

I have signed up. Can I change the size of my T-shirt ? If yes how ?

→ Reply

4 months ago, # ^ |

You’ll enter your shirt size when it’s about to be shipped to you later in the contest season

→ Reply

aayush249

4 months ago, # |

Why is it so hard to create a new Facebook account?

After uploading a verification photograph, an appeal is automatically sent, and my account is disabled for no reason whatsoever within the next 24 hours.

→ Reply

puupil

4 months ago, # |

can someone tell where can I practice past problems and submit them?

→ Reply

SkibidiPilav

4 months ago, # ^ |

Here, you can select any of the past seasons, and any of the rounds and participate. You do have to be logged in to submit though.

→ Reply

SkibidiPilav

4 months ago, # |

Hi, what is the minimum age requirement?

→ Reply

4 months ago, # ^ |

18 years old

→ Reply

pathetique

4 months ago, # ^ |

any rationale for this, or any chance this could be changed?

→ Reply

3 months ago, # ^ |

+10

There's lots of scrutiny about showing the names of minors who competed on a publicly available scoreboard, and Meta didn't want to open themselves up to needing to defend doing so either in court or in public opinion. For what it's worth, this has always been in our Terms of Service.

→ Reply

pathetique

3 months ago, # ^ |

Thanks for the clarification! I guess I've never read the ToS carefully enough in the past.

→ Reply

3 months ago, # ^ |

-8

So it's ok for minors to have a Facebook account and be shown to the world but it's not ok to appear on a competition scoreboard?

→ Reply

3 months ago, # ^ |

You can tweak privacy settings when you create an account. You can’t with a leaderboard.

→ Reply

3 months ago, # ^ |

But you can't hide your name

→ Reply

3 months ago, # ^ |

You literally can

→ Reply

3 months ago, # ^ |

I didn't know that. But still meta can add option to hide name from standings

→ Reply

3 months ago, # ^ |

I guess they could fair enough

→ Reply

Beatrix-0

3 months ago, # |

What's the difference between human and ai track?

→ Reply

3 months ago, # ^ |

-8

In the human track, you write code to solve the problems yourself.

In the AI track, you write a code generation system before the contest to solve the problems itself.

→ Reply

Ahamed_TJ

3 months ago, # |

I'm so excited about this year's contest!! Thanks for the post!!

→ Reply

Amit9

3 months ago, # |

Was registration opened for only one day?

→ Reply

3 months ago, # ^ |

It's still open. You just need to sign into your facebook account to be able to register, as it says on the site.

→ Reply

chiranjivikeshav

3 months ago, # ^ |

For human track, will there be only competitive programming type questions or other?

→ Reply

3 months ago, # ^ |

All tracks will have questions similar to the kinds of things you'd see in competitive programming contests. You can see our questions from previous years as an example of what you can expect.

→ Reply

Monis_Azarus

2 months ago, # ^ |

Is it still open now or has it closed? I can't see any registration button for round 1, for practice round it is still open though...

→ Reply

codevengers

2 months ago, # ^ |

I tried creating an account, but it got disabled citing it doesn't follow community standards, and that I can't request a review of the decision. So, is there anyway you can help with it?

→ Reply

spectacle_

3 months ago, # |

Finally, a chance to get t-shirt :))

Excited !

→ Reply

Monis_Azarus

2 months ago, # |

Can anyone please tell me if the registration for meta hacker cup round 1 has started or is it already finished, I am able to register for practice round but there is no button to register for round 1

→ Reply

2 months ago, # ^ |

It has started and hasn't finished yet. You must log into Facebook to be able to register.

→ Reply

harshsingh2

2 months ago, # |

What Kind of questions can i expect in Round 1 ?

→ Reply

2 months ago, # ^ |

You can look at our questions from previous years to get an idea of what types of questions will appear in each round

→ Reply

Julea

2 months ago, # |

I am unable to register for round 1

→ Reply

2 months ago, # ^ |

registration has not started yet i think for main rounds!

→ Reply

2 months ago, # ^ |

If you register for the practice round, we'll automatically register you for Round 1 afterwards.

→ Reply

TCL

2 months ago, # |

I cant update Email Address. I already updated my email at facebook account, but hacker cup site dosent get my email and I cant even write my email.

→ Reply

https://codeforces.net/blog/entry/133675

2 months ago, # ^ |

← Rev. 3 →

I just looked into the code that shows emails on the profile page and the Hacker Cup code doesn't even store your email. We're just reading from your Facebook user's primary email. So technically it seems like is/should be impossible for these to get out of sync with your Facebook email.

Can you please send me a direct message explaining why you think these are out of sync and what you did to update your facebook account's primary email?

→ Reply

TCL

2 months ago, # ^ |

I dont know how to put picture in comment, so I write at blog

→ Reply

surgutti

2 months ago, # |

i bet that problemset will have over 100k words

→ Reply

HVCYM

2 months ago, # |

I just opened hacker cup website and got to know that I finished in top 2000 in round 2 last season. I missed the t-shirt claim process. Can I claim now?? Hopefully not!! :)

→ Reply

2 months ago, # |

← Rev. 3 →

I have my email linked to my Facebook account, but it still doesn't appear on my profile.Could you please advise on what to do now ?

→ Reply

2 months ago, # ^ |

mine has not updated also ! Did your email updated ?

→ Reply

2 months ago, # ^ |

It seems like the issue hasn't been resolved yet. As SecondThread mentioned earlier, they get the email from Facebook’s primary email. However, I'm still unsure what the exact problem is. My email is public on Facebook, so there shouldn't be any issues with visibility.

→ Reply

2 months ago, # ^ |

i updated my email just 1 hour ago! Is it necessary to have an email to get a t-shirt ?

→ Reply

2 months ago, # ^ |

I guess the only way for communication regarding prizes is email.

→ Reply

2 months ago, # ^ |

Bro,if your problem get fixed then plz inform me also!

→ Reply

2 months ago, # ^ |

The contact info we use here is your primary facebook email. The easiest way to update this is in the facebook app if you go to Menu > Settings & Privacy > Settings > Search for "email" > Email, and then updating your primary email.

→ Reply

Ghulam_Junaid

2 months ago, # ^ |

I have clicked on Email, but I don't see any option to update the primary email. All there is, is some notification settings.

→ Reply

Ghulam_Junaid

7 weeks ago, # ^ |

can you update "Edit this in your Facebook settings" to the correct place?

→ Reply

6 weeks ago, # ^ |

← Rev. 4 →

Open Facebook mobile app (not Facebook lite), and then

1) click on Menu (3 horizontal lines)

2) Settings & Privacy

3) Settings

4) Search for "email"

5) Email

6) Primary email

7) Enter the email you want to add, if you get an error then try adding other email.

8) You get an OTP on mail , Enter it

This worked for me !

→ Reply

puspendra_09

6 weeks ago, # ^ |

Every time it is saying wrong email , even after entering 3 different mails.

→ Reply

gouravg

2 months ago, # |

I have successfully registered for the practice round, but when I attempt to access Round 1, it shows on the left side that I have not registered for this round. Could you please tell me with the registration process for Round 1?

→ Reply

sahaun

2 months ago, # ^ |

You will be registered automatically if you registered for the practice round.

→ Reply

bansalarpan4

2 months ago, # |

Can I register for it now ? Or am I late

→ Reply

2 months ago, # ^ |

You can still register.

→ Reply

bansalarpan4

2 months ago, # ^ |

← Rev. 2 →

I have solved 2 problems in the qualification round in Human track but I don't know if they are accepted,they are just showing a question mark there

How can I register, can you please guide me ?

→ Reply

7 weeks ago, # ^ |

During the contest, you won't know whether or not your submissions are correct. Now that the contest is over, you should be able to see the result.

→ Reply

2 months ago, # |

← Rev. 2 →

What are the rules about AI for the human track? Can I ask for hints, solution, boilerplate code, debugging, refactoring, etc even though I’m participating in the human track? If not, how would that be enforced? Thank you.

