Meta Hacker Cup 2025
Meta Hacker Cup is back! We’re excited to announce our schedule for our 2025 season, kicking off on October 10th!
- Practice Round: Fri. October 10th, 10am Pacific (120 hours, 5 days)*
- Round 1: Sat. October 18th, 10am Pacific (3 hours)
- Round 2: Sat. November 1st, 10am Pacific (3 hours)
- Round 3: Sat. November 15th, 10am Pacific (3 hours)
- Finals: Sat. December 13th, 6am Pacific (4 hours)
*While optional, we recommend you participate in the Practice Round to familiarize yourself with our submission system before Round 1, when time will be at a premium.
The contest will be held on the Meta Hacker Cup site. Registration will open Friday, October 3rd. Register here.
You can expect familiar prizes in the human track, including T-Shirts, Elite T-Shirts, and cash prizes for finalists. We’ll announce more prize details closer to Round 2.
Continuing the Meta Hacker Cup AI Track
We'll be continuing our AI track this year, but with some modifications. We'll share specific details on Friday when registration opens, but at a high level, you can expect a single AI track with time limits matching those in the human track. We won't impose any limits on compute resources or energy usage in the AI track this year.
Additionally, we have an open invitation to companies developing AI to compete in our AI track, including OpenAI, Google, Anthropic, xAI, and any other companies that are interested in competing. For large companies that elect not compete, we'll run their top publicly available model as of our registration date with default settings on a single inference attempt per problem to represent them instead.
If you're interested in competing in the AI track, you can join our discord server to learn more.
Update: Registration is now open -- Register here!
Auto comment: topic has been updated by SecondThread (previous revision, new revision, compare).
FINALLY ITS HAPPENING
Hi SecondThread,
I just saw that MHC is happening again this year and am so excited! However, it seems that Round 3 is going to conflict with the ICPC regionals for the Pacific Northwest region: http://acmicpc-pacnw.org/ which contains some historically high-performing teams from Stanford, Berkeley, and some other UCs. Given this, would there be any possibility to move MHC back a few hours or to a different day?
I completely understand if this won't be possible-- just wanted to alert you of the conflict. Thank you for bringing MHC to us every year!
SecondThread Thank you so much for giving us the good news. MHC is happening, let's go.
Finally :) hoping problems will be anti-gpt
holy heck it's happening
It's happening , lets go !!!
Auto comment: topic has been updated by SecondThread (previous revision, new revision, compare).
can we join like in 2023 if we aren't 18 yet? Not quite sure about the policy as not yet 10/3
Technically, Meta's official policy on this has always been that you need to be 18 to register for the contest
ChatGPT mentioned that there was a "Meta Hacker Cup High School Division (HS)". Is that not accurate? I would love to participate.
Key points: • High school students under 18 are eligible to participate in the HS Division, which runs in parallel to the main competition. • The HS division typically uses the same problem sets, but results are filtered to rank only high school participants. • The prizes and legal agreements are different from the main track (e.g., HS division often requires parental consent if the student is a minor). • Participants under 18 cannot enter the main division, but their performance in the HS division is publicly ranked and recognized.
A quick google search reveals this is probably a hallucination.
oh wow !! let's go
Lets gooo. Thank you so much for this.
Thank you, SecondThread!
That's very exciting! Been waiting for this news for some time! Thanks to the Meta Hacker Cup team for organising the competition :D
LETS GOOOOOOOOOOOOOOOOO !!!!!!!!!!!!!!!!!!!
is usage of AI allowed in the human track?
why? do you want to cheat ?
he cheats every day
It is not cheating if it is allowed....Unfortunately we are in the state that AI being allowed or not makes even my contest strategy different, and so a decision either way is good for me to decide.
In 2024, the rules had: "You may not use any AI system for assistance during the contest."
Thanks
Dkijstra?
dkijstra.
djiskara!
maskara
Hi can you update the site to have the 2025 schedule?
Are meta employees allowed to participate ? will we get t-shits ??
Meta employees are not allowed to compete, but you can participate by helping write problems, or help in other ways! Ping me on Work Chat and we can get you set up!
Based and Zuck-pilled
Finally! I was worried it wouldn't happen this year because of all the cheating through AI and all...but extremely glad to know that they are continuing with this!
Big love!
It's happening!!! YAY
SecondThread, sorry for the notification but I registered for meta a few days ago, with my full legal name, main email, phone number, and with the recording of my face (the thing where you move it to the side), and my account got auto disabled, in fact the "read more about account integrity" page does not even work, so I don't know what to do, and frankly I don't know what I've done wrong either? if you need specific details I can message them.
Also yes I'm 18+ and reported my birth date accurately
Hi, same here,
Did you find a solution?
Yeah as the comment below says, using mobile worked first try.
Creating the account from my mobile , fixed this for me , didn't even ask for verification. I used email for verification.
Okay Thank you sm!
omg omg omg !!!!!!!!!!!!!!!!!
tuff
Looks like 10/18 will be very busy? Div1 CF 7:30-9:30 PST, Hackercup 10:00-1:00 PST
Have the registrations started?
Not yet -- it'll be open in about 24 hours though!
Aight thanks
Exciting month ahead.
Counting down the days to Meta Hacker Cup 2025. Long live Hacker Cup!
Have the registrations started?
Yep! You can register here: https://www.facebook.com/codingcompetitions/hacker-cup
Wow!!! let's gooooooooooo
Finally it's happening!
Thank you!
So I was giving a last year mock as it is my first participating,
https://www.facebook.com/codingcompetitions/hacker-cup/2024/round-1/problems/A
for this problem the downloaded case is so large that it is not working for me on vs code, codechef ide, ideone. So any suggestions how to handle this?
I feel like it's not working because you are trying to copy paste the input in the usual cf format. If this is the case (if it isn't idk how to help you), you should instead take the input from the downloaded file itself, something like:
I used this template last year, for sure it can be improved (I will for this year hacker cup), but it should still solve your problem.
Thanks it worked
Do you have the template for maximizing the stack size?
i use Clion and do it in Cmake file like this:
Btw you can ask llms to check for other IDEs.
i use sublime txt and LLMs told me to make a custom build. Just making sure is this build correct? Its mine first time at hacker cup...dont wanna ruin my qual chances
Hey , I wanted to know , does this work ever for very large inputs , it worked in practice but i am not sure if it will work all the time , Could you please clarify?
For AI track, will there be two tracks — open model track and close model track this year?
is this a team contest?
No, all rounds are individual contests
Thanks a lot for the update, sir. Excited for the first time.
