Hey, Codeforces!
We have great news for you.
Harbour.Space University is excited to announce a new contest for all interested programmers who want to start their studies in Barcelona, Spain or Bangkok, Thailand, and join our ICPC team.
The contest will be hosted on the Codeforces platform 22.07.2021 17:35 (Московское время). We have prepared 9 problems of the joined (Div1 + Div2) level. The round will be rated and open for everyone.
UPD: Scoring distribution: 250-500-750-1250-1750-2500-3000-3750-5250
For the next academic year (2021/22), we are recruiting a fascinating community of competitive programmers from the top prize-winners of international Olympiads to join one of our competitive programming teams at the university.
In the next few years, our goal is to win SWERC and compete at a high level in the ICPC globally. Therefore we want to invest a serious amount of our energy, resources, and talent into these activities.
That’s why we are inviting you on an exciting journey in the company of like-minded people, outstanding coaches, and our ongoing partnership with Codeforces.
We have already organized over 100 educational rounds, so we think the time has come to test our joint efforts and reward the most diligent.
Here’s what you win if you place in the contest:
The monthly living allowance throughout the entire duration of studies depends on the overall performance of the candidate. As a guidance, it is in the range of 500-1500 EUR (it might be applied to the contestants who win 4th-10th places).
*No application fee is required for any of the awards listed above.
We would like to say a word of appreciation to:
- Round Coordinator: BledDest
- Authors: 244mhq, bthero, Adel_SaadEddin, Zaher, Errichto, mnaeraxr.
- Testers: nooinenoojno, 998kover, Utkarsh.25dec, Um_nik, adedalic, Wind_Eagle, abunyawa, aniervs, hitonanode, Aidos, thenymphsofdelphi, ptd, ZetaZomB, rumazius, madxmad, I.AM.THE.WILL, Vladik.
- MikeMirzayanov for Polygon, Codeforces, partnership with Harbour.Space and making this round happen :)
UPD3: Congratulations to the award winners!
Note1: All the winners get an application fee waiver.
Note2: All of you eligible for Competitive Programming Scholarships (cf rating > 2000) may apply directly through our website and go through the general admissions process.
Guaranteed full scholarship + monthly living allowance + free 3-week course:
- 1 (49). Ali.Kh
- 2 (58). Yousef_Salama
- 3 (175). early-morning-dreams
A full scholarship with interview process + free 3-week course:
- 4 (184). sunyx
- 5 (200). amanbol
- 6 (203). IMRED
- 7 (219). Meijer
- 8 (226). loan
- 9 (265). kpw29
- 10 (267). Huah2
Free 3-week course:
- 11 (292). dhruvsomani
- 12 (299). adamant
- 13 (326). Kaitokid
- 14 (355). RetiredPlayer
- 15 (264). c8kbf
Good luck, and may the code be with you!
Harbour.Space University
Sir Errichto is an Author, seems Good!
Can i ask why so many downvotes?
please don't downvote this message :((
because people think that downvotes are better than upvotes :(
My rating is 1351. If i wanna become a specialist after this round, (to gain 50 rating points) 4000th place will be enough, right? (giving the fact that all div1 participants (about 1200 users) will be above me)
I think yes, good luck :)
Thank you, you too;)
i also fear of that , to be honest u will not achive today but some day sure
why not today?
because he is not consistent yet thats why
people downvote everything, just like stack overflow ahah, anyway dont take it personal
Get a higher rating and you'll get upvote — CF hidden laws No.1 -
Yes, get higher rating and stop writing stupid comments and you will get upvoted.
Like this guy.
Only 3 testers? Kinda weird O_o
UPD: LostCoder was right, more testers were added
only if they were grey
Probably new testers would be added in few more hours i think.
tourist prepared the last contest all himself. So, i think 3 testers isn't too bad
tourist counts as 3 people
Now there are 1015 testers.
Niceeeeee prizes, but absolutely they're not mine :(
Erricto means we r getting Geometry Problems. change my mind!
If they aren't A,B, or C, i am good
Errichto is one of the authors = good round
Weird terms of agreement when registering for the contest:
Maybe it makes sense to fix this? So that the cheaters don't have a formal "I didn't break any rules" excuse.
Albeit really appreciative for the CF partnership, I highly doubt this recruitment strategy will be particularly effective. Limiting the choices to only top 10 in div 1 just seems absurd and suicidal. Unless it was never intended for the prizes to be collected?