→ Reply

2 months ago, # ^ |

+10

The FAQ has been updated — https://www.facebook.com/codingcompetitions/hacker-cup/2024/round-1/faq

Human Track — Can I work with other people / AIs, or use pre-existing code?
You may use any code or online information that had been written before the start of the contest. However, you may not communicate with anybody else about the contest, including non-competitors, while it is running. You may not use any AI system for assistance during the contest.
If you submit the same code as somebody else, or if the Hacker Cup team in any other way has reason to believe that you have communicated with another competitor, you will be disqualified.

Yes, AI includes o1

→ Reply

2 months ago, # ^ |

Thanks for the answer!

→ Reply

asiad

2 months ago, # |

← Rev. 2 →

+14

MHC Round 1 conflicts with the 2024 ICPC North America Qualifier (NAQ) which happens at 11:00 — 16:00 PT on October 5th. Considering many universities in NA use NAQ for team selection for ICPC Regionals, will there be another bye system setup like there was last year? Not sure how this would work since the only round before round 1 is the qualification round.

And alternatively, can future MHC rounds happen on Sundays since it seems like there will always be ICPC conflicts on Saturdays in October/November every year.

→ Reply

cry

2 months ago, # ^ |

→ Reply

2 months ago, # ^ |

+13

We won't be offering alternatives to Round 1 this year. We hope to see you on the scoreboard; of note, you could likely use the first hour of the contest to qualify for round 2 even if you're participating in NAQ.

→ Reply

asiad

6 weeks ago, # ^ |

SecondThread since you are not offering alternatives to round 1 this year, can you at least change the requirement for qualifying to round 2 to be a points lower bound (like it has been for nearly 10 years)? I know last year the requirement was originally top 5000, but changed to 4 points due to system issues. But considering the NAQ conflict, 1 hour may not be enough to solve enough problems to secure top 5000 (especially if there are system issues in the first hour).

→ Reply

6 weeks ago, # ^ |

Nope, sorry. I believe in you though!

→ Reply

2 months ago, # |

← Rev. 9 →

Hi, SecondThread. I am not able to submit for the older contests that I participated in. As opposed to a Submit for Practice button for the contests I didn't touch, I see a Submit on Home Page button now for the contests I've submitted once for, but there is no Submit button on home page.

Can you help please? Here are pictures. (I'm unrated, so I can only link them.)

UPDATE: They fixed it.

→ Reply

dio_2

2 months ago, # ^ |

For me it has an option to "Download encrypted problems and input" but when I press it shows an error.

→ Reply

2 months ago, # ^ |

← Rev. 5 →

I'm seeing the same now. It wasn't there last night. Anyways, it's still not working for some older contests.

UPDATE: They fixed it.

→ Reply

2 months ago, # ^ |

Yep, should be fixed now. Sorry for breaking it!

→ Reply

2 months ago, # ^ |

No problem, thank you for fixing it :)

→ Reply

LLG

2 months ago, # |

There is no countdown for the timer before it expires, How much time we have to run the solutions ?

→ Reply

dio_2

2 months ago, # ^ |

In human track we have 6 minutes. Below the validate and submit button it shows the time remaining.

→ Reply

algoplutus

2 months ago, # |

I hope I get a t-shirt this time!!

→ Reply

BayernForce

2 months ago, # |

The input files in the HackerCup are quite large in size. VSCode crashes if I run my source code with the large input file and there is no output. Does anyone know how to resolve this ? Also tried using Sublime but didn't help much.

→ Reply

DC_H

2 months ago, # ^ |

Consider increasing your stack size https://nor-blog.codeberg.page/posts/2023-09-25-increase-stack-size/

→ Reply

jay_kay

2 months ago, # |

← Rev. 2 →

SecondThread Do the problems in the Practice Round have a TIME LIMIT? I couldn't find any info related to this on the contest platform and after submitting the solution to a problem, we can't see the judgement of our submission. If yes, how to know if it is 1 sec or 2 sec or 4 sec?

→ Reply

sahaun

2 months ago, # ^ |

There is no official TL because the solutions are not run by the system.

You just need to submit before the 6-minute timer expires. Considering you need a minute to download the input and upload the output, the TL is somewhere around 300 seconds.

→ Reply

2 months ago, # ^ |

If it's helpful, all of our judge solutions run in a few seconds to all problems on a modest laptop

→ Reply

2 months ago, # ^ |

But is it ok to submit a solution that takes a minute for example?

→ Reply

2 months ago, # ^ |

Yeah, if it takes less than like 5 minutes, there's no way we're going to DQ you.

→ Reply

Misa-Misa

2 months ago, # |

How does the round system work? Is each round independent and just for the qualification of next round. Or There is some accumulation of results of each round (like prefix sum) that is taken into account for next round?

→ Reply

sahaun

2 months ago, # ^ |

+11

Each round is independent, as mentioned in the contest page.

Anybody can also enter Round 1. If you place in the top 5,000 participants in Round 1, with ties broken by penalty time, you will advance to Round 2.
In Round 2, the top 500 competitors will advance to Round 3.
In Round 3, the top 25 competitors will advance to the Final Round.
The winner of the Final Round will be the 2024 Hacker Cup champion!

→ Reply

AshrafSustS19

2 months ago, # |

-11

Why do I have to go through such a long and convoluted process to just upload a submission? This is annoying

→ Reply

Post contest discuss stream for practice round

2 months ago, # |

→ Reply

2 months ago, # |

Btw how is checker of Fall in Line implemented?

→ Reply

2 months ago, # ^ |

+14

It's less exciting than you'd think. When we generated the solutions, we naively found the best lines and counted how many points were on it. The checker is given the optimal answer, so it doesn't have to calculate it on the fly, it just checks if you're within x and 2x the number it is given.

→ Reply

2 months ago, # |

← Rev. 2 →

In Problem 3 Fall in Line (very nice), luckily for me I think the trivial case with 3 ants not in line was not in the test cases. My code output is 3 (n instead of the smarter n-2) — correct answers are 1 or 2. I realized after submitting and I was almost sure it would have ended up in WA. This year favourable test cases :D

→ Reply

A_Le_K

2 months ago, # ^ |

← Rev. 4 →

In this case any pair of 2 distinct points $$$P_i, P_j$$$ is a "witness" for a line with $$$p==2>=3/2==n/2$$$ points lying on that line. Maybe You've just rounded $$$n/2$$$ incorrectly while doing the comparison $$$p ?>=? n/2$$$.

Anyways, problem C was a real "little delight"! ^_^

→ Reply

2 months ago, # ^ |

← Rev. 2 →

Getting random points 3 by 3 to find clues for possible lines with >n/2 points. As there aren't 3 points aligned, it outputs n. It always works, except with n=3. Outputing n-2 it would have always worked.

And yes, C and D2.

→ Reply

2 months ago, # ^ |

Happy to hear you liked it!

→ Reply

shadow9236

2 months ago, # |

I am surprised that intended solution for D2 is AVL/Splay/Treap instead of this geniosity

Solution

void testcase(){
	int n,targ;
	cin >> n >> targ;
	vector<int> a(n);
	for(int i = 0; i < n; i++){
		cin >> a[i];
	}
	
	for(int i = 0; i < n; i++){
		a[i] -= n - i - 1;
	}
	
	sort(a.rbegin(),a.rend());
	
	for(int i = 0; i < n; i++){
		a[i] += n - i - 1;
	}
	
	int sml = 1e9 + 60,who = -1;
	
	for(int i = 0; i < n; i++){
		if(abs(a[i] - targ) < sml){
			sml = abs(a[i] - targ);
			who = i + 1;
		}
	}
	
	cout << who << ' ' << sml << '\n';
}

→ Reply

jtrh

2 months ago, # ^ |

Mine is very similar (I do not know splay/treap):

code

struct solution {
    solution(int tc) {
        int n, g; cin >> n >> g;
        vector<int> e(n);
        for (int& x : e) cin >> x;
        for (int i = 0; i < n; ++i) {
            e[i] += i;
        }
        sort(e.begin(), e.end());
        pair<int, int> ans(inf, inf);
        for (int i = 0; i < n; ++i) {
            ans = min(ans, make_pair(abs(e[i] - n + i + 1 - g), n - i));
        }
        cout << "Case #" << tc << ": " << ans.second << " " << ans.first << "\n";
    }
};

→ Reply

2 months ago, # ^ |

Can you please explain the idea behind your D2 solution? Thanks.