When I am trying to submit my output file, a pop up comes saying "Something Went Wrong". Does anyone else have the same issue?
try using incognito tab . I too faced the same
I can upload files when I use edge, but it doesn’t work in chrome or firefox. I don't know why :))
didn't able to submit in practice itself , add more servers :) ss attached
Why aren’t people under 18 allowed to join?
There are many talented programmers under 18 who have done really well in the IOI, and it would be sad if they couldn’t participate.
i have few doubts regarding the competition
What if people did the following in the contest
1) do multithreading , and somehow make their O(n ^ 2) code run for an O(n) indented solution
2) someone created an another account and than downloaded the test case files, or simply just borrowed it from their friends
3) for every question we are getting 5 minutes , which is a lot like normally on codeforces we get approx 3 minutes
so that means codes with time complexity 500 * n , n logn * logn * logn will work
I think for point 2, the input is rearranged and is a different permutation for each participant. The only way you can use the input of some other account is if you store hash values of each test case with their answers(which itself might not always be feasible), and will be caught as they ask you to submit the source code as well.
When O(N) is intended the bounds are a lot higher (e.g. 10^7)
Can't submit anything including validation and clarification today. It pops up a "Something Went Wrong" and seems nothing could be done on the platform, and I tried different network which didn't help.
Here's the screenshot: Google Drive.
See you at 18th
Will we get an option to download test cases directly without downloading encrypted file in contest as well ?
Can someone provide a problem from past MHC rounds which is expected to throw an error of stackoverflow if stack limit is not properly increased? Just in case if you remember. Thanks.
I remember almost failing this one because of this: https://www.facebook.com/codingcompetitions/hacker-cup/2023/round-2/problems/A2
Thanks.
just check warm-up problem of the 2025 hacker-cup practce
I have tried to create a Facebook account to participate in the Hacker Cup. I was told to upload a video selfie, which I did. After an hour or so, the verdict came:
"We disabled your account We reviewed your account and found that it still doesn't follow our Community Standards on account integrity. You cannot request another review of this decision. To learn more about the reasons we disable accounts visit the Community Standards. You may have the right to challenge our decision in your country's courts, or refer our decision to a certified dispute settlement body."
Any chance to review this in light of Hacker Cup participation, or no Hacker Cup for me this year?
(the email is the same as the one I used for all previous Hacker Cup participations)
(to be clear, I do not have any other Facebook accounts, I have deleted the previous one in the end of 2024. I have only entered my name, birth date, gender and email in the standard account creation flow, and immediately got the video selfie verification, I did not even get to use the newly created account before it got disabled)
I've managed to create an account by changing as many things as possible from the first attempt (desktop vs mobile, home WiFi vs mobile hotspot, email vs phone as username), and registered for the Hacker Cup. I hope the new account does not get disabled while the Hacker Cup is still ongoing :)
Did he ask you for your face for identification? It’s just that I also tried to register, and it asked me for my face. After that, it banned me :(
Same problem for me... I probably should have tried to make an account before today :) I guess no hackercup for me this year :(
Same I tried to submit and it kept saying "something went wrong" so I tried to make a new account and it disabled it.
Hey , I wanted to know , does this work ever for very large inputs , it worked in practice but i am not sure if it will work all the time , Could you please clarify?
For practice round:
Problem C was cute.
Problem D was great. Was not expecting to require Eulerian cycles for the optimal split.
Anyone can provide editorial for problem E? It's the only one I didn't solve (I did not think about it a lot so maybe that's just my fault)
Here is the outline of my solution.
Do it offline. So, it is easy to retrieve the position of the left and right walls of each query. WLOG, we can assume that the left wall is at position $$$0$$$ and the right wall is at position $$$L$$$.
Also, process the queries in decreasing order of their robot ID. This way, your robot of interest will always pay off everyone (in your data structure) it passes.
Also note that each pair of robots will hit each other twice in their $$$2L$$$ period (before returning to their initial configuration). The case of interest will be the case where $$$s \lt 2L$$$.
Clearly, there is a closed form of hitting time of robots, which depends on the distance of your robot of interest, let's call it $$$t(d)$$$. Observe that out of the robots to the left, $$$t(d)$$$ is monotone. Same thing for robots to the right. This way, you can find the position of the last robot.
What you say makes perfect sense. Thank you for the explanation. I suppose you implemented a treap for this.
Also, the $$$2L$$$ argument can be reduced to $$$L$$$ (after $$$L$$$ seconds the whole setup reverses) so you can do your solution modulo $$$L$$$ instead I believe.
You don't need treap. Sets suffice.
I used Segment Trees (after coordinate compression), which seems unnecessary as the inverse of $$$t(d)$$$ is just a linear function.
Two meets are actually $$$L$$$ apart, as you noticed. Modulo $$$L$$$ should work.
Could you explain the ideas/observations behind your solution with Eulerian cycles?
I understand that for each vertex of even degree, we want to have an equal number of adjacent edges which are colored 1 and 2, and for a vertex of odd degree, we want the counts of adjacent edges of colors 1 and 2 to differ by 1. I can see how Eulerian trails/circuits could be relevant here, but how exactly did you make them fit in? Thanks!
To minimise the global cost you can try to minimise local costs. This is not always possible. We will however try to obtain local minimums. Local minimum is obtained when a node of degree $$$d$$$ has $$$\frac{d}{2}$$$ white edges and $$$\frac{d}{2}$$$ black edges. For odd degree, one of the two colored degrees has to be one bigger.
This can lead us to the think about Eulerian cycles. If we have an Eulerian cycle of even length, we can alternate the colors to get every color degree to be precisely half the degree. This happens because for every incoming edge, the outcoming edge will have opposite color, thus perfectly balancing the white degree and the black degree.
What about odd length Eulerian cycles? We can apply most of the same argument, except for the "starting node" that will get two extra white edges than black ones. However, if we have an odd length Eulerian cycle, that means that $$$M$$$ is odd. If that is the case, you cannot achieve local minimums for every node, one will be sacrificed. So you should select the node that gives the least increment. That can be computed to actually be the one with least degree.
Ok, but what about non Eulerian cycle graphs? Let's assume we have $$$2K$$$ nodes with odd degree. Then we can create $$$K$$$ imaginary edges and reduce the problem. Now just find the cycles and compute each one as mentioned above (almost).
This is fine, since if $$$M+K$$$ is even, then we have local minimum everywhere and when removing the imaginary edges we still do. However, if $$$M+K$$$ is odd, then this might not lead to an actually minimal solution. To do that, we just want to start alternating from one of the imaginary edges. This way we don't lose minimality for any node (the only node that is not minimal in the imaginary case has odd degree normally and we remove one of the excess edges).