I understand that it's top-10 among those showing interest.
Yes, that would make way more sense. I misinterpreted it.
rules are broken again
Thanks, fixed.
"City where you live ot city you are representing", In social settings.
My only achievement at codeforces
And my only achievement is gone now, thanks to my new friends :(
It's not gone, you have been given a new target, chase it up ;)
Meanwhile me:
you are guys? i thought you are computers with legs :D
I looked at my contribution and my fans' number and didn't feel like saying anything :(
And then your contribution dropped :(
wait should i be happy or sad lol
well thanks for the upvotes and for my 2 new friends xD
Woah!!! Behold, Here comes the true legend
Every action has equal and opposite reaction? Btw, nice coincidence!
(/◕ヮ◕)/As a tester, I want to wish you all good luck & high rating!
P.S. Yes, it is rated.
Praying for specialist! Also orz to the authors, the testers, and MikeMirzayanov for the Platform!
Imagine tons of unnecessarily tag for Mike!!!
my bad lmao
You will get there, don't worry.Remember me?I was in ITMO camp.
yeah ofc
As a tester, I ...
am late.
Is it possible to decline the scholarship(not that i'll win it, just curious)
Hope to reach pupil. Should i solve previous contests of these authors? Would that be helpful?
Why so many downvotes? I don't understand.
Hoping to be an expert for the first time in this round. Got stuck around 1500-1600 for too long.
Looking forward to it! interested in your university. i ma still a bit young but i have plenty of time to practise for your future contests!
When, will the scoring distribution be announced ?
curious to know how did authors from different nations came together to write a round? As far as I know, CF coordinators don't accept individual problems
Hope I can become an International Master after this round.
Sad.I could become an International Master if My D were not FST.
i just wanna congratulate tourist, benq and radewoosh for getting the first three places and winning the scholarship
How in the world did you spell tourist wrong?
To avoid copy-right issues.
Is it a 3 hour round?
Score distribution??
It's been a while since I entered a contest after a time skip, I'm looking forward to it <3
The purpose of the scholarship is to attract the top-ranked coders to join the university and to win ICPC, right? So, if they don't fulfill the requirement for ICPC (e.g. age limitation), does it mean they are not eligible for the scholarship?
Good day, foolishgoat We carefully take a look at each case separately. Winning the ICPC is one of the priorities of our university, but also, we are investing in a community of like-minded competitive programmers. We have had the cases of several top-rated programmers who have been already awarded scholarships for one of our technical master's programmes
update: 9 problems in 2h 30m :)
I think this round will have best problem ever I have seen.
Score distribution?
UPD: Scoring distribution: 250-500-750-1250-1750-2500-3000-3750-5250
A 250pts task...?
Only once I saw this score
Guys i don't really understand the downvotes. please don't downvote unnecessarily :( Tnx.
I have never seen 250 pts for problem A. It will be interesting!
tourist ans Benq both are participating today and the fight for rank 1 continues!
Well Benq did it!
Speedforces coming up
So glad to see mnaeraxr as a problemsetter!
A 250 -_-
Amazing!!Contest
stringforces
Does someone ever had this feeling, that you didn't register for the contest and thought you will after you solve till C (pupil goal), and then realize you can't register anymore (for the first time)? (sad emoji). Anyways will upload in practice.
why did you have to use so much tactics??That you didn't register then waited until solving C.Just fucking register before the contest and give the contest like a winner.Don't be so timid!!
For The First Time. Also, Regarding the tactics everyone should have there own Freedom of choice.
saying a person who stopped giving contest after becoming specialist once.Show the same courage yourself if you think you can sustain specialist by giving contest.
It's a dumb comment from you.How do you know that I am not giving contests because of fear of loosing specialist position??I am very confident that the next contest I will give I will be expert.Besides,I don't consider being specialist as an achievement.For me it's just a node on the path.
A kid who took 71 contest to be specialist once is confident to be expert in his next contest.Just stop lying to yourself bro.
and people like you can only comment like these...I don't care whatever you say.3 years later when I will become red after giving 200 more contests..Please on that day comment some stupid stuffs too..I am inviting you.Till then take care and stop barking
At least i will be able to participate in div 3 round.....
Me too :) we had same ratings before contest xd..
HackForces
Just hacked some1,feeling good. xD (Only doing memes,plz don't downvote me)
Soon it'll turn into FSTForces :(
Saw a tough Fight between Tourist and BenQ...Congrats BenQ for winning.