→ Reply

jtrh

2 months ago, # ^ |

← Rev. 4 →

Assuming you already solved D1 and know that we only need to find the final positions of the stones:

Each stone we throw will either end up or end up pushing another stone to the $$$E_i$$$ th empty space.
Every stone that was pushed leaves the stone that pushed it behind one position from where it was.
This is better visualized as filling up the $$$E_i$$$ th empty space and shifting the positions before it backwards once (this is where splay/treap comes in).
Instead of thinking about positions we think about the empty space between them. Let $$$S_i$$$ be the empty space between the $$$i-1$$$ th and $$$i$$$ th rocks in sorted order with $$$S_0$$$ being the empty space between $$$0$$$ (exclusive) and the $$$0$$$ th stone.
Now watch what happens when we throw a stone:
--X-XX----X we throw a stone with $$$E_i=5$$$.
-X-XX--X--X
- The empty space inserted into splits into 2 because of the shift.
- $$$S_0$$$ is decremented.
- All other $$$S_i$$$ do not change.
This means a new prefix sum $$$S_0+S_1+...+S_i$$$ is inserted with all prefix sums being decremented. We can handle the decrement of all prefix sums using an offset however notice that this offset is just $$$i+1$$$ since it is incremented once each time. Now the actual empty space before the $$$i$$$ th stone is $$$S_0+S_1+...+S_i-(i+1)$$$.
Instead of storing $$$S_i$$$ we store its prefix sums $$$P_i$$$. From the first observation we know the prefix sum to be inserted is $$$E_i-1$$$, but we need to account for the offset $$$P_i-(i+1)=E_i-1$$$. So we sort $$$P_i=E_i+i$$$.
We can now calculate the actual position of the $$$i$$$ th stone by adding the empty space before it $$$P_i-N$$$ (subtract the final offset $$$N$$$) with the number of stones before and including it $$$i+1$$$.

→ Reply

2 months ago, # ^ |

Can you please explain your D2 solution?

→ Reply

abous

2 months ago, # ^ |

Wow that's brilliant. Wondering how this works though, would love to know the intuition behind it

→ Reply

2 months ago, # |

hey SecondThread , in the practice round my ranking in my country is not showing but it is showing in global , in my profile and in the certificate . what is wrong ?

→ Reply

ojojo

2 months ago, # |

Will every round of the Hacker Cup have a fixed 3-hour time slot, or will a 24-hour window be provided in which I can attempt the round for 3 hours?
Are the test cases given during the 6-minute duration only sample test cases? If not (which I believe is the case), what are the time and memory constraints for each problem?
Which version of the g++ compiler is used to check our solutions?
What is the qualifying criteria to advance to Round 2?

If anyone knows the answers to these queries, please comment.

→ Reply

2 months ago, # ^ |

← Rev. 3 →

Why does it matter how they check our solutions, if your output.txt is correct, it’s correct no matter how they check it.

→ Reply

ojojo

2 months ago, # ^ |

Please answer other questions too

→ Reply

h_squared

2 months ago, # ^ |

There will be a fixed 3-hour time slot.
No, they are the actual test cases. I think there is no strict time and memory requirement for hackercup problems. If you can run the test cases and submit within 6 minutes its ok.
I think code is only used for plagiarism and it is not actually run. Maybe someone else can confirm this.

→ Reply

https://www.facebook.com/codingcompetitions/hacker-cup/2024/practice-round

7 weeks ago, # |

In this year's Practice Round, nobody practiced harder than Neal Wu. He was the first to solve each problem, at 3, 7, 12, 18 and 37 minutes respectively.

→ Reply

tyristik

7 weeks ago, # ^ |

-17

That is the only round where he could prevail, LOL

→ Reply

Fanarill

6 weeks ago, # |

+16

SecondThread, hi, sorry for bothering you, but I need help.

I ran into some problems registering facebook account. Intially I had one account and now it's perma-banned. Don't know the reasons, yet I participated with it previous year. Decided to register new and got insta-disabled. I didn't worry intially, I just sent appeal.

Today, they reviewed my appeal (which is just photo of me) and still decided to perma-ban me.

The question is how to register facebook account without getting banned. Or how do I restore my previous accounts, if that's possible.

I hope it's not too late, thanks for help in advance

→ Reply

kyooz

6 weeks ago, # ^ |

+16

same problem for me :(

→ Reply

Fanarill

6 weeks ago, # ^ |

← Rev. 2 →

I managed to get around that disabling BS. Both of my previous accounts were registered just using an email. And I decided to register a third account, but this time I did things a bit differently:

Registered from phone facebook app.
Used phone number instead of just an email.

I think the second step is key, though I'm not 100% sure if it's sufficient on its own. Try it at your own risk, lol.

→ Reply

kyooz

6 weeks ago, # ^ |

yeah worked for me aswell, unfortunate situation

→ Reply

MathModel

6 weeks ago, # |

SecondThread my FB account is linked with my real age 17 , and I'm not allowed to participate

Any solutions or T_T

→ Reply

6 weeks ago, # ^ |

Unfortunately you can't participate until you're 18. Lawyers say so.

→ Reply

tyristik

6 weeks ago, # ^ |

Laws are trash :/

→ Reply

ChuckNorizz

6 weeks ago, # |

All math?

→ Reply

6 weeks ago, # ^ |

+29

How so?

Unless we say that all CP is math (which is true but useless), D and E are not math, and calling even A math is a stretch.

→ Reply

asdasdqwer

6 weeks ago, # |

How to solve C?

→ Reply

6 weeks ago, # ^ |

+57

Let $$$E_L$$$ be the expected no of moves to reach $$$W-1$$$ from $$$W$$$ and you are not allowed to go more than $$$W+L$$$.

When we reach $$$W-1$$$, the expected no of moves required to move to reach $$$W-2$$$ from $$$W-1$$$ is still same $$$E_L$$$, since we are not allowed to go more than $$$W+L-1$$$.

Therefore the answer is $$$(W-G)*E_L$$$.

Now let focus on $$$E_L$$$ now,
If we go left, we are done, if we go right, then we will need $$$E_{L-1}$$$ moves to come back and then another $$$E_L$$$ moves to go 1 unit left.

$$$E_L = 1+0/2+(E_L+E_{L-1})/2$$$
This gives $$$E_L = 1 + 2*L$$$.

The overall answer is $$$(W-G)*(1+2*L)$$$

→ Reply

asdasdqwer

6 weeks ago, # ^ |

"This gives $$$E_L = 1 + 2 * L$$$"

I don't see how you can get from the previous step to this.

→ Reply

6 weeks ago, # ^ |

Subtract $$$\dfrac{E_{L-1}}{2}$$$ from both sides

→ Reply

6 weeks ago, # ^ |

$$$E_0 = 1$$$, since we are forced to go left.

$$$E_L = 1+0/2+(E_L+E_{L-1})/2$$$
Moving stuff around, this simplifies to $$$E_L = 2 + E_{L-1}$$$.

$$$E_L = 2 + E_{L-1}$$$
$$$E_L = 4 + E_{L-2}$$$
$$$E_L = 2*i + E_{L-i}$$$
$$$E_L = 2*L + E_{0}$$$
$$$E_L = 2*L + 1$$$

→ Reply

asdasdqwer

6 weeks ago, # ^ |

Thanks!

→ Reply

AwakeAnay

6 weeks ago, # ^ |

+31

If you find the expected number of moves to reduce the minimum achieved weight by $$$1$$$, you can just multiply this value by the difference of the weights to find the answer. Memorylessness.