To recap:
Add imaginary edges so that every node has even degree. Split the graph into disconnected Eulerian cycles. For a cycle: if it has even length, just alternate the colors; if it has odd length but contains an imaginary edge, alternate starting from an imaginary edge; otherwise, (if it has odd length and no imaginary edge) alternate starting from the node with least degree. Now compute the cost only with normal edges.
It's not necessary to pick the least-degree node in the odd-cycle case; the additional cost is always 2.
You might be right. Thank you for the clarification. In my code I just wrote the formula for the increment and did not care to simplify it. I assumed it would reduce to smallest degree
I like Problem D: Plan Out, especially in a you-only-submit-once contest.
.
What is the solution for the Problem D: Plan Out?
I solved it looking for a Eulerian Circuit.
Consider the graph with ai-bi as its edges. You have to label these edges 1 or 2 depending on which day you want to put it. Labelling adjacent edges with the same label is bad so it has to be avoided as much as possible (the edges incident at each node is desired to be split evenly between the days). Consider the block cut decomposition of this graph and consider only the biconnected components. For each biconnected components consider its ear decomposition. It is possible to add alternating edges by looking at the end point of each ear. Finally, for the bridge edges you can do the same.
I was trying to solve problem A, but when running the code on the testcases it took 4m29s! like I know my laptop is a 13 years old piece of junk but I should be able to optimize this in someway right? help me out if possible. remark:I had only 30s to submit the output but after submitting it the site gave me "Something Went Wrong" and GG I couldn't submit anymore, let's just say the entire protocol of the contest doesn't meet the hype of the contest at all, and I'm saying that because the whole registration-participation systems are completely ass.
Can someone please help me resolve the submission issue on Meta Hacker Cup? I’ve been trying to submit my solution for the past three days, but I keep getting the same error message: “An error occurred on our side. Please try again.” I’ve also tried logging out and logging back in, but the problem still persists. I’m attaching a screenshot of the error message for reference. Could someone please tell me how to fix this issue? Your text to link here...
Is it just me?
Not working mate, what is wrong? Even requesting clarification fails. Already done hard refresh.
no. i too am facing this issue
Me too
I tried safari and it worked!!
Tried on three different browsers and two different networks already. SecondThread is your technical team aware of this issue?
I am facing the same issue; using chrome!
got the same problem in practice, somehow worked with incognito (suffering from light mode)
Had the same error in Safari, switching to Chrome fixed the problem
why is the website so finicky
Wasn't able to submit my solution for C due to this error message "Oops, an error occurred on our end while processing your submission. Please try again." Tried several times before my timer ran out and got this same error message every time.
Actually, the submission page shows the code/output that I submitted, but the leaderboard shows 0 points. Hopefully this gets fixed so that I can do round 2.
Update: leaderboard correctly reflects my score, looks like it is fixed
It happened with me too. I got error on the website and because of that my 6 minutes timer for Problem A2 got over and now I am unable to submit the solution. Please look into it!!!!
why is it so buggy lmao i cant even download validation input
Kept getting "an error occurred from our end while processing your submission while my submission was pretty innocent in the sense that it was in the expected format" in problem B2. And now it timed out. What is this behaviour SecondThread
How to submit my B2 ? It keep get error until 6 minutes timer finished.
Was unable to submit B2 because of the error message. Tried multiple times, but the 6-minute window got closed. SecondThread
Please try to make the contest work before fucking hosting it again. so many participants including myself weren't able to submit a1 for like 45 minutes or something and even then I had to use incognito to submit. this round feels like a gpt battle. pay to win if you've a pro gpt. the system feels like as if an unpaid intern was forced to build that. I'm raging and for obvious reasons.
THIS IS THE SHTTIEST CONTEST I'VE EVER PARTICIPATED IN MY LIFE!!!!
then you've never seen an actual shit contest
Hey , even after submitting first 3 questions. My score shows 22, but sum of scores of first 3 problems are 32. Can anyone explain why?
Same
Same
Same
let me download validation input for D TT
Have been continuously attempting to submit a clarification for almost 10 minutes re: B2 since my submission did not go through. Even the clarifications are broken
Anyone else lacking permissions to chat in the Discord server?
My rank is falling so fast 2.5 hrs into the contest. I went from around 1000 to now 4000. Hope the plagiarism check is strong ;_;
same solved 3 in 25 mins, then just saw mine ranking go down
wow 25 minute is fast... did you binary search in A2
Dont tell me thats wrong :)
no no I also did binary search but that idea came a bit late to me !! ...
i was thinking if there is some other simpler solution which is superfast to code coz of your speed.
good luck for system test
thank you, had a stroke for a second
whoa !! apologies
i did it using a weird dp, not too sure if it's correct
oh very cool ... i can kind of feel why that might be right .. thanks for sharing your approach
can you please post here if it gives wrong answer, thanks
just checked, it worked luckily lol
wow !!! congrats .. thanks for update
I think i overkilled A2 by creating a graph with edges between 0 to every i and each i to i+1. Then converted the graph to MST and printed the max weight in the MST as answer.
i was having exactly the same idea lol ... but we need to be careful that we touch the ground also .. which i guess is your node 0
but i was not able to fully work this idea, got confused with proving it to be right. like the tree will be MST or not
it will be
I can imagine how Prim's work for A2. I'm not sure about Kruskal's though. Can you or anyone explain how that will work specifically? And what property of A2 makes MST's largest edge the answer — I never heard of MST keeping largest edge small.
i did same lol
A simpler idea works (for A2). For every "hillock", choose the second largest ladder. Then compare all such chosen ladders (with some subtleties / edge cases). Worked for me in the validation input at least.
A simpler idea works (for A2). For every hill, choose the second largest ladder. Then compare all such chosen ladders (with some subtleties / edge cases).
ok, thanks for sharing... will think about this to see why this works.
Meta GPT Cup
I know that loads of work goes into these things but that was poor. A lot of frustration and wasted time due to technical issues server-side.
How long will it take for them to reveal the results for Round 1?
few minutes
i guess might be delayed !!! ... they asked people to manually submit solutions so it is going to get delayed !!!
I am just near the 5000 mark though .. have less hopes :(
my code for C. but website kept crashing for last 2 minutes straight. Probably, I wasn't meant to get AK !!!
Anyone else getting presentation error for B2 even though the output format seems to match exactly
ha ha !! this image is very funny
Can anyone give some hints how to solve B2?
Fix the number that's gonna be after $$$N$$$ steps (there are $$$\mathcal{O}({\sqrt{B}})$$$ different options; consider only those that are $$$\leq A$$$) -- for each one, you know what count of every prime factor of $$$B$$$ is going to be in the left, and how many in the right part. Use some combinators to calculate the number of ways for that to happen.