I trolled on F because of a one char error for like 40 minutes -_- but the problems were good overall.
The observation on E was pretty nice (albeit maybe a bit convoluted?).
What is the intended solution for F? I am doing like some $$$O(n \log MAX ( \log MAX + \sqrt{MAX} ))$$$ shit and runtime is ~3950ms. Likely not going to pass sys test.
I had a solution with the same complexity, but couldn't debug in time :/
If you use sqrt decomposition with prefix/suffix sums in each block then the $$$log$$$ factor disappears because the intervals are disjoint.
I was able to do $$$O(n\sqrt{n})$$$
I used Fenwick trees to compute the sum and count of numbers $$$\ge \sqrt{K}$$$, with $$$K=500$$$ it passed systests in 888 ms.
How to solve E?
Why my E does not work?
Foreach p[i] we calculate the k if that p[i] was an unswapped one. Then we collect the frequency of all those k.
Each k with a freq>=n-2*m is a possible k.
But that fails on pretest 2, why?
Nope, you still need to do further check. Try this input:
Your code outputs
2 0 5, but this is incorrect. Can you change2 3 4 1 5 6to1 2 3 4 5 6with only two swap operations? You can't; you need at least three.I see, thanks.
At most 3 different shift values are possible. Check each of them (we run dfs and check if the number of bad-ordered vertices — number of cycles <= m).
I will always remember to read special constraints...
There are at most 3 possible k's (you only consider k's where we have >= n-2*m fixed points, you couldn't move more), and m <= n/3.
For each k, just check how many swaps you need to sort it.
I will always remember to read special constraints...
EDIT: Ignore, incorrect
.
Why would it be $$$O(N^2)$$$?
I'm an idiot, for some reason I thought $$$O(TN)$$$ was $$$O(N^2)$$$, which makes no sense. I'll edit the comment.
I feel like I'm missing something obvious, but how does having atleast n/3 fixed points imply that only 3 k's can satisfy that constraint ?
Using swaps you can change at most $$$2m$$$ positions. That means that each valid k has to have at least $$$n-2m$$$ fixed points. This is smallest when $$$m$$$ is maximal, which is $$$n/3$$$. That means you have to have more than at least $$$n-2\cdot n/3 = n/3$$$ fixed points, which is possible for at most 3 different k's.
Thanks. I got to the "we need n/3 fixed points" criteria but could not come to the conclusion that it means we can have atmost 3 K's.
The rather simple piece I was overlooking was that each position can only be fixed by a unique 'k'. So the set of fixed points for any two K's do not intersect and therefore having 3 of them have sets of size >= n/3 means there's nothing left for the rest of the Ks.
Is problem E related to counting inversions? I think I was close to the idea but couldn't implement it on time xd.
Too bad problem B had such weak pretests. :(
Thank you for the contest!
Also, I'm glad to see a contest with hacks, for a change.
E was such a troll problem.
Why so ? . I just want to know why do you think it is . I wasn't able to solve it,Can you explain your solution ?
The basic idea is that for each k, you perform the shift, and then you count the number of swaps needed to go from one permutation to the other. If it is at most m, we found a good k.
But this solution takes quadratic time. The key insight is the constraint on m <= n/3. This means that we only need to consider options where at least n/3 numbers are already fixed. And there can be at most 3 of these.
Please check why it is giving runtime error for Problem-D :- 123349910
Do not do
size() - 1without typecasting. Because size() returns unsigned int, so when size = 0, size() — 1 becomes arbitrarily large.Thanks rifatrraazz !!
Find my D is wrong and resubmit in the last minute!!!!!
And hope my F will pass although it takes 3.96s in pretests
Bruh!
Too many incoming FSTs!
what is the hack testcase for B?
1
aaaabcddddddddddda ddddc
It should output YES
1 fabcb abcbaf
How to solve D?
Scoring distribution: 250-**1900(500+14×100)**-750-1250-1750-2500-3000-3750-5250.
What testcase did you use for hack?
NO NO YES NO
is it right?
One more FST :D
You still missed 4 hacks in your room lol
I did my best in hacks too :)
I thought there was a B test case that I was missing after only hacking 1 B solution in my room and seeing others hack 5+ B solutions, but after system testing I see that others were simply way more lucky with the rooms they were in.