Let's find this value. You start at $$$0$$$, You move left or right on the number line with equal liklihood, except when you are at $$$L$$$, you only move towards the left.

Expected number of moves to reach $$$-1$$$ is needed, which can be found by solving a system of equations. This is the approach used by most contestants, I believe.

I would describe another approach. Let's say that you are well aware of the following fact:

If you are on the number line at $$$0$$$, moving left and right with equal liklihood, and want to stop moving after you reach either $$$-a$$$ or $$$b$$$, then the expected number of moves needed is $$$ab$$$. This can be proven by solving a similar system of equations. But, this fact is more well known (look into gambler's ruin).

In the problem, you seem to have only one choice when you are at $$$L$$$, that is to move left. Say you remove this restriction and allow yourself to go to $$$L+1$$$. Assuming that your world has been reflected. In this reflection, $$$-1$$$ will appear at $$$2L+1$$$.

Now, this has been reduced to the version of gambler's ruin bounded on both sides, with $$$a = 1$$$ and $$$b = 2L+1$$$. The expected number of moves is thus $$$2L+1$$$.

→ Reply

HynDuf

6 weeks ago, # ^ |

Excuse me, where can I find the proof of this version of "gambler's ruin bounded on both sides". Basically, "If a and b are positive integers, then the expected number of steps until a one-dimensional simple random walk starting at 0 first hits b or −a is ab."?

→ Reply

hello_india

6 weeks ago, # |

C is same as these

→ Reply

tyristik

6 weeks ago, # |

+25

Problem D is something...

→ Reply

dasdull

6 weeks ago, # ^ |

Edge cases must be insane. I thought I covered all, but still WA...

→ Reply

eduardische

6 weeks ago, # ^ |

← Rev. 7 →

Second part of the problem is simple DP after you replace every ? with 1 as it always yields a string with maximal number of splits.

First part of the problem was to split the string in groups of length 1 or 2, with each group having independent choice of options. My cases were:

Every 0 in string forms a group of length 2 with previous character and itself. If previous character is ?, options are 20 or 10; otherwise there is 1 option.
After cutting out 0 groups, we get everything else split into sub-groups.
Note that it's never valid to replace ? with 0.
If the sub-group is of length 1: if the character is ?, we have 9 options; otherwise there is 1 option.
Otherwise process the sub-group from right to left.
If the character is not ?, it forms a group of its own with 1 option.
If we have ? at the end of the sub-group, if the previous character is ? a group of size 2 and 15 options (26, 25, ..., 21, 19, ..., 11); if the previous character is 2 a group of size 1 and 6 options (6, ..., 1); otherwise it's a group of size 1 and 9 options (9, ..., 1).
If we have ? not at the end of sub-group, if the next character is 7, 8 or 9, we have 1 option (1); otherwise we have 2 options (2, 1).

The key intuitions here was to split out 0 groups, and after that 3 — 9 can only be put at the end of subgroups, 2 can be put only if the next character allows it, 1 can always be put and 0 can never be put. So I think that ??0 was trickiest case as it has 18 options, because first character will always be on its own.

→ Reply

6 weeks ago, # ^ |

-6

What edge cases did you already cover? I can't say that my solution explicitly "covers" any particular edge case.

→ Reply

6 weeks ago, # ^ |

To expand on my solution here a bit.

If the string is known, the problem is classical. It is the number of paths from $$$0$$$ to $$$n$$$ on the following graph:

an edge $$$i \to (i + 1)$$$ is added if $$$1 \le s_i \le 9$$$;
an edge $$$i \to (i + 2)$$$ is added if $$$10 \le s_i s_{i + 1} \le 26$$$.

It should be clear that all edges that this graph can ever have are present if every ? is replaced with an 1. However, some of these edges might not be part of any path.

Find all edges that are part of at least one path and discard all others. Now you have a bunch of constraints that the string you construct must satisfy. Strings that don't satisfy all these constraints will certainly have fewer paths from $$$0$$$ to $$$n$$$. I believe that this step is the key to avoiding having to explicitly handle many cases in your code. At least, doing this you make sure that you don't have to think about it.

The rest of the solution is standard. Let $$$\mathrm{dp}[i][k]$$$ be the number of ways to fill the characters starting from the $$$i$$$-th, under the assumption that the $$$i - 1$$$-th character is $$$k$$$ (where we can treat $$$k = 0$$$ as a stand-in for $$$k \ne {1, 2}$$$). If $$$\mathrm{dp}[i][k]$$$ is too large, treat this as an infinity.

Code to calculate the dp

  for (int k = 0; k < 3; k++)
    dp[n][k] = 1;

  for (int i = n - 1; i >= 0; i--) {
    for (int k = 0; k < 3; k++) {
      for (int j = 0; j < 10; j++) {
        if (s[i] != '?' && s[i] - '0' != j)
          continue;

        if (good[i][0] && j == 0)
          continue; // can't put 0 here

        if (good[i][1] && j != 1 && j != 2)
          continue;

        if (i != 0 && good[i - 1][1] &&
            (k == 0 || (k == 2 && j >= 7)))
          continue;

        // dp has a data type that rounds all numbers higher than 1e9 to 1e9
        dp[i][k] = dp[i][k] + dp[i + 1][reduce(j)];
      }
    }
  }

→ Reply

ayhan23

6 weeks ago, # ^ |

what does 'good' array stands for in your code

→ Reply

6 weeks ago, # ^ |

good[i][0] is set to true if edge $$$i \to i + 1$$$ was not removed;
good[i][1] is set to true if edge $$$i \to i + 2$$$ was not removed.

→ Reply

tfgs

6 weeks ago, # ^ |

How do you construct the k-th lexicographically largest string after calculating the dp?

→ Reply

6 weeks ago, # ^ |

Go from left to right. Try putting every character in the current position. Using the DP values you've calculated you can find out how many solutions are possible with this character. If too few — try the next character instead.

→ Reply

tfgs

6 weeks ago, # ^ |

Thank you!

→ Reply

ank7845

6 weeks ago, # |

← Rev. 2 →

Are the problems available for practice now that the contest has ended? I'd really like to know if my solution to D works as I was late by just a few seconds in downloading the input.

→ Reply

6 weeks ago, # ^ |

They will be in a bit

→ Reply

Geothermal

6 weeks ago, # |

+164

After thinking about E for a little while, I was bored and decided to just write O(2^N * |S_i|) and see if it passes. It did, taking about 5m to run on my PC. I think this is a pretty bad situation for the MHC format: I have a pretty strong PC and I am almost certain that my solution would have TLE'd on the median contestant's PC. (I also think my solution would have TLE'd on stronger test data, e.g. 105 copies of the case with 25 strings of 100 question marks.) In general I think it's bad for a possible solution to be on the border of TL (even if that solution is not intended), but it's especially harmful in the MHC format because whether or not you pass is largely based on how good your computer is, not just how well you do your constant optimization.

I hope there's a faster intended solution, though even if there is, I think this situation is still bad: I'm pretty sure a large fraction of the E ACs used the 2^N |S_i| approach and it's clearly an advantage to be able to pass with this solution.

(Incidentally, has anyone tried writing code to execute the test cases in parallel? I'm not an expert on parallel computing, but I'd imagine that on PCs with many processor cores this could sufficiently speed up execution.)

Thanks for the round!

→ Reply

rgnerdplayer

6 weeks ago, # ^ |

tourist did.

→ Reply

6 weeks ago, # ^ |

I'm using golang and my E took about 60seconds.

→ Reply

TennojiRina

6 weeks ago, # ^ |

← Rev. 2 →

-42

That's what makes MHC "fun". People should be allowed to use their own resources to solve problems.

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

-53

Author's solution is something like $$$2^{N} \times (L / 16)$$$. We expected powerful desktops to handle all test cases successfully for $$$O(2^{N} \times L)$$$, and that's exactly why this problem is in Round 1: for this problem not to decide participants, who advance to the next round.