It has lot to do with combinatorics.
First factorise the final value 'B'. Now, we need to spread these factors across 2*N places while binding to constraints.
Constraint: first half should be less than 'A' . So lets find out all the divisors of 'B', which are less than 'A'.
let's say there
ksuch divisors which divide 'B' and which are less than 'A'. call them [ d1, d2, d3 ... dk ]so, first N places should spread factors of d[i] ( for each factor we need to iterate this ), and for second N places we need to spread factors of B / d[i].
This is standard stars and bars related problem. there is small catch here, N is pretty large, so we won't be able to calculate nCr() with our standard way,,, instead we have to use little trick ( although, that trick is also very famous over 1900 rating, so that's all ).
Oh..hope plag check is there because too much submissions for B2, tried my best but was not able to solve
Will obvious cheaters be pruned before the leaderboard is finalized?
How long does it usually takes for the result of round 1 to come out?
Impossible to say. In smoother-running contests in the past, the leaderboard has updated automatically straight after the contest. However, since the servers were clearly unable to cope with this contest, the organisers now have to process presumably a lot of 'clarifications' which are actually legit submissions. No way of knowing how many, and no way of knowing whether there's an intelligent way of doing this, and no way of knowing how many people are working on it, so it could be hours, maybe even days.
How to do D?
Reverse the string, replace B with ( and A with ). If for each ) there is a ( to pair with, then Bob wins simply by taking the paired bracket. Otherwise Alice wins, since she can do a turn and make the state of the game a correct bracket sequence.
Also could be easily guessed and checked with exponential dp, my code was just checking if there's a suffix where the number of As is greater than the number of Bs
Had the same idea, just stated differently to give my intuition behind it. Basically, if you have substring AB, you know that Alice will only eat that A if forced (because it guarantees Bob's win). Similarly, Bob will never eat that B unless forced to. So if we just remove that occurrence of AB we get string with same answer. Recursively applying this rule we get string that has all Bs before all As. If A is the last character, Alice can eat it and win, otherwise, there are no As and Bob will win.
Neat! I couldn't see how would someone come up with bracket sequences without already knowing the solution (and I guessed it with exponential solution, found the proof only after 30 minutes). Now I see how someone could've proven the solution without guessing
The first observation was that if the string ends with A, Alice wins. Then with strings of the form AAA..BBB.., you need more As than Bs. This gave me the idea of suffix count. After this, it's easy to prove that, for Alice, choosing the first position where this cnt becomes > 0 is the winning move.
These are trivial cases, but when there are more complex combinations of As and Bs like AABAAABBBAABBAAABBB it's not so easy to show without the idea of bracket sequences which comes out of nowhere if we're only talking about these trivial cases
But if Alice removes the first position where the count is > 0, for any subsequent move Alice can choose the first A and there is no way for Bob to choose a B which makes #A <= #B
Is it correct that "first" means first suffix from the right direction (shortest suffix)? Reasoning is that after A's move, whatever move B takes, makes the remaining string more A than B.
Yes
this would be the correct implementation.
can you please explain the intuition behind this
i couldn't get how bob will win by picking the paired bracket
and how to solve B2 for a<b
if a>=b we can place any prime factor along with 1's using stars and bars but for a<b how to pick prime factors such that their product stays < a
How to solve $$$B2$$$ ?
here are my ideas to solve problems $$$A1, A2, B1, C, D$$$
for $$$A1$$$, just print the maximum difference between any two platforms heights
for $$$A2$$$, my solution is O(N * log(1e9)), I'm not sure if it will pass or get TLE, my idea was to binary search, and run a BFS from the platforms that can be reached from ground, if I reached all the platforms, return true, otherwise return false
for $$$B1$$$, just print (n * 2 — 1) 1s, then print $$$B$$$
for $$$C$$$, the summation of all the subarrays lengths is $$$n * (n + 1) * (n + 2) / 6$$$, we will maintain a prefix xor, and for each prefix, we will decrease the result with the summation from $$$1 ... x$$$ where $$$x$$$ is the number of occurrences of this prefix before, then update the prefix
for $$$D$$$, maintain a suffix sum of $$$B$$$ s, and for each consecutive $$$A$$$ s, see if the length of this segment is greater than the suffix sum at this segment or no, if yes, Alice will win, otherwise, Bob will always win
I hope not to get FST with some of these solutions xD
for something like:
AABAABAABAABAABAAB....AABBBB
Alice can win, yet the number of consecutive A's is never greater than the suffix sum... (a smaller case is something like "AABAABB").
By the way I did the exact same solution but discovered this case after downloading the input :(
My solution for B2:
The number you reach after N turns can be any of the factors
FofBsuch thatF <= A.Then the product of the numbers in the first N turns isFand in the second N turns isB/F.So precompute for all the factors
FofBhow many ways you can spread its prime factors over N turns (use stars and bars for repeated prime factors) — call thisQ(F), and then iterate over allF <= A, multiplying each byQ(B/F)and then adding to the total. With modular arithmetic this can get a bit messy, and care is required.Thank you for the nice problemset!
Are nlogn solutions not supposed to pass for C or is my laptop too slow?
your code might cause problem due to default stack size being too low.
My N log N solution passed within 10 second or so
No, actually the problem turned out to be that I was using
-D_GLIBCXX_DEBUGcompilation flag and apparently it has massive overheadWhen will the testing happen?? Can I get a rough idea about when did it happen in past cups??
I have asked about that in the clarification section, it goes like:
It may be another 24 hours before the scoreboard is ready.
Best,
The Meta Hacker Cup Team
Meta hardly check the submitted code so i doubt they can remove the GPT solutions of which there are thousands,there is no point in participating in Round 2,i'm just going to skip it.
.
I highly doubt that there were thousands of AI-generated solutions, even though there obviously must be some cheaters. Problems A1, A2 and B1 were so rasy that it's not surprising that so many people have 32 points.
Btw seeing the AI meta, i doubt i will able to win a tshirt this year :)
Thanks for updating my ranking after i submitted my clarification request :).
Given there were less than 15K users, why has the system failed to respond to requests? It shouldn't be handling lots of traffic, they probably don't care about this competition anymore, since there was not even enough server capacity allocated for it.
AI solutions are bothersome for harder problems, people can at least have an initial guidance of how to proceed, perhaps some effort is required up to making it work flawlessly.
Is this the end for online competitions?
Obvious Cheats:
Is this based on you already knowing them or something else? Coz we can't see the submissions right now? just wanna know.
.
honestly a1, b1, a2 were doable for newbies/pupils/specialists but not anything more. There were truly many cheaters , though a1,a2,b1 were easy solves , that's why their numbers are higher than usual i guess.
