Round 2 of the 2020 Facebook Hacker Cup is less than 48 hours away!
The round will begin on August 29th, 2020 at 10am PDT and will last for 3 hours. The contest can be found here.
You're eligible to compete in Round 2 if you scored at least 25 points in Round 1.
Everyone who solves at least one problem will receive a Facebook Hacker Cup t-shirt, and shirts for the top 200 contestants will include a “Top 200” badge! We'll begin shipping out t-shirts by early 2021 or earlier, at which point the winners will receive emails with tracking information.
The top 200 contestants will also advance to Round 3, taking place on Sept. 12th. Please note that all submission judgments will only be revealed after the round end, and that time penalty will be used to break score ties. More details about rules and other information can be found in the FAQ.
We wish you the best of luck, and hope that you enjoy the contest!
The Facebook post can be found here.
Update: The round has ended, thank you for participating! The scoreboard can be found here, and the solutions here.
How to increase stack size in windows 10? (I am using c++ in sublime text 3)
Was solving https://www.facebook.com/codingcompetitions/hacker-cup/2018/round-2/problems/B
I used g++ -Wl,--stack,268435456 -o setup.exe setup.cpp in command prompt
But still this error pops up:
terminate called after throwing an instance of 'std::ios_base::failure' what(): basic_filebuf::underflow error reading the file
You can find some information regarding increasing the stack size here, and more generally on the internet. Though it's worth noting that it's not obvious that your error is a result of your stack size.
yes got it done, i don't know what was the problem in Windows, but in Ubuntu, I added ulimit -s unlimited inside sublime build, then it was fine.
Thanks
LoneFox Can you please share the file size for each problem? It'll help me decide if I'm going to use file I/O and what I should prepare for.
Will the input have the exact same formula as last time (or if there is a new formula, is it possible for you to share it with us)? I'll template, if possible, as the last thing I want to be debugging is input reading.
Thanks!
We'll keep the size of each input file under 8MB (generally much smaller), and some problems will feature expressions for input generation similar (not necessarily identical) to those present in Round 1.
But why keep changing the input method?
I assumed the input method would be the same as round 1 (on the premise that you'd want competitors to spend their time solving problems and not dealing with input idiosyncrasies), only to be met with an extremely subtle and completely unannounced change, forcing me to debug for 10 minutes. In a competition where penalty is sum of times, this was extremely costly.
Problems may have different requirements for how to reasonably generate input for them.
Not the case here. The shift from 1-indexing to 0-indexing from round 1 to round 2 was unnecessary -- I don't see how having tables with 0 people changes problem A in any way.
And for the different modulos in problem C, you could have simply supplied the modulos in the input file, and then we would have been able to read all input arrays with the same code.
As a tester, I very much enjoyed the problems. GLHF to everyone competing!
Hello Mr. David Harmeyer,
How to get involved in Facebook hackercup's tester-contribution conspiracy?
Regards,
Aryan Choudhary
Hmmm.
Why does the link look so familiar to me?
What happened to those fun problems? Did you decide to save A, B and D for the next round in the last moment?
That's a pretty cool idea NGL, my only concern being that now it looks less cool to laymen since they'd imagine there's less than 200 ppl with this shirt.
Asian coders starting their round at midnight be like:
Nothing's gonna stop me from going straight to bed after 1AC
My plans revealed xD
But you won't know if its an AC until the contest ends! (and that's enough to keep me from going to bed ;-;)
opens Q1
No, I don't think I will.
If someone solves only a subtask, can he get a tshirt?
Assuming subtasks exist
There are no subtasks or partial marks per problem.
Input format is so disgusting(except second problem), statements are so long and hard to understand, problems(at least first 2, that I read) are so boring, so I decided not to participate in this round.
I hoped that the problems would be somehow more interesting than in round 1 and qualification round, but I was wrong.
Very upset and demotivated.
Just solve the A and get a T-shirt
Since I didn't get last 2 or 3 T-shirts, I believe that I won't get it this time too.
Last 2 problems weren't better imo
Why can't I see Petr, Gennady or Mikhail in the standings? Are they already advancing to the finals?
I wonder if Tourist, Petr, Errichto, neal and ksun are not submitting just because they know something that people who have completed the set don't? Like, they're thinking about something bigger
Also, ecnerwala made his final submission first, and yet he is shown to be behind Benq Can anyone clarify how are rankings calculated?
Penalty as sum of times is stupid.
Why? ICPC does the same thing
But ICPC is a team competition and in my opinion that mitigates some problems that arise compared to individual competition. I agree that sum of times is stupid for individual. You have no way of catching up if you were 5 mins slower on first two problems.
I think it's different because:
Overall I think the ICPC experience is incomparable with this contest, and I've never had the experience of finishing an ICPC contest in 100 minutes without any wrong submissions just to wait for another 30 minutes for people to surpass me, with literally nothing to do but blame myself for not having pre-written templates and triple-checking the input parsing. It might sound arrogant, but I'm sure I'm not the only one in this situation, to feel like the contest was very poorly thought through, with this penalty system.