→ Reply

Dominater069

6 weeks ago, # ^ |

Why not remove it then? All it does is leave a bad impression about the round (not that i had any good impression left after last year)

→ Reply

Geothermal

6 weeks ago, # ^ |

+20

+1, I'm not really sure whose experience is improved by the existence of this problem? I'm really disappointed to hear that the intended solution doesn't involve an algorithmic improvement from the brute force; my solution path looked like "immediately come up with $$$2^N \cdot L$$$, assume that surely that isn't intended, spend 40 minutes thinking, and eventually decide to just implement the solution, see if it passed, and move on with my life." This is only R1, it's okay for a lot of people to AK; I think it's far worse to have ranks decided by who has a better PC (especially when the problem isn't interesting enough to compensate).

→ Reply

6 weeks ago, # ^ |

-42

I don't know what you mean by algorithmic improvement, but it seems that you couldn't find an algorithm that computes value for each subset faster than in $$$O(L)$$$. While there are few of them, actually. If you want more like mathematical puzzles, you probably should ignore problems like that and find problem you like somewhere else.

→ Reply

tourist

6 weeks ago, # ^ |

+229

My experience in this round was improved by this problem. I liked the problem statement idea. The $$$O(2^N \cdot L)$$$ solution was easy for me to come up with, and even though it was enough for me to pass using a multithreaded template, I enjoyed thinking about ways to optimize it. Optimizing it $$$16$$$ times is a legit "algorithmic improvement" in my eyes.

I would not want to see this problem in Round 2 or any other round where it would decide anything. It's perfectly fine for Round 1, and I was happy to see it.

I really wish sometimes that we, as a community, were mindful and left more positive comments about the work of problem writers.

→ Reply

redpanda

6 weeks ago, # ^ |

+79

I really wish sometimes that we, as a community, were mindful and left more positive comments about the work of problem writers.

look at the author's previous comments about other problem writers lol

→ Reply

6 weeks ago, # ^ |

-131

My comments were mostly about coordinators, who haven't filtered really bad problems that gave too much freedom to cheaters and affected rating. Those people have done their job poorly and couldn't properly prevent massive cheating. Examples are pure mathematical problems of level Div2D+, where you can simply share formula and there is absolutely no way to catch cheaters after that. Also, poor round testing and wrong point assignments. So I was mostly furious because of good problems were wasted as a result of poor coordination and authors were more like victims there.

→ Reply

6 weeks ago, # ^ |

+51

Funny you say this and C is literally a one-line math problem

→ Reply

Beaboss

6 weeks ago, # ^ |

that's complexity is per test case right? I TLE'd with the same thing. Although yeah it doesn't really matter because its round 1

→ Reply

Golovanov399

6 weeks ago, # ^ |

+33

has anyone tried writing code to execute the test cases in parallel?

I dug up my codejam parallel template from like 2015 or 2016 and got E passed in like 3 minutes. Then I ran it normally in 1 thread and it took 6:34. So yeah, in my case AC and TL were separated by the number of threads, which is probably not good (on the other hand, this is a direct consequence of the contest format I guess)

→ Reply

6 weeks ago, # ^ |

+48

Other than your PC being fast or slow, this highlights another issue. You can't be sure what actually is in the test case. In theory, the file might consist of 105 test cases with $$$N = 25$$$ and $$$|S_i| = 100$$$ for all $$$i$$$ where your code actually runs in $$$\Theta(2^N |S_i|)$$$. But that's likely not the case. For many solutions I can think, many maximum tests are "easier" and there might only be one or two "hard" test cases. Finding out if this is actually so is a massive gamble.

→ Reply

6 weeks ago, # ^ |

-27

I would say, in this problem there are three separate states:

You solved this problem and your solution works way faster than 3 seconds per test case.
Your couldn't solve it in a normal way, but found a good way to implement slower solution and got your points (it gave you nothing but a higher rank in the standings of Round 1).
None of the above. Not critical at all.

It could've been critical, if the problem affected someone's CF/AC/whatever rating or spot in the next MHC Round. But not the case.

Anyway, congrats to everyone, who is in the first state.

→ Reply

jeroenodb

6 weeks ago, # ^ |

← Rev. 3 →

+17

I think another solution is that you take some slightly slower solution, and run it on $$$16$$$ testcases at the same time. This is essentially the same as doing the total time $$$/16$$$. What is your view on this? You can calculate that your solution should run in time, only you don't use bit parallelism but actual parallelism, is it really that different? (Of course it is much less 'cool', but what makes this not a legit solution?)

→ Reply

6 weeks ago, # ^ |

Within this competition's format it sounds totally legit to me. I agree that it makes a lot of sense to have parallelism at your sleeve for cases like that. This way or another, you will get these points. Then it's everyone's choice if they want to find a solution, that works for ICPC format as well.

→ Reply

AwakeAnay

6 weeks ago, # ^ |

Threading worked.

I was implementing a $$$O(2^N \times 26)$$$ solution. Using meet in the middle and breaking bit strings into two $$$64-bit$$$ numbers.

For a binary string with only '?'s and characters. One can compute the number of nodes required by splitting the string into $$$4$$$ $$$25$$$-bit numbers, having precomputed counts for all $$$25$$$-bit numbers and merging these by linear combinations.

Somewhere around, I realized that I wasn't sure whether this would pass. I was using an M3 Pro with $$$11$$$ cores, so I ran each testcase in a separate thread and it passed. Took 3 minutes. Meet in the middle was still required, however, for memory limits.

→ Reply

Tlatoani

6 weeks ago, # ^ |

I ran them in parallel (Kotlin), took like 40-50 seconds. Serially it seems like it would have taken 7-8 minutes, so I actually wouldn't have gotten it.

→ Reply

Wielomian

6 weeks ago, # ^ |

← Rev. 2 →

I also wrote a $$$2^N |S_i|$$$ dp, but didn't bother at all with optimization. I used STL maps and sets, everything was single threaded, my PC is a 6 yo laptop and it run about 4 minutes (minGW + WSL). I liked the problem even though this solution was a little bit to straightforward to come up with for the last problem in the contest.

I'm in fact surprised that so many people TL'ed this problem :thinking

→ Reply

elsantodel90

6 weeks ago, # ^ |

I checked your code, and unlike many of us that just did a straightforward for (mask = 0; mask < (1<<N); mask++), so we always explicitly iterate all 2^N possibilities and then for each one we intersect those patterns in order to do the literal exclusion/inclusion formula, your code uses two maps as queues and does some interesting swaps so it is quite more involved in the way things are computed.

I assume that because of that it only looks at "interesting" subsets, and this is bound to be much faster in practice (the straightforward inclusion-exclusion formula always has 2^N terms to add/subtract, no matter what are the input strings).

If I am not mistaken, your code can be described as an "optimized version of inclusion-exclusion formula", where you go element by element and try "intersecting it" with all previously existing sets to construct new sets, while critically deleting anything that becomes empty.

So I think you did bother with optimization! At least compared to mine / other people's brute force version :)

→ Reply

Wielomian

6 weeks ago, # ^ |

Thanks for the exhaustive reply. Very interesting point, I would think that the tests would blow the number of interesting masks straight to $$$2^N$$$ after just a few iterations. So maybe in practice this really is somewhat faster.

The queues are "masks for current character index" and "masks for the next character index", in a fashion similar to 0-1 BFS.

→ Reply

6 weeks ago, # |

→ Reply

6 weeks ago, # ^ |

Connections have recovered

→ Reply

6 weeks ago, # ^ |

+11

Any reason why the "Validate Solution & Submit" is grayed out? Would love to upsolve https://i.imgur.com/BRLWJ9t.png

→ Reply

6 weeks ago, # ^ |

It is working now. Happy upsolving!

→ Reply

hxu10

6 weeks ago, # |

← Rev. 2 →

+26

Extremely weak pretest for problem D. Nearly every false solution can pass.