I'm not talking about a and b. I'm talking about people rated < 1500 who solved every problem (all killed) in under 2 hrs.
I'm sure there are cheaters who only solved a and b, but i've only listed egregious cheaters who are blatant.
Nobody goes from running a 8 minute mile one day, then the next runs a 4 minute 15 second and beats the top runner in their college. Same with competitive programming, you don't go from getting 8,000th place in division 3 one week to placing top 100 in the Meta hacker cup beating math PhD's
The winner of the practice round: https://codeforces.me/profile/SinhaSatwik
.
There's no way to validate apart from monitoring tools. Chess probably only evaluate if you're making similar moves to how a machine would. If you occasionally play random moves, you're indistinguishable.
What if I generate a code and change it to a random thing that looks more human? If I solve only problems close to my rating, it looks quite real.
I can also work on my free time to do the same and increase the rating here.
.
that's so obv , also i guess all of these cheaters learned' pollard rho' today .
I created a Codeforces account with the handle "inksamurai" years ago for practice and testing purposes, and I often use this handle.
.
you should have also tagged them/their ids
Having my name there without any proofs makes no sense @based , I solved all my problems myself. If you wanna check, I have solved over 3500+ problems over the last 2 years , Never been skipped. So I request, remove my name.
.
hello based, although I appreciate you finding cheaters I only cheat on your mother.
on a much serious note b1 was just printing 1s and b I think its a 800 rated question,A1 was a 1200 A2 was a 1600 If I am being nice. based if you were a little more diligent you would have been nice enough to go through someones profile and see what questions they practice and their submissions.
.
He's neither 1400 below nor top 1500 in round 1?????
iitprep got caught in a cheating scandal last year itself
What do you mean ? You literally wrote iitprep3 in your list of cheaters
yes, because he was caught in university last year itself
This still does not answer the question???? How was iitprep3 part of this list in the first place. He is above 1400 rated on cf and his rank in round1 is currently 5000+?
.
I did not even read your list , i do not think any of these guys are my "classmates' , i just defended a false baseless accusation that's it
.
you aren't seeing the question level and pointing out me as a cheater even I'm out of 5k rank how can this justify and how rating below 1400 proves me as cheater(does this means I've to cheat now to increase my rating above 1400 and mark myself as legit contestant :( ) please check once and remove my name its a request.
Why i am added to cheating list? A1,A2,B1 i did in 30 mins. C took some time. Then just considered worst case a ans. then tried to reduce the ans using prefix mapping technique. didnt did B2 + mine D is wrong....... it would been better if you would have done this after final verdict.
at the time when you wrote this comment, ranking was only on basis of submissions, not correctness/optimality of those submissions. submitting a problem (whether correct/incorrect) does not make someone a cheater. It might be possible they hardcoded the validation input test case to get AK. I would encourage you to reconsider your list now that system testing has been done.
.
but I know few people on the list who hardcoded initial validation testcase or submitted sub-optimal code to get AK. Now their code has been rightfully marked incorrect. Isn't it unfair to them?
How do you propose fairness then?
by not publically shaming and blaming users of cheating without any concrete or deterministic evidence. That's the least we can do to ensure fairness. Most countries follow "Innocent until proven guilty" by law, then who are we to disregard this?
This is an endless war we must fight!!
C looks somewhat interesting. How to solve?
Let $$$X(i, j)$$$ be xor of all elements within i to j.
Hint 1: If $$$X(i, j)$$$ is not 0 then it's impossible to make all elements 0. Xor of a subarray is always the same regardless of operators.
Hint 2: $$$X(i,j) = X(1, i-1)$$$ ^ $$$X(1, j)$$$
My idea was that if the xor of a sub-array is 0 then we split it into chunks such that each chunk xor is 0. We save one operation per chunk. So, the answer for this sub-array will be
length of sub-array - max. no. of chunks. This is $$$O(N^2)$$$.We can get a better solution by finding the answer for all sub-arrays ending at
iat once. Suppose, we are processingA[0..=i]currently. Find max.ksuch that a suffix ofA[0..=i]can be split intokchunks, each having a xor of 0. For example, for[3, 1, 1, 0, 2, 2]kis 3 for suffix with chunks[[1, 1], [0], [2, 2]]. We ignore sub-arrays ending atiand starting atj < 1.Now, chunk
[1, 1]will be present in all sub-arrays starting at $$$j \in {1}$$$ and ending at $$$i$$$.[0]will be present in all sub-arrays starting at $$$j \in {1, 3}$$$.[1, 1]will be present in all sub-arrays starting at $$$j \in {1, 3, 4}$$$.The answer for all sub-arrays ending at $$$i$$$ comes out to be: $$$\sum_{j=0}^i (i-j+1) - \sum_{t=1}^k t$$$. The first part is the cost for all sub-arrays ending at $$$i$$$ without doing any operation. Second part is how many operations we can save.
$$$k$$$ can be found for each $$$i$$$ using DP and prefix xor.
Let me know if this solution is wrong.
When will system tests finish? I hope I get to know if my solutions are correct or not before they start the plagiarism checks, as that can take a long time. Only B1 requires a verifier, rest can be checked using token matching.
Additionally, did anyone stress test their solutions? Not knowing the verdict immediately after submission is new to me. It's very stressful.
It’s surprising to see that cheaters are the ones complaining about cheaters. I don’t see any top coder doing that. Obviously, a cheater rated around 1500, who solves all the problems by cheating, is only upset because another cheater rated 800 also solved everything the same way.
Using GPT-generated templates for segment trees, DSU, or modular arithmetic isn’t really cheating — it’s just using standard tools.
There’s no real way to stop cheating. So just solve problems for fun. If not, then maybe it’s better to move on. I’ve already stopped caring about ranks.
Please, Meta — there shouldn’t be any T-shirts or prizes except for onsite contests (you never know if reaching the top 25 is done fairly). No prize means no stakes, and without stakes, I don’t see why cheaters would even bother.
Ultimately, these programming problems are just toy puzzles. Research and theoretical computer science will always matter more. I doubt any world-class researcher would spend time competing with toy puzzle solvers.
The fact that you've used GPT for the above message
Obviously, don't call it cheating.
I can see a lot of anti-cheating sentiment — cheaters are downvoting me. AI is incredible; if it can solve problems, there’s no point in solving them just for the sake of competitions. Some blue and purple coders have an issue with this because they can’t reach red, and now everyone can reach purple with GPT.
Let’s be honest — even before AI entered the scene, people used to get help from friends to climb the ranks. For example, many Indian Googlers @Rubalien cheated in contests to boost their credibility and later did the same in online interviews.