I hope you have good plagiarism detector because i participated in this contest lolololo...
This was funny 6 months ago.
You gave him the attention he desperately wanted to get. Better to completely avoid such comments, he will surely get bored and leave it. XD
_
The input format is getting more and more ridiculous (eg Problem C).
Suggestion: Provide an optional helper script that converts the downloaded input into a larger file with the decoded input. This would be similar to how GCJ provides a helper script for interactive problems see here.
This would simplify the input processing for everyone, and still keep the downloaded file size under control.
LoneFox SecondThread
Thank you for the suggestion! We'll certainly consider alternatives such as scripts or encrypted zip files (at least for next year), bearing in mind that the addition of an extra step in the process and interaction with different software may add a different sort of confusion/complexity for competitors.
It was common in IPSC to give a python script to generate the input, and removing all the hassle from contestants. Also an encrypted zip file looks good too!
I think any of these strategies will improve the UX, and any new process/interaction/software will be easily learnt by contestants in a 72 hours qualification round.
How to solve D? I couldn't even solve an easier version where $$$h[ p[i] ] > h[i]$$$.
Just convex hull trick with min-max merge.
Just take the code from https://github.com/kth-competitive-programming/kactl/blob/master/content/data-structures/LineContainer.h , add some glue code, and a whole lot of trashy input parsing.
First time, in many years of competitive programming, that I boredom-quit a contest.
Problems B and C were already not the best, but problem D is on the highest level of boredom! What is the point of giving a problem about known techniques that is hard because:
A contestant who does not know the technique will never solve such a problem. A contestant who knows the technique reads the problem, sees the solution in half a minute, and then spends half-an-hour wondering what he did wrong in his life to deserve this...
To FHC organizers: Please find some better problem-setters/testers (maybe outside facebook). You are organizing one of the most important contests in competitive programming (and, I guess, you have a budget that is way bigger than the usual atcoder/codeforces contest). Don't waste this opportunity.
p.s. I agree with the replies that B, C were not "bad problems". I did not want to say that they are bad, just that they are not good enough to make me forget about D.
My solution for D was parsing the input, copy-pasting the code from https://mirror.codeforces.com/blog/entry/51275, and a nice and short actual solution that uses that data structure. So it was at least possible to solve D without implementation-heavy techniques (assuming copy-pasting the above data structure is not counted as such).
So copy-pasting data structures makes problems fun?
:) I did not say that this problem was fun, I said that it was possible to avoid heavy implementation.
May I know which implementation did you use? I had issues with swapping objects using this implementation.
I've actually used two, and added asserts that they return the same outputs: https://mirror.codeforces.com/blog/entry/51275?#comment-351410 and https://mirror.codeforces.com/blog/entry/11155?#comment-539199
I had the same issue with the same implementation (as of now I don't know why). I fixed this by avoiding swaps on the structure and instead storing
id[u]— the index of the structure used by node $$$u$$$. Then I can simply swap two numbers —swap(id[u], id[v]).Let me know if you figure out the issue, I will investigate myself.
I didn't use this implementation, but I see one suspicious thing in it: the lambda at line 30 implicitly captures
this, which is used by the call toend().When the contents of two
HullDynamicare swapped theLineobjects are moved to the otherHullDynamicwithout being changed, i.e.Line::succstill remains the old lambda. This old lambda doesn't work properly now, because it still comparesnext(y)to theend()of the oldHullDynamic(that was captured throughthis).Yeah, this is indeed true. Thanks for clarification.
I used this and passed. Some people told me that swapping the structure may not break the complexity, but I’m not that much of C++ expert..
Do you mean that you swapped the structures and passed? Complexity wasn't an issue for me, but I was getting garbage values for queries when I swapped the structures themselves.
Ah sorry. I mean I also had the same id[v] stuff because I don’t know how to do it otherwise.
I struggled in a lot of problems with this implementation. It seems like the multiplications overflow when the limits are high and then everything goes wrong.
UPD : Sorry misunderstood the problem
I believe that kactl's LineContainer is neat and efficient. It also didn't have such problems.
Also, you can use
object1 = move(object2);if you don't needobject2anymore. I think it may be better thanswap().I thought C was ok, at least you have to think a bit or draw something on the paper.
Otherwise having tree+queries problems in every round is just beyond boring.
Maybe they should change the format and do something completely different. It just can't get worse.
I thought B was nice (no ugly input and above average quality for its level).
But everything went downhill from there and you are absolutely right about D.
Good that someone liked it but B requires literally 0 observations/ideas, just a standard approach.
I found the video from D rather interesting.
Curious to know more about B solution
My solution for problem D is successfully validated, but crashed on full input...
Just to be sure, is it maybe the stack size on your machine? Some trees in the main input were quite deep.
Mine crashed even on validation input before increasing stack size
Rank 200 :P
LoneFox Please tell there won't be any changes in ranklist XD
In general you can only go up after the contest (because of people getting disqualified for plagiarism), so you should be safe :)
Rank 201 by one second. Please accept one more person into Round 3 :D
If you want to know the value of one year, just ask a student who failed a course.