Can someone please help why my solution of B giving wrong answer, got stuck for 1.5 hours:(
link
1 gamfhvaz av -> YES
I solved A,C. Solved C first time in codeforces. Thanks for the contest.
Hello! Why does this solution for C is failing on the second pretest? I was debbuging it for like two hours, but my debug method has a fatal flaw -- i'm brainlet. Pls help. https://pastebin.com/xqysHjZu
Test with
1010101010. Your code outputs7, but the correct answer is6.let's consider answer for k cyclic shifts Let the number of elements in the array that are in their position be p and that are not in their positions be t
Thanks for sharing your observation...!
I thought about it for hours, but couldn't get it :(
What is Test3 in problem F
Pretests for D are very weak.
abcde
abcd
And my pretest passed solution also fails on this test case and I realized that after hacking someone else solution and locking my own T_T.
Nvm, I thought mine would fail, but it is AC. I thought it should be YES. Nevermind.
Ans should be NO ryt?
correct
Problem B & D weak pretest :(
i think standings is gonna change very much :/
I've got a really nice new achievement in this contest. Which is I'm the first participant to get WA on test 23 on Problem E, and almost the only, actually there are just two participants who managed to do that. Thanks codeforces for this new achievement.
How to solve F? I think I figured out how to calculate all pairs apart from ones with structure (i < j a[j] % a[i]).
Add numbers from left to right. When we insert a number $$$a$$$:
The other 2 cases are a little more difficult
In order to count $$$c \mod a$$$ for numbers $$$c > a$$$, lets notice that $$$c \mod a = c - i \cdot a$$$ for numbers $$$c$$$ in range $$$[a \cdot i, a \cdot (i + 1) - 1]$$$. Just loop over $$$i$$$ and count using the formula. Total iterations over $$$i$$$ for all values will not exceed $$$A \cdot log(A)$$$ (similarly to Sieve of Eratosthenes), because all of the elements are distinct.
In order to count $$$a \mod b$$$ for numbers $$$b < a$$$, we do similarly to above, but instead we precalculate for $$$b$$$. So we do something like add $$$b$$$ to position $$$b$$$, add $$$b$$$ to position $$$b \cdot 2$$$ and so on... Then for $$$a$$$ we will add to the answer the number of smaller elements $$$\times$$$ $$$a$$$ — sum on range $$$[1, a]$$$.
We can perform all the operations fast using some fenwick trees.
unbalanced contest + weak pretests, lmfao.
tourist was happily securing second place ... And then the hackers came :)
And then week Pretest came :)
123348935 Problem D Why this gives WA on pretest 17?
1 wtcy w -> NO
Hacks on problem D should be a lesson to future problemsetters that you should either have more than 1 brain cell or use $$$t \leq 10^4$$$.
What do you mean?
You can either make all possible test cases with small $$$n$$$ and a small alphabet size(2, 3 or 4) or do the brainless approach of using $$$t=10^4$$$ and making completely random test cases by partitioning $$$1.25 \cdot 10^5$$$ into $$$10^4$$$ parts and making the size of each part the size of the first string for that test case, making the alphabet size a random number between $$$1$$$ and $$$min(n,26)$$$ and making the size of the second string a random number between $$$1$$$ and $$$n$$$. Both of those will definitely make a test case like
ab a, which is the only D hack that I know of.Well, setting $$$t \leq 10^4$$$ is something different than setting higher $$$t$$$ and just making a few random tests, but I guess the "more than 1 brain cell" part is kind of about it.
Bruh, all pretests on D except 2 (and sample test) have $$$t=1$$$. These 2 pretests have $$$t=10$$$ and $$$t=500$$$ respectively.
How to solve E?
Suppose someone hacked any solution. My question is will that test case be included in system test?? Or it remains the same?
yes it will , every hack case is included in system test
f**k the sh*t weak pretest
FST Forces!!! T_T
Is this real? Red coders getting FST on B?????
It seems hacker also get hacked :3
Good job on making an Educational Hackforces Round!
yeah i can feel the same pain..
Cool.. Educational Round is rated for Div. 1 now, meanwhile AtCoder Regular Contest isn’t for some reason.
In B, except for the edge cases, not getting covered in system tests. There is also another hack case.
The system tests don't include even a single case of "YES" with
sz(s) == 500andsz(t) == 999, because there are solutions (which got AC'ed), with dp array parameters as 505, 505. (dp[505][505]).:(
Why not $$$10$$$ tasks of "write a cycle"-style?