It’s shameful but let me be honest. > 90% cheaters are Indians
I won't cheat on you though
Nobody would coz I won’t give a chance
mommy please >///_///<
What is the purpose of this website if you remove the competition? Those who already study for the sake of learning doesn't care about it. Most people here love to compete for rating, but that's not going to last long.
Look at a sample for what AI can already accomplish:
What if the green guy here grows up to red? How do you check if I had studied or just used an AI tool?
do it yourself and become red!!
My brain is unfortunately not that good, AI is going to study for me though...
what color are the tshirts this time?
My intestine is a bit blocked so far, will get back here with a preview soon.
It should match the quality of their servers, also their ethical choices on their hiring process. It seems like every employee is receiving 6 figures to keep practicing this behavior I'm eager to do. Maybe that explains the interview rejection.
Everyone said everything about the website, cheaters and other well-known issues. Let's talk about the real issue of HackerCup — grossly overestimating the amount of interest in the competition (which makes it even more ridiculous that the servers can't handle the demand, but that's irrelevant). There are 8400 non-zero scores. Alright, maybe there are some zero scores from people who tried to participate, we can even use 14000 registrations as an estimate for the number of participants. Why spend an entire round to cut it down to 5000? Wouldn't it make more sense to split the eliminations more evenly (in geometric progression)? Something like 15000 -> 2000 -> 200 -> 25. It would allow for smoother difficulty distribution in each round.
Since you’re at the caliber that AI cheating rarely affects you it’s a non-issue for you. However within the next few years you probably won’t even be able to rank top 2,000 to advance to the second round without cheating detection. I guess maybe you still won’t care, but at that point why even waste your time “competing” when you’re facing a chatbot
Why are you crying so hard, stop participating. You also cheated you are just frustrated because others cheated too. That’s the difference, have you ever talked to any researcher… I have talked to red coders and even they themselves say, CP is heuristic and research is completely different from solving hard toy puzzle in a time constraint. Only a mediocre coder like would think CP is something to accomplish a big milestone.
I have talked to researchers and I also work at a FAANG company. Many CP problems are based on cutting edge research problems of the past.
Nobody is claiming it’s the same domain, but to say all the problems are trivial is a bit naive.
Can you mention even one… how solving it is equivalent to research? What is this FAANG thing? Does it give you a credibility to talk non-sense on codeforces. I bet you are an Indian. They use FAANG and then spit out non-sense. Of course they are cheaters and of course they complain about cheaters too.
There are many such examples, maybe the most famous one is IOI 2017 — Wiring and the paper "Efficient Many-To-Many Point Matching in One Dimension" by Justin Colannino et al. Note that this problem had already appeared in other competitions, some of them as early as 2007, which is the same year as the publication date of the paper.
Why would researchers use this platform instead of doing the research themselves? If it's valid, then publish a paper with the details of their knowledge.
This was supposed to be a competitive tool to rank people based on their skills, now that we have AI, that is not valid anymore. Anyone can fake results and achieve higher ranks.
Exactly what I meant. It’s silly to assume that CP will help them in research. There is no competition, people can still solve it for fun. Now there is no point of rating or ranking. And crying and expecting somebody to not use AI is just bullshit. Indians used to cheat before AI. Its hardly unlikely they won’t cheat now.
How do you define fun by yourself only? That used to play a core part of my studying stimulation. It seems like companies also can't care less about complex data structures.
I have seen it once, AI models can probably solve medium questions on LeetCode and will eventually solve the hard ones as well. Everyone learned it from some source, it will do just the same but without excuses for being gray.
It's impossible to validate if someone has changed the generated code to look more realistic. It may look irrelevant since the candidate is not improving, but it is what it is.
Not sure why people are upvoting Um_nik's post, as he's basically advocating for cutting off Round 1, which is only interesting to LGMs that feel "bored" with Round 1 problems (and also run the risk of getting out-typed by faster but weaker competitors). I do appreciate LGM efforts in the past to improve round 1, as we went from a pseudo practice round to an actual competition.
For the majority of the CP community, and especially anyone outside of CP, Round 1 is actually the most interesting and fun round. At my level, I enjoy round 1 and round 2 equally. If MHC is going to grow in participation, it'll grow from Round 1 problems.
It is also important to note that the gross number of participants can be very misleading. If participation went from 8400 to 16,000 or 30,000, most of the additional participants would be weak. The skill in MHC is super top heavy. Modifying the contest in the interest of that top heavy crowd that will show up anyway won't grow CP and won't grow MHC.
But the gross amount of participants is not actually (currently) increasing...umnik is arguing for round 1 to be more selective (choose lesser people — 2000) instead of removing it.
Google Code Jam, which everyone benchmarks for corporate contests as the gold standard, had three round 1 contests. If one engaged in “metrics”, one might suggest that half the problems Google created for Google Code Jam were Round 1 problems.
The path to MHC growing is multiple round 1s. I don’t see how taking a single round 1, keeping it at a single round 1, and then cutting the pass threshold from 5000 to 2000 will accomplish anything.
why do you keep saying this is how MHC will grow? even if it is true, it has not currently "grown". In the current set of ~10k participants, it makes no sense to have a round cut by 2x, when the round 3 cuts by 20x factor.
I am not an insider, this a purely outsider view: MHC was almost cancelled like Google Code Jam was when budgets got tight, but a ragtag group at Meta believed in a different outcome. The fact that they announce only a week before the practice round that MHC 2025 would happen shows how fickle the current situation is (while Meta stock price at all time high), despite the event originating in 2011!
Although you may view the round as getting cut by 2x, what is actually happening in round 1 is that the world's population is getting cut to 5,000. The harshest cutoff, when viewed globally, surely is entry to round 2. ;)
no, most of the world does not want to participate. Only 10000 do. It is important to keep that in mind when designing rounds, instead of such nonsense.
How do you know abt the wants of so many ppl ?
finally decided to reply to such takes, I see no relation between this and having round 1 as it is currently. MHC is not getting extra participants due to having easy round 1 (clearly). Thus, it should adapt to the set of participants it has.
The monetary constraints and having easy round 1 are definitely not connected.
Also, lets stop pretending like MHC is such a great gift. Most of the rounds are worse than regular codeforces rounds, and then all that is left are some prizes for Top 10, and some t shirts.
MHC is a great gift, and the only people who don't see that, from my perspective, are LGMs.
I looked through Round 1 standings (ignoring FST), and it was not smooth sailing for a lot of veteran CF folks, including GMs.