→ Reply

dasdull

6 weeks ago, # ^ |

+42

I think that was part of the challenge :)

→ Reply

6 weeks ago, # ^ |

E test data is also quite weak, I managed to pass all but one of the test cases even though my solution outputs the wrong answer on:

1
2
A
B

(4 instead of 3) since I forgot to delete some if statements from my solution. Anyway, not a good idea to rely on validation data being strong. :)

→ Reply

6 weeks ago, # |

That feeling when you realize the solution is O(t * 2^n * |S|), with t = 105, n=25, |S|=100, and you are using Python.

No chicken dinner

→ Reply

Beaboss

6 weeks ago, # ^ |

+10

is this the intended solution? lol i couldn't get this to be fast in C++

→ Reply

6 weeks ago, # ^ |

I think so. Only after the end did I realize that with some bitset magic, you can probably speed it up like a factor of 16 by using that you can store 16 alphabetic characters in a 64 bit int.

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

64/log_2(26) is 13.6 :susge:

→ Reply

maomao90

6 weeks ago, # ^ |

← Rev. 3 →

My solution works in $$$O(B2^B+T2^N(\frac{|S|}{B}+3))$$$ lol. It managed to run in 55 seconds with $$$B = 25$$$.

→ Reply

LowPill

6 weeks ago, # ^ |

Python 3.13 is on the way

→ Reply

TooTiredRetiring

6 weeks ago, # |

SecondThread would there be plagiarism checks?

→ Reply

6 weeks ago, # ^ |

Yes, we haven't started them quite yet.

→ Reply

TooTiredRetiring

6 weeks ago, # ^ |

-13

Thanks, please do them, disappointed in seeing quick influx of submissions at the end and my rank dipping so quick

→ Reply

_HUECTRUM_

6 weeks ago, # ^ |

So you're basically disappointed that people solved the problems?

→ Reply

distinguished_dunker

6 weeks ago, # ^ |

← Rev. 2 →

→ Reply

AltHaiBhai

6 weeks ago, # ^ |

-8

IDK why there is a need of getting the code file, if that is not checked after running whether it is giving correct output or not. There are some users who just get the output files maybe from some of their friends or just buy them and then just submit any random code. Below is an example, check his submission in Problem A. Not only that, he had done same thing in Hacker cup 2023, round 2 also. You can check any of his source code file.

See image here: https://ibb.co/K688bPt

Please do something about this, there are some genuine users who don't get to qualify by 2 or 3 ranks.

→ Reply

theRealChainman

6 weeks ago, # |

Great problemset, I enjoyed problem B and C a lot. Problem D looks like a pain to implement, if the approach I came up with is correct, which I'll see in practice mode later. Thanks for the contest!

→ Reply

6 weeks ago, # |

-16

My ranking is 5055! 1st problem was quite difficult for me. I hope at least 55 people get rejected for cheating(lol)

→ Reply

san9blin

6 weeks ago, # |

Hello, is it possible to account for the problem if it is correct but i forgot to compile the latest version before saving answer to output file on my PC. The source code I added is correct.

→ Reply

6 weeks ago, # ^ |

I think they Don't run the code. So as long as solution file is correct you are good!

→ Reply

6 weeks ago, # |

For some reason my score is 31/100, despite solving the first 3 problems? It should be 51/100 or am I missing something?

→ Reply

6 weeks ago, # ^ |

What's your handle?

→ Reply

6 weeks ago, # ^ |

Benjamin Kleyn (kriepiekrollie)

→ Reply

Edvard

6 weeks ago, # ^ |

I reported the same issue (handle: Edvard): in the last 15min of the contest scoreboard was showing correctly that I submitted 4 problems but the score was 51 (as if 3 problems were submitted).

→ Reply

heavenly_

6 weeks ago, # ^ |

Did you fail systest on C?

→ Reply

6 weeks ago, # ^ |

Doesn't look like it

→ Reply

heavenly_

6 weeks ago, # ^ |

thats a really weird bug, hope that SecondThread will fix it soon!

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

+25

I was scrolling through the score board. Found one with the opposite issue, 2 solves but has 51 points. Maybe Navin stole your points =p

→ Reply

6 weeks ago, # ^ |

→ Reply

nkelkins

5 weeks ago, # ^ |

← Rev. 2 →

Actually, I am that Navin who got the extra points lol. There were server issues when I was submitting the code and answer. The solution was correct, so i guess they considered it.

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

Scrolled through all of top 5000. In total there are 6 people with a buggy score. One with more points than indicated by the solved problems, and 5 with fewer points. So you are not alone.

→ Reply

iadityakumar

6 weeks ago, # |

When are the problems be available for practice now?

→ Reply

6 weeks ago, # ^ |

Yes, you should be able to submit to practice now.

→ Reply

TwentyOneHundredOrBust

6 weeks ago, # ^ |

+27

It's not working for me, it just says "Timer expired"

→ Reply

iadityakumar

6 weeks ago, # ^ |

Same

→ Reply

hxu10

6 weeks ago, # ^ |

I try to submit at the front page, it tells me Unable to create submission. I cannot even submit the test case in the practice round which is 2 weeks ago.

→ Reply

my_new_acc

6 weeks ago, # ^ |

Can someone submit for practice ?

I can't see it, please tell where to do it

→ Reply

Dart-Xeyter

6 weeks ago, # |

← Rev. 2 →

Hello, where can I find the upsolving of the First Round?)

→ Reply

6 weeks ago, # |

-10

Hey,plagiarism checks finished?

→ Reply

distinguished_dunker

6 weeks ago, # ^ |

← Rev. 3 →

-15

→ Reply

kelzzin2_fan

6 weeks ago, # |

← Rev. 2 →

RIP to those who break early before reading all input and passed all the validation TC for A...

→ Reply

hxu10

6 weeks ago, # |

The interesting thing is that, when I try to submit the problem D in the last minute, and after revealing real result, I get a wrong answer on problem D and .... my ranking even increased a little.

→ Reply

Eslam_Ahmed

6 weeks ago, # |

why can't submit problems after contest ended?

→ Reply

6 weeks ago, # |

I had no time to do the contest fully and I had a floating point inaccuracy in the first problem final submit testcase (36.000000 <= 36.000000 was false, I thought setprecision remedied this but I was wrong). From rank 2200 to 5500. Oof. I forgot that the feedback to the submission is given after the contest. If only, the format was like Codeforces. Eh I guess better luck next year.

→ Reply

6 weeks ago, # ^ |

+14

That is a perfect example of why you should use rational numbers instead floating point numbers. Especially on a problem like this where it is straight forward to solve it with rational numbers.

Python solution

from fractions import Fraction as f

minspeed = -1
maxspeed = 10**18

for i in range(n):
    minspeed = max(minspeed, f(i + 1, B[i]))
    if A[i] != 0:
        maxspeed = min(maxspeed, f(i + 1, A[i]))

if minspeed <= maxspeed:
    printer(float(minspeed))
else:
    printer(-1)

→ Reply

Synaptic_Savant

6 weeks ago, # ^ |

How reliable is decimal.Decimal for these problems?

→ Reply

6 weeks ago, # ^ |

There are many reasons I would avoid decimal. I guess decimal is fine if you want a really high precision floating point number, and you don't care about speed. But if you want provably correct solutions, then you should stick to fractions.

→ Reply

6 weeks ago, # ^ |

For problem A was it necessary to use fractions to pass?

→ Reply

TianyiChen

6 weeks ago, # ^ |

Passed with double, I did try some corner cases before submitting.

→ Reply

sparsetable

6 weeks ago, # ^ |

As far as I know, setprecision only affects the amount of digits outputted. pajenegod's solution using fractions is very clean, but for future reference, you can also use an error term eps (i usually set it to something between $$$10^{-7}$$$ and $$$10^{-9}$$$) and check if a > b + eps instead. This is usually enough to mitigate precision errors (though I also spam long doubles just to be safe)

→ Reply

6 weeks ago, # ^ |

I was using only long doubles. Most times that is enough. I was aware of the epsilon fix but was too lazy and did not realise I had a precision problem because it was in the final testcase and delayed feedback of that. I guess now I have an actual example to fallback on why to use epsilon.