I was wondering for a while. Did you mean "loop"?
I was wondering for loop a while loop. Did you mean "loop"?
Man, what is going on, my submissions are not being checked
Same problem
Same..My all submissions are in the queue.
Your text to link here... I want to know that why problem B is always routime error
I think the line "int i=s.length()-2" is the reason for getting runtime error. s.length() always returns an unsigned integer. If the value of s.length() is 1 then s.length()-2 will not be -1. It will be a big unsigned integer and your string doesn't have that big size. That's why you got runtime error.
Hi Everyone, what is the correct approach for problem D, I was trying to greedily find the next matching characters but it's not the correct approach. Do we need to think around dp for this question?
Do greedy on the reversed strings and you will succeed. To prove it: let's say that in the correct solution we do not match a pair of symbols and do backspace instead. Then after some even number of symbols, we will face the same symbol. We could have just deleted all the skipped symbols + this one and just pick the very first one we faced. Thus, our greedy works.
Sir, thank you for hacking my B, only if you could've done it a little earlier :(
some people were hacked in the middle of the contest and they were able to resubmit? i dont see this fair but ok..
Yeah, It seems unfair because getting hacked is just telling somebody that your solution will fail system tests .Obviously he/she will resubmit and escape FST .
yeah but that allows you to find the mistake and submit a good solution which will pass. and what is FST?
FST-Failed System Tests
Is B the problem with the most FST in cf history ?
The pretest are f*ck*ng weak.
With the help of Problem D，I can't get Grand Master！
I would worry about losing rating, but fortunately div.3 rounds are unrated for me.
unfortunately*
I think he wrote "fortunately" veeeerrry intentionally!! :)
A meme:
B & D
Currently I am new here started few months ago and did few problems of Div 1.
should I solve 100 problems of Div 1 or I should solve 30-40 and then proceed to Div2
I did the below problem B
But I got late to submit
Can anybody suggest if this is a good approach or a lengthy one
Did anybody explain you not to insert code directly in the comment?
Why do I get So much downvote though I already wrote that I am newbie, you code masters treat a newbie this way to encourage them
This blog is about the certain contest, so people only discuss about that contest. Avoid discussing about other stuff.
There are many other people who have asked the same questions. Do some searching online.
People's experiences are different. No one can tell you the single best way to learn.
I have a question but I might get downvoted,
What is the importance of pretests if they don't tell me if my solution is correct or not?
Just thought the same thing..
You can enjoy the contest till it lasts. :D
Well, there is a reason they're called PREtests
So I spent more than 1.5 hours on F and after reading some accepted solutions, I realized that I overlooked the "distinct" part. Please press F guys.
But I think there is only a slight difference if written cleanly(which I wasn't able to do so).
WTF I literally just found that I had overlooked it because of your comment.
F
:O I didn't read that too and just hoped my $$$\mathcal{O}(N*(N/B)*log(N/B))$$$ works lol
Me too! I had a lot of trouble with fitting my $$$O(N * \sqrt {N * \log (N)})$$$ solution in the TL, but eventually I did try deleting the
#define int long longline, which helped a lot. Will reconsider using it in the future.Very bad contest, very bad pretests! Worst contest I've ever seen.
Woops I didn't realize we had to end an even number of positions away from the end in problem D so I tried both parity sequences, muh bad.
Lol, now I got my mistake. Thanks!
yosupo Sir, how were you able to submit both D and F at 00:31? Was F already known to you?
why my submissions are not getting judged?
Don't forget to check my submissions. My all submissions are still in the queue. :(
mine too. :(
Does the $$$==$$$ operator on strings return "false" immediately if the lengths are different? I assumed that it doesn't and tried to hack a guy which would have $$$O(n^4)$$$, but it was unsuccessful.
I think it does.
tried the same with same result, so might be
I found this: https://stackoverflow.com/a/32486610
Damn! Hackforces :/
)
same...I was going to gain 93 rating (and reach expert!), but then after system tests, I'm going to lose 7 rating.
My F passed with 3.993s. Wow.
(Probably the redemption of last round's F)
Can someone point out the edge-case of B, though I passed FST, would like to know what it was?
1
cabad
cabac
And why should not be "YES" ? "cabac" = "cab" + "ac" ? right?
Yes, output should be "YES" but my code printed "NO".
hackforces going great :3
why soooo weak pretest!! :'(
I think that it's the most fstforces round on cf
no, but yes maybe in terms of a number of redcoders getting FSTs.
ok, then which one?