I disagree, I think MHC is a great gift. From my experience, it is the one competition where people compete annually even after quitting cf. Secondly, it is a prestigious competition where you can show your results to companies and distinguish yourself from the crowd. Third, and of course, most importantly: shirts. For people who don't win prizes regularly, meta shirts are a source of happiness and pride. It might not appeal to some people, but there are definitely people who love Meta Hacker Cup.
Note: My thoughts are totally unrelated to the discussion. I just wanted to object to your idea about MHC not being a gift.
Maybe because I'm low-rated, but I love MHC and think of it like a gift.
I would be extremely sad if seeing MHC being stopped.
To me, MHC is more like a annual huge (I know it's declining) CP event while CF rounds are just some normal training practices. Therefore, I would still enjoy the vibe of MHC even after not doing CF rounds (intensively) many years.
It's hard to see it as an annual HUGE CP event when normal CF rounds attract way more people.
I think it used to be? Nonetheless, I agree it's no longer "HUGE" especially from the number of participants and how they publicize this year.
A unique T-shirt on MHC is definitely better than nothing on CF.
If you can't distinguish between real people and AI bots, you will remain with some close people group that trusts each other and finds it funny to keep repeating their habit.
This kind of online event lost its purpose, the prize used to have a meaning attached to it. That is not valid anymore.
Who told you it was about budget? They are one of the most profitable companies in the world and can't spare less than 1M to host some servers and allocate some employees times?
This kind of competition became as useful as solving jigsaw puzzles. At the very best, it's a great achievement with little to no relation with work.
how long till rankings are finalized ?
I don’t know what Meta is smoking, why the system testing is still pending.
I am not able to download test for practice.
The Meta Hacker Cup team seems to be on a Diwali break — system testing will resume once the festivities are over. So for those waiting eagerly… might as well enjoy your Diwali too
Happy Diwali, Everyone
I'm afraid this is taking longer than we expected. We're making progress, but we don't have an exact time. I would expect at least another 24 hours.
Best, The Meta Hacker Cup Team
Thanks for updating my rank after I submitted a clarification! I went from 2,800 to ~500th!
I'm pleased to see this ridiculous take being downvoted.
Why are you happy this is being downvoted, I couldn't submit during the contest, so they accepted my submission after considering everything
That isn’t what’s happening. People submitted clarification requests live, during the contest, with timestamps to prove they did what they did when they said they did.
That's exactly what im saying
It was literally announced during the contest by means of a banner notification instructing participants on what to do.
And if someone uploaded their output and source code as a clarification at HH:MM, then obviously that’s the timestamp they get. I would have thought that was obvious.
How are you going to have time to submit a clarification within a 6 minute window? I submitted mine as soon as i saw the banner (which was after my timer expired)
i dont care about results anymore, just tell me the verdicts of my submissions, I understand you want to do plag check, include clarification submissions, but its not so hard to public the results of those who submitted their solutions normally.
st
how much more time for the standings to be finalized , any idea anyone?
I missed the announcement, but will still warn everyone to not participate in this scam as you won't get your T-Shirts anyway.
"your t-shirts"?
I never understood the fad around t-shirts and stickers. Does it hold so much value to people that they start cheating for it? If they do cheat to get it then should it hold any value to them at all?
You need to place in top-200 in round 3 to get a T-shirt, but Meta scams you by not sending an announcement when T-shirts are available to order for a week or so. As result you don't order the T-shirt that you fairly won and Meta saves $10. This happened to me multiple times already
On an unrelated note I totally agree that participants who cheat should not get anything, but in this case Meta scams participants who won T-shirts fairly.
I never finished top 200, I have two T-shirts, and both times I was prompted by Meta to order them.
I actually received all of my 3 shirts for free, just had to use a coupon to bypass any payment requirement. Still, you have to access your e-mail to open the link and order the shipment to your address.
I kind of agree that with the sentiment that they could just ship them if they have an address on file. But also would probably result in X% going to old addresses and I reckon most people do check their email.
They just don't send emails, last year it was announced during the finals livestream, which no one is watching.
thank god, this was a relief for someone participating first time. :)
I think you missed a zero there, it's top 2k that get tshirts. Top 200 or so get special tshirts.
I wouldn't call this a scam, but I agree with the frustration. I won shirts in 2021, 2022 and 2023, I got the first two, but did not get the third one, as I got no Email Notification and then the deadline already passed and I was quite bummed, too. See also my message back then here: Link to my complaint.
So, if I were to win a shirt this year I will be checking weekly :P
I recall it taking ages. But it’s also free, so I’d hardly call it a scam (re the original comment, not yours)
Yeah, I usually place in Top-200. And it's a scam, because it's actually takes a lot of time to participate in 3 rounds and to win a T-shirt you need at least 1-2 hours per round.
I mean, if you’re viewing the T-shirt as recompense for your time invested, I’d suggest you don’t bother. Just go and buy a T-shirt and spend your time doing something else.
T-shirt is a reward for doing well in a competition and when Meta scams people by not sending an announcement to order T-shirt it feels like a scam. 1 hour of my time definitely worth more than $10, so I was not doing this competition just to get a T-shirt.
Another rewarding result which would compensate for T-shirt scam is winning the final round, but unfortunately I'm not good enough to win.
Did they reject you at job interview? I’m struggling to believe all this invective is because of a T-shirt.
It's just a bad organization when people don't get notifications about T-shirts availability. I'm trying to bring my point to organizers so they can do better going forward, but they have much bigger issues each year with problem submissions, so maybe when that part is sorted out they can get to T-shirts.
Why do you participate if you never reach the finals? Quite bad website for practicing if that is the reason.
I find this question very odd but I’ll humour it:
For a combination of enjoyment and educational purposes. I took up CP as a lockdown hobby during Covid. I learnt loads and made it to orange. I was not quite good enough to reach red. Perhaps with years more effort I could have, I don’t know, but this was a secondary aim behind fun and learning.
I understand the goal of learning to have fun. But to join Hacker Cup meant something else, otherwise you would have stick to Codeforces only.
A T-shirt used to mean a physical representation of an achievement, it's the only thing different than here... If the rank is enough, what is the purpose of participating there instead of on Codeforces?
The fun of a different format? The higher profile, the rounds structure, the fact that the problems are often fun too?
result published everyone
Why I solved question A1, A2, and B1, but only had 22 points? Did anyone have the same issue? You can search my name, "Bao Hoang" in the scoreboard to see that. Accepted A1, A2, and B1, but only 22 points.
looking at your penalty, it looks like it only accounted for your A1 + A2 solves as well. will require asking for a clarification probably.
They did not reply to my clarification. What should I do in this case SecondThread? I solved question A1, A2, and B1, but only had 22 points.