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

Would you mind sharing your solution code as well as the fractions that were being compared incorrectly? Because I just implemented the straightforward floating point solution without any epsilon and it passed.

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

My solution

#include<bits/stdc++.h>
using namespace std;
#define int long long int
#define double long double
void IO(string s = "") {
    ios_base::sync_with_stdio(0);
    cin.tie(0);
    if (!s.empty()) {
        freopen((s + ".in").c_str(), "r", stdin);
        freopen((s + ".out").c_str(), "w", stdout);
    }
}
void solve() {
    int n;
    cin >> n;
    double speed;
    vector<pair<double, double>> segments(n);
    for (auto &[x, y] : segments) {
        cin >> x >> y;
    }
    speed = 1.0L/segments[0].second;
    for (int i = 1; i < n; i++) {
        speed = max(speed, (i+1.0L)/segments[i].second);
    }
    bool f = 1;
    for (int i = 0; i < n; i++) {
        if (!(segments[i].first <= (i+1.0L)/speed && (i+1.0L)/speed <= segments[i].second)) {
            f = 0;
            break;
        }
    }
    if (f) {
        cout << fixed << setprecision(7) << speed << endl;
    } else {
        cout << -1 << endl;
    }
}
int32_t main() {
    IO("A");
    int t;
    cin >> t;
    for (int i = 1; i <= t; i++) {
        cout << "Case #" << i << ": ";
        solve();
    }
}

It was my check that had the problem, "(i+1.0L)/speed <= segments[i].second)" was the condition where 36.000000 <= 36.000000 was false. I understand that I could have just kept the information of both minimum speeds and maximum speeds and used them but because the time complexity allowed it, I thought that my check would have been better and well it came first to mind than the minimum and maximum speed check. Oh, how wrong I was. It is a bit scuffed that I took ints in as doubles but that should not be that problematic?

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

I see. If you had compared minimum / maximum speeds there wouldn't be any issue.

Reasoning: If you want to compare two fractions of the form $$$a/b \le c/d$$$ where $$$a,b,c,d\le 10^6$$$, then double precision suffices since it has $$$15>12=2\cdot 6$$$ digits of precision

→ Reply

6 weeks ago, # ^ |

Interesting. I used epsilon then it gave me correct answers. What about long double precision, I used long doubles. I defined double to be long double.

→ Reply

6 weeks ago, # ^ |

The precision of long double vs double is platform-dependent. E.g., on codeforces you will get 18 / 15 digits of precision from long double / double respectively, though on my computer I just get 15 printed twice.

#include <iostream>     // std::cout
#include <limits>       // std::numeric_limits
int main(){
    std::cout << std::numeric_limits<long double>::digits10 << std::endl;
    std::cout << std::numeric_limits<double>::digits10 << std::endl;
}

→ Reply

6 weeks ago, # ^ |

← Rev. 2 →

I gained new insights because of my mistake. Thank you for some of them! Now I am just confused about how even <= works for doubles and how many digits does it compare, because sometimes 0.0 does equal 0.0 but sometimes not, I understand that it did not matter whether I used double or long double, as the important digits were accurate, <= would read the inaccurate digits too, I think but I am not sure any more.

→ Reply

https://en.wikipedia.org/wiki/IEEE_754#Directed_roundings

6 weeks ago, # ^ |

← Rev. 2 →

Unless specified otherwise, the floating-point result of an operation is determined by applying the rounding function on the infinitely precise (mathematical) result. Such an operation is said to be correctly rounded. This requirement is called correct rounding.[20]

Basically, if you are computing the quotient of two integers in double precision, the result is rounded to the nearest possible double. So it is safe to compare quotients of integers if you have enough digits of precision, because rounding occurs only once.

However, in your solution you have two divisions:

speed = 1.0L/segments[0].second;
...
(i+1.0L)/speed

Which effectively causes rounding twice instead of once. This introduces imprecision (regardless of whether using double or long double).

→ Reply

6 weeks ago, # ^ |

← Rev. 4 →

Oh okay, Thank you for taking the time to answer. I understand the problem that my code faced now a bit more. I guess long double was not enough to remedy this, probably is the same as double in my operating system (Edit: I checked, long double is 18 and double is 15 for me).

→ Reply

UP84

6 weeks ago, # |

3000 AC for problem C,seems a bit much.

→ Reply

distinguished_dunker

6 weeks ago, # ^ |

← Rev. 2 →

→ Reply

rumike

6 weeks ago, # |

B killed me, the of idea of twin primes directly came to my mind, but I read $$$10000000$$$ as $$$10^8$$$ and lost about 40 min, thinking and googling a method which can find twin numbers in $$$O(\sqrt{n})$$$

→ Reply

6 weeks ago, # ^ |

+13

You can just store all of those primes in memory ;D

→ Reply

_HUECTRUM_

6 weeks ago, # ^ |

← Rev. 2 →

+13

O(n) (precomputation) is enough for 10^8, considering you have like 5 mins to run it.

→ Reply

YoussefMorad1

6 weeks ago, # |

+10

Please fix the practice issue. Problem D and I have some unresolved issues!

→ Reply

hypeskynick

6 weeks ago, # |

-11

please rejudge my C submission, the servers died and I couldn't submit my code :((

I am 不是nick

→ Reply

tickbird

6 weeks ago, # |

-10

on AI track 1st-pl AI AK R1 in 22min, is this real...?

→ Reply

puupil

6 weeks ago, # |

+19

feeling so sad right now, I was qualifying but didn't realize that my code for B failed at $$$ n = 4 $$$. My first ever experience, was nice. Thanks for great problems!

→ Reply

ayush_n

6 weeks ago, # ^ |

Same but my both answer went wrong Problem A due to precision and Problem B due to n=4 :(

→ Reply

ASHWANTH_K

6 weeks ago, # |

+14

Can we upsolve problems for Round 1 meta hackercup? I dont find an option to submit solutions now

→ Reply

6 weeks ago, # ^ |

I think they forgot to turn on the practice mode SecondThread

→ Reply

4 weeks ago, # ^ |

Should be enabled now

→ Reply

TheBhediyaOfDalalStreet

6 weeks ago, # |

← Rev. 2 →

Had fun deriving my solution for C (after guessing the solution in 15 minutes and then banging my head for 1 hour deriving it to figure out what's wrong, only to realise at the end that I just had to take the modulo of the input as well)

My derivation was just writing down a bunch of recurrence relations and solving for the expectation at current state. An observation I made while deriving was that if you move downwards in the random walk, ie from $$$W$$$ to $$$W-1$$$, you would end up in the exact same situation as before with lesser distance to cover. So we could use this symmetry to avoid writing a recurrence for those states.

Concretely, I made an assumption that the expected steps would then be linearly proportional to the distance between the states ($$$w$$$ and $$$G$$$):

$$$\quad$$$

$$$ e(w) \propto w - G + 1 \quad \forall \quad G < w \leq W $$$

This let me simplify:

$$$ e(W) = 0.5 * (1 + e(W + 1)) + 0.5 * (1 + e(W - 1)) $$$

By calculating $$$e(W - 1)$$$ using the assumption, and finding the value of $$$e(W + 1)$$$ in terms of e(W) using recurrence.

Issue is, I still can't rigorously prove why the assumption of linear proportionality would hold. I mean, it is apparent that via symmetry we end up in the same situation, but with lesser distance to cover. But why would it be linearly proportional to the distance?

Does anyone who rigorously solved the problem have a proof for this assumption/observation?