I gathered some statistic on this topic once (well, not exactly this, it was about hacks, but I suspect those with a lot of hacks should also have a lot of SFTs except for educational rounds )
https://codeforces.com/blog/entry/83599
What does FST mean?
Failed System Tests
TLE on 87th test case :( .
Habour poor test case.
This contest told me never locked the problem if you don't want to hack others. I knew why i was wrong when I was hacked by others, but I cannot make any submission anymore.
score prediction after contest +69
score prediction after system test -69
such a big difference between system-testing on B vs E
lesson learnt, never participating in any harbor space contest again.
sanghai2000 mind explaining all those xxxxx=1 in:
https://codeforces.com/contest/1553/submission/123313319
https://codeforces.com/contest/1553/submission/123326933
https://codeforces.com/contest/1553/submission/123334705
?
what's TC 19 of problem D?
If you have WA 19, then you do not take into account the fact that you have to type the entire line. Most likely in some cases you leave the last character unused
what is wrong in this approach with problem D : every two adjacent letters in t string should have even distance between them in s
so the path in s should be even odd even odd ...... or odd even odd even odd even.
how is
cabad cabaccorrect in problem B: Reverse String?the steps are 1 2 3 2 1
c
ca (Going right)
cab (Going right)
caba (Going left)
cabac(Going right)
You have been warned.
Looks like a strong testcase for problem D.
Almost certain it is not. You can check it by yourself if you want.
The test
Does anybody know what was the problem in test 19 problem D?
That's a large test case, I don't know. But looking at your solution, what you're missing probably is that the position at which $$$t$$$ ends must be at even parity relative to the end of $$$s$$$.
hey can you debug my D code pls WA on test 18
Ahhh... thanks, I missed that.
Thanks for the contest! It was nice. Really enjoyed solving it!
What is wrong in this approach with problem D : every two adjacent letters in t should have even distance between them in s. so the path is even odd even odd ...... or odd even odd even odd even.
there is no or. There is a single possibility. D was solvable by a single for. It was honestly ridiculous for a D.
It's a combined round, not ridiculous. In most Div1 + Div2 A-E is solvable by pretty much anyone >= 1700 rating, it's mostly based on observations.
take the index of the last character that you choose, call it p
the number of characters from p + 1 to the rest of the string has to be even
thanks
So the solution to D which checks if $$$t$$$ is a subsequence of $$$s$$$ fails on the second testcase in the sample test and then on test $$$17$$$ if anybody is interested.
I feel that D could have been worded better. "If you type the string s and press backspace instead of typing several characters of s."
This made it seem like using at least one backspace was compulsory.
Can Anyone please explain Why This Solution is giving TLE when being run on GNUC++17(64) compiler and is getting accepted on being run on GNUC++17 Compiler only Link To Accepted Submission
Poor contest. G is trash, don't understand why it is approved.
That and bad pretests on B and D. D was literally 10 lines of code that were basically nearly guessable as logic by pupils that hardly have any experience with such contests. B was probably in the end harder than C or D. Also, that left quite a big difference in difficulty between D and E.
There wasn't a gap between D and E
In fact, if you look at number of solvers and the fact that E is after D, E is barely more difficult at all
I'm curious--why do you dislike G? I think the weak pretests in this round are problematic, but I enjoyed solving G. The main observation (that the answer is at most two) was fairly nice and well-motivated, and my eventual solution was relatively concise. The authors could perhaps have done more to make the round as a whole less focused on speed and on avoiding FSTs, but I thought this problem itself was a fun one to solve.
Quite straight forward, the idea is trivial, I can solve immediately after reading it.
I agree that G is simple and one can come up with a solution in a few minutes (which is bad if you then need to code for 10-15 minutes). That wasn't the case for many testers and organizers though. The number of ACs confirms that the difficulty was fine.
Even though I was very involved in preparing this problem, I think it's the worst one out of E-I. EFHI are good problems.
Also, I think that this contest would be much more enjoyable for div1 without the first few problems. I stated many times that combined rounds are bad. What about using just D-I or E-I, maybe with problem G replaced with something more interesting.
Combined rounds are bad but contests with prizes can't be split into divisions.
To not keep you waiting, the ratings are updated preliminarily. In a few hours/days, I will remove cheaters and update the ratings again!
Thanks.