I solved 3 problems A1, A2, and B1, but only got 22 points. Here is the Google Drive Folder for evidence: https://drive.google.com/drive/folders/1aPAVN2W5FyPH6SnDC7ALqB_5pgB5RRnA. I am not sure why.
Too bad they didn't catch any cheaters, My clarification request brought me to top 500, but if they actually did a cheat detection i'd probably be top 300.
I went from 187th to 133th place, there were some cheaters they caught :)
This kinda proves my point though tbh.
So you’re saying ~40% of the top 200 competitors were cheating.. but from rank 200-5000 they only caught < 3% of people.
So cheating is only a problem when it effects top competitors!?
The problems were pretty easy for the AI so it would seem logical that cheaters would get high
.
bro you are being too much with your comments?
a legit person would never be so desperate about such things.
.
same happens at my college. If i could possibly do something about it, I would. But you see we can't really do anything about this at this point. so we better wait.
They may be FSTs, are you sure they were cheats?
No, I'm not sure. I forgot that problem D had weak pretests and many people failed on it
How do you plan on catching cheats apart from cheap gossip? If I create an account and cheat from the very beginning, you will never notice it. At best you're behaving like those people's parents.
.
So, if I adjust the code style to follow a pattern and wait for tourist to pass a problem, I can bypass your rules. There's no need to be very smart for doing such a thing...
should have participated in the practice round, got left out just because of wasting time in submitting my answer ;/
Hi, this is my first time doing MHC, so can someone (preferably who has done it in prior years) explain the process after qualifying for round 2? My place is <5000 on the scoreboard, and I haven't seen any new button to register for round 2. Do I need to wait a few days for registration to open or something?
After the scoreboard is finalized (there currently appears to be glitch where some contestants haven't received credit for their submissions even though those submissions show as accepted; the organizers may also still be in the process of e.g. removing cheaters from the leaderboard), participants who qualify for Round 2 will automatically be registered. There's no further action you need to take beyond showing up at the start of the round.
Geothermal I solved A1, A2, and B1, but didn’t receive any credit for B1. I believe I’d be in the top 5000 if B1 were counted. What should I do? I sent a clarification to the Meta Hacker Cup team but haven’t received a response.
I’m not affiliated with Meta, so I can’t really give you any definitive guidance. Submitting clarifications sounds reasonable. Other people are having the same issue, so hopefully they are aware and will fix the problem before the start of R2.
When will the registration start for round 2?
I'd like to know this as well.
I think the answer is "you will be automatically registered once round 1 standings are super-duper-ultra-finalized" but I don't want to accidentally forget to register if that's not the case.
Some explicit clarification from the MHC team would be nice
https://codeforces.me/blog/entry/147893?#comment-1321846
I think they are finally removing cheats/ fixing scoreboard , Coz rank's getting better :)
Just now I noticed that date for the round 2 has changed to 15th November 2025 from 1st November 2025.
There should have been an email about this lol.
It's Meta, lol :) They never send notifications about anything.
I'm surprised they moved the round forward instead of backwards. It would have been totally Meta to have a contest happen a week ago and a bunch of people hoping to win a T-shirt showing up today. Perfect scam.
Hi community!
I have a doubt that there is an issue in Problem B of Meta Hacker Cup Round 2: “Defining Prizes.”
I took some solutions that were accepted during the round, and I tried to “hack” them.
For example, many codes give a final answer of 6 for the following test case:
6 3
1 1 2 2 3 3
6 6 1
But in reality, we cannot award all 6 of them, since the merch distributed to the same person must be different. The answer would be 4 instead.
The same issue happens with this test case:
4 2
2 2 3 3
5 1
Some accepted codes give 4, others give 2, but the correct answer should be 2.
Accepted solutions have different answer for this kind of testcase.
Am I the only one that noticed it ?
No, you're not
From the hacker cup terms of service:
They score based on "input sets". Under these rules, there would be no rejudging UNLESS your source code produces a different output from what the output file that you uploaded contained
8. SCORING SYSTEM. All output files will be judged based on the following scoring system: Each problem is worth the number of points stated at the top of the problem. A competitor's total score for a Round will be the sum of the point values for all of the input sets he or she correctly solves in that Round. In the event of a tie between two or more competitors, those competitors will be ranked in ascending order according to the penalty times associated with their solutions. In other words, as between the competitors tied on point values, the one with the lower penalty time will be ranked first while the one with the highest penalty time will be ranked last. A competitor's penalty time for a given Round is equal to the sum of the Submission times of the last output that was submitted by the competitor and judged correct for each Problem in the Round (time is measured from the start of the Round). If, after the close of any Round, an alleged discrepancy is discovered between the source code and the output file for any of a competitor's Submissions that were judged correct at the conclusion of the round, a panel of two or more judges consisting of employees of Meta and/or its subsidiaries shall examine the source code for all Submissions of the competitor for that round. The judges shall determine, in their sole discretion whether a discrepancy exists, and if so whether the discrepancy is trivial or non-trivial. In the event of a trivial discrepancy, the competitor shall be assessed an additional 6-minute penalty for that input set. In the event of a non-trivial discrepancy, the competitor shall forfeit the points for that input set. In the event the judges rule that there is no discrepancy, no change will be made in the competitor's score for that input set and no penalty minutes shall be assessed.I think it's definitely unfair, but also changing the rules after the competition is unfair but in a different way (especially since the input validation was weak as well)
How to order the top 2000 shirt ? Can't find the option... Can somebody help (screenshots will be really helpful)
Usually have to wait a little bit for this, can't remember when it becomes available.
Usually after 1.5 months you can expect to get tshirts telling from my previous exp.
Hey SecondThread, is the standing for Round2 finalized? Or is there a last minute change in Round3 schedule?
does top 25 still get possibly automated invites for job interview at meta? @secondthread and others.
how to redeem Tshirt, is there any link
The profile page https://www.facebook.com/codingcompetitions/profile is broken, at least for my account. Initially the page loads, and shows "Meta Hacker Cup — 2023: T-Shirt" information, which I missed in 2023 and didn't claim. In about 10 second, the page automatically refreshes itself and shows a blank page "This page isn't available". The profile page has been this way for over a month and I assume the problem is not specific to my account.
The website is broken, it says that I have "No email address" while there is clearly one in my Facebook settings, and the organizers aren't replying... any idea how to sort that out?
I would have liked to be contacted by Meta recruiters :(
cooking the t-shirt?
searching the kitchen for it, maybe
Hi , SecondThread why India country is not available on the tshirt vendor website ? It had been there atleast for some past years.
[UPD:] type India and it is found ...
Am i the only one who didn't like the design and the color of this "limited edition." <> ?
will MHC happen this year too ?
Any news on MHC 2026?