If you want to know the value of one month, ask a mother who gave birth to a premature baby.
the person who just escaped death in a car accident12tqian.(c) Marc Levy
Why you have to torture us with problem like C?
the video in problem D is more interesting than all the problems (:
Consistently bad after Round 1
D: Almost same as https://mirror.codeforces.com/contest/932/problem/F
Except for input parsing.
Solved it in 18 minutes that time but not now :(
Now I feel bad for not reading D lol
The only creative thing about problem D is the pun "log-istics team".
Never thought I'd find myself as the 201st, with the difference between 200th and 201st place being only 1 second. :(
RIP Round 3
I made 6 bugs in reading input in D. SIX!!! Very glad that they show what is actually the valid input that we want to get.
https://www.facebook.com/codingcompetitions/fb-hack what is this competition? Is it a team-based CP competition or some other stuff? LoneFox
FB Hack is an event that has a CP component, and also an ideas hackathon. It's made for teams of 5 university students, and it's held locally in various regions (though of course online during COVID).
CP+Hackathon?! Wow, sounds interesting.
Everybody seems to use dynamic hull, but isn't this really easy with plain static hull? Process tree from bottom to up and build a segment tree over preorder and ech time you get a new node and want to know its value (and some queries for it) just ask about some base intervals and that's it. Only tricky thing you need to care about is that for each base interval all queries regarding it come after all insertions, so you need to build it from what you gathered up to this point first time you query it.
With dynamic hull you don't need a segment tree — simply merge the hulls of children via small-to-large and answer queries at that node. This gets you extremely short extra code (excluding parsing and copy-pasting structure, a single DFS with ~15 lines).
Well, having dynamic hull over static hull may indeed simplify things, but if somebody can't find ready-to-use dynamic hull then solution I suggested is far easier than coding dynamic hull and proves that having dynamic hull was not a prerequisite here.
Was not a prerequisite, but the problem asks "write dynamic convex hull" which makes it a bad problem in my opinion.
Need help in finding error in my dp for problem B.
Should be
if (i && j)Thanks.
Couldn't find this silly mistake during the contest. Disappointing.
Maybe next time use
if (i > 0 && j > 0)This is equivalent to
if (i > 0 & j > 0)in case you mis-spell the&&operatorAm I the only one who cannot access the scoreboard? (for both "Everyone" and "Friends") It keeps loading forever.
Tried clearing the cookies, still getting the same result. But opening the scoreboard while in incognito (not logged in) works. Maybe something is wrong with my account.
My screencast, 7th place https://youtu.be/Gpsn2P5UY_A (very fast ABC then quite slow D because I didn't copy anything)
Can we expect very generous t-shirt distributions in future rounds too?
Did anyone receive the email saying that the Round 2 would last for 24 hours, instead of 3 hours?
I planned to compete the next day, as the round started late (1am) where I am living now, therefore I missed the round.
My mistake is not checking the time again in the announcement FB post and this post. But this inconsistency costs me a chance to participate.
Edit: I cannot add the image. Link to the image: https://pasteboard.co/JoGsyHj.png
There was one more day later :)
The best thing about this round was free T shirts. I did really like problem A and B, not much implementation and stuff, but problem C was, at least in my opinion, a straight out implementation problem, with zero thinking needed (parse inputs, figure out which edge is where, repeated set operations, etc.). I thought it was just me, but even the editorial didn't offer a better approach. And from other comments, it seems the same goes for problem D. Not that I personally care as I wouldn't have made it into the top 200 either way, but its sad that in such a prestigious competition tougher problem implies implementation oriented problem.
PS :- I did like the effort in trying to make the problem statements funny.
Till which date can we expect the t-shirts to come. Really excited for it :P
It's mentioned in the post:
We'll begin shipping out t-shirts by early 2021 or earlier, at which point the winners will receive emails with tracking information.
Thanks!
why is tourist not participating in this? Just curious...
He is, I think he was exactly 1 rank above me in this round.
I have found a strange mail last night. I think this is fake. Is it from hacker cup official?
I too received such an email, though my name was present in the email instead of the ascii characters. All previous round announcement emails I got were from the same email address as this one. But your mail seems strange so I guess it maybe better to wait for someone from FB to confirm.
I see, those characters are actually my name. My facebook account name was written in my local language using some font. So, now i am afraid of my name. is it make some problem while delivery? I thought this was fake because they mentioned this prize for begin top 2000. But after that i found they use the same mail address they sent announcement mail.
If your name is correct when read in your local language it should be fine. Maybe the delivery mail will ask for the details once again.
i think it is fake because In my email, it is written about top of 1500
oh sory , after rechecking my email, that is written about top of 2000 ... sorry
I've received similar email from no-reply@jniwebshop.com with next content:
I assume this is a fraud, but it would be great if facebook could confirm that. Unfortunately I found no possibility to contact them in facebook
I received it today.
In 2019 there was also some mail with jniwebshop link, and I got my T-shirt, so I think it is not a fraud.
LoneFox