If cf-predictor is correct, We will see Benq on top replacing tourist.
... For the first time in my case.
How does that matter to you?
At least he/she didn't tag Benq and tourist.
say hello to weak pretests on D
what's TC 19 in problem D ?
One of them is checking if the postfix after the match of the last char of t[] is of even len.
1
aba
ab
Correct Output: NO
Great I TLEd on the system test of D with PyPy because of slow input even though the same solution was less than 300ms with normal Python.
By the way, is stdin.readline always faster than raw_input (in Python 2 / PyPy 2)?
Should I always use raw_input = stdin.readline?
Is output speed important?
Basically B and D:
In D, https://codeforces.com/contest/1553/submission/123351567 I am getting WA on pretest 4, any idea what is the issue? I can see in my submissions but pretest 4 is too huge, so portions of it has got cut out, so I cant understand my mistake. can someone help?
Your mistake is on line 31. You don't need to check this condition and change i and j. We just need to get rid of lines 31-34. But after that, you will have TL. The test will look something like this AAAAAA...AAA AAAAAA...AAB You can get rid of this if you do not check all positions where s[i] == t[0], but only the first one standing on an even and the first one on an odd position.
Thanks a lot. Got AC now :)
No problem
what is the 85 th token of test case 19 for problem D??
How idiotic of me!
Submitted correct solution, for D — Solution
3 minutes after submitting, realised that it would be better to traverse backwards, and not to miss some edge cases (because the pretests seemed week cause of hacks), in case, if any, so changed a bit in the current code, and submitted this — Code
I thought I had covered the case of even parity left out characters in the end, in my latest reverse traversal submission. But, when I checked it again just after the contest finished, and I was like :(
PS: I had hacked 2 solutions in my room, for the same "odd parity case", which my last solution didn't cover.
Press F :(
F
Решил F еще во время контеста. Использовал Дерево фенвика, но оно падало по времени(.
Уже после контеста подсмотрел наибыстрейший ввод/вывод, выйгрыш соизмерим с секундой! ((
ЗЫ: Как оказалось, дело в компиляторе (GCC 17 x64 не проходит, а GCC 17 x32 проходит ) :)
How is this code not getting RE on any of system tests with "YES" as answer and sz(t) >= 505?
Link
It gives RE on my local machine.
I even tried to hack it after the contest, but it didn't get RE, and passed the hack.
How is this possible? Can someone explain?
хороший раунд, одобряю
Претесты еще лучше)
I was so certain that in D we didn't have to type the whole s and could skip its suffix and it even passed the pretests
No idea why I thought this way
errichto
PROBLEM LINK:https://codeforces.com/contest/1553/problem/B
my solution to this problem got hacked during the contest after it passes the pretests(4).I want to know is there a possibility suppose the hacker didnt hack my solution during the contest and in system testing my solution get accepted because the tests prepared for the question was not good enough or lets say it excludes the test case through which the hacker hacked.??
Also i want to know after somebody succesfully hack other solution do code forces add that particular hack test to their set of tests for system testing or code forces set of tests already has that test case covered.?? please answer
Thanks in advance
In what system testcase B is failing ?
its hacked not failed on system test
how to share the link of your all submissions for a contest to others
1
abbb
abbba what should be the ideal ouput in this test case for problem b of this contest
rubbish pretest on problem D. How dare you put only 1 string in a proble which have at most 100000 test cases !!!????
rubbish pretest on problem D. How dare you put at most 100 strings in the pretest in a problem which have at most 100000 test cases !!!????
I loved the contest despite getting system test failed on B and wrong test 17 on D , I learned to debug my code and it also taught me to test your code on different edge cases before submitting and dont truely be dependent on pretests .
Many people are downvoting due to weak pretest which according to me is wrong .
One more positive thing due to weak pretest is that many cheaters also got system test failed or hacked on their solution as they just copy(cheat). And I guess they are the ones who are frustrated and downvoting lol!.
It's our fault that we cannot come up with correct solution for given problem . Those who understood the problem and solved by themselves will figure out their mistakes and are happy from the round .
And we have div 3 round today to regain our lost ratings :) Learn from your mistakes and peace out !
Is there a problem with the 26th test point of question F? The code of the person I have used is still wrong.
has harbour announced or contacted any winners?
Hi, Krause. We will announce the winners later today. Stay tuned!
That turned out to be a salary, not the living allowance
Don't repeat my mistakes
What's the particular difference?