→ Reply

TheBhediyaOfDalalStreet

6 weeks ago, # ^ |

Update: Figured it out via this comment

So basically, The path $$$W \rightarrow W - 1 \rightarrow W - 2 \rightarrow ... \rightarrow G $$$ can be broken into steps of 1, and via symmetry we know that going 1 state down on the path from any arbitrary state will take the same number of expected steps, ie $$$E_L$$$. This immediately proves the "assumption"

→ Reply

6 weeks ago, # |

Thank you for the nice and particular problems, especially C that left me something. Just a question, is practice mode available? The validate button seems disabled here

→ Reply

A_Le_K

6 weeks ago, # |

+11

The most important question of life, meaning and existance:

~~ArE wE gOnNa GeT rEpLaCeD bY AI?~~

Were these results in AI open/closed track expected?

→ Reply

6 weeks ago, # ^ |

I think good contestants just didn't bother with AI track. Is AI open rules a superset of human track?

→ Reply

6 weeks ago, # ^ |

I think the AI results are coehrent with what we currently see in business. Surprisingly good help for simple tasks, still ways to go when a bit complex reasoning is needed. Let's consider also that here there are sample cases, which is not the case in real world. For example, problem C: I am quite sure that none of the 4 that solved it "deduced" the solution from the actual math underlying the problem. The solution was simple enough to be found with a few reasonable trial and error attempts (answer proportional to L and W-G with some constants).

→ Reply

Post contest discussion stream Round 1 ABCE

6 weeks ago, # |

-9

→ Reply

TwentyOneHundredOrBust

6 weeks ago, # |

← Rev. 2 →

Submitting for practice seems to still be closed. Could it be enabled? SecondThread

→ Reply

TianyiChen

6 weeks ago, # ^ |

For those who probably missed it: though the practice submission is not on, there's a link on the solution tab for downloading input and output files.

→ Reply

6 weeks ago, # ^ |

+10

Working now.

→ Reply

htoshiro

6 weeks ago, # |

Just a quick question here, but I qualified for round 2 but in the round 2 contest it says that I am not registered for the contest. Is this intentional? Thanks.

→ Reply

wjomlex

6 weeks ago, # ^ |

Official advancement usually happens about a week after the round, after plagiarism checks have been done.

→ Reply

6 weeks ago, # ^ |

-10

So plagiarism checks has not done yet?

→ Reply

5 weeks ago, # ^ |

+22

There are less than 3 days left until round 2, but it's still showing You are not registered for this contest. for round 2, even though I'm among the top 5000 of round 1.

Is that expected?

→ Reply

chuka231

5 weeks ago, # ^ |

same, hope they update the system soon

→ Reply

wjomlex

4 weeks ago, # ^ |

Advancements will take place today, probably in about 3 hours.

→ Reply

tushgupt0512

4 weeks ago, # |

-7

I have been checking the scoreboard of round 1 everyday. I just got to know that my solution to problem A was incorrect even though it was showing as accepted till yesterday.

Q1: Why is the contestant required to submit the source code file + output file? My understanding till now was that the source code file might be run by the hackercup team if they find the need to do so only on the input file for which the output file is also submitted. If the team is going to run the source code again on every testcase(that are not even the part of initial test cases) what is the need of output file from contestant?

Q2: How is the team detecting TLEs in the solutions to problem 5 or in any other solution for that matter? What would have happened if I had submitted a solution which would have taken eternity to execute for that particular case for which it is currently failing? The time limit, memory limit is mentioned nowhere in problem description. If the solution is being run only on the input file, there is still 6 minutes time limit in worst case.

Q3: There is leetcode that tells you after 4 days of contest that your solutions to the problems were incorrect. And then there is the hackercup team who is telling you after 13 days(just 1 day before the round 2) that your solution is incorrect. I don't know what to say. If the test cases were not prepared beforehand, there is no point of having such a problem in such a contest. And if they were prepared beforehand, why were they not present in the input files during the contest? I am not sad about codejam and kickstart are no longer happening but am happy because they happened. Not sure if I feel the same about hackercup though.

→ Reply

tushgupt0512

4 weeks ago, # ^ |

SecondThread wjomlex Please reply as there are only a few hours left before the contest

→ Reply

4 weeks ago, # ^ |

The test cases were prepared beforehand. There was one problem for which one our checker was being overly generous (our oversight) by more than our allotted epsilon. We identified this after the contest had ended, and regraded the submissions of those affected, (it was under 100 contestants iirc).

imho this is less of an issue than it would be in say a leetcode contest with because final graded results are not known in a hacker cup round until after the contest ends, so this doesn't affect decision making. But we've fixed the underlying issue and this specific thing won't happen again regardless, apologies for giving you any false hope that A was right.

If your solution would have taken an eternity to execute, it'd be very difficult for you to upload the output which you don't have to your input within 6 minutes, and you won't get points for that problem. Not sure I understand your question here.

We don't necessarily run every submission to every problem, but we will check individual solutions if we suspect they either do not run in time, we suspect foul play, or we decide we want to run your solution. But in general (at least this year), we give you the freedom to run your own solutions however you like, so long as they complete within 6 minutes on a modest computer.

→ Reply

tushgupt0512

4 weeks ago, # ^ |

My bad, I misunderstood the problem here as it wasn't clearly mentioned anywhere why all of a sudden an accepted solution got rejected. Thanks for the clarification SecondThread

→ Reply

MosesTheRedSea

4 weeks ago, # |

+35

Algo professor promised an A in the class for Top 500 LOL

→ Reply

4 weeks ago, # ^ |

+35

Based

→ Reply

Snapper_001

4 weeks ago, # |

is tshirts is uptill 2000 in round2?

→ Reply

4 weeks ago, # ^ |

Yes, and top 200 in round 3 will have a special badge on their shirt.

→ Reply

piyush_pransukhka

4 weeks ago, # |

← Rev. 2 →

+27

What color are the t-shirts this year?

→ Reply

ryucc

4 weeks ago, # |

← Rev. 2 →

+27

I feel there are some sketchy submissions. not having "Case" anywhere in the code could be a useful feature.

Sorry in advance if I misread any of your code, or if the submission was correct, and I just couldn't understand it. I don't want to imply cheating, it might just be a misunderstanding of the contest rules (for example uploading inputs instead of code). Let me know, I will remove the line.

Sumit Patel (s_mt — Code for A1 and A2 are exactly the same, algorithm looks like geometry.

Soumya Ranja — Code for A1 only contains input/output, no algorithm

Amit Kumar (Amit7777) — source code is just input data

Swadhin Majhi (Nihda — source code is just input data

Gangaram Sudewad (Gangaram_25 — Irrelevant algorithm, doesn’t print “Case” anywhere in the code

Suraj Kumar (Gsudewa — Irrelevant algorithm, doesn’t print “Case” anywhere in the code

→ Reply

chomotngayloihuanohoa

4 weeks ago, # ^ |

Soumya Ranja bro literally copy paste something, I suspect that too

→ Reply

4 weeks ago, # ^ |

Thanks! We'll be doing plagiarism reviews over the next couple days and will look into these ones specifically.

→ Reply

chuka231

4 weeks ago, # |

+20

Finally won a t-shirt after many tries

→ Reply

Swistakk

4 weeks ago, # |

+54

Not everyday I fail a problem on systests and raise from ~350th to 36th place in the result of that xD

→ Reply

orz

4 weeks ago, # |

-66

Dear SecondThread,

Please could you by chance add me automatically to round 3? I am currently in UTC+8 timezone, so for me the start of the round was 1AM, and I simply overslept it. To not have you made an impression that I simply want to skip a round and qualify without any hard work, here's a screenshot from my alarm clock list where you can see 0:55 and 0:56 scheduled:

They were on yesterday, but still somehow I unconsciously turned all of them off.

→ Reply

4 weeks ago, # ^ |

As much as I would love to, unfortunately we can't make exceptions for certain people after the fact because otherwise it'd be unfair to everyone who participated and didn't make it to round 3.

→ Reply

Aditya8676

4 weeks ago, # |

← Rev. 4 →

Dear SecondThread , I have my gmail linked to my facebook account but its not showing on my hacker cup account.Kindly fix this bug,so that you can contact me during t-shirt avail process.

Update : Issue resolved.

→ Reply