Hello! Codeforces Round 903 (Div. 3) will start at Oct/12/2023 17:35 (Moscow time). You will be offered 6-8 problems with expected difficulties to compose an interesting competition for participants with ratings up to 1600. However, all of you who wish to take part and have a rating of 1600 or higher, can register for the round unofficially.
The round will be hosted by rules of educational rounds (extended ICPC). Thus, solutions will be judged on preliminary tests during the round, and after the round, it will be a 12-hour phase of open hacks.
You will be given 7 problems and 2 hours and 15 minutes to solve them.
Note that the penalty for the wrong submission in this round is 10 minutes.
Remember that only the trusted participants of the third division will be included in the official standings table. As it is written by link, this is a compulsory measure for combating unsporting behavior. To qualify as a trusted participant of the third division, you must:
- take part in at least five rated rounds (and solve at least one problem in each of them)
- do not have a point of 1900 or higher in the rating.
Regardless of whether you are a trusted participant of the third division or not, if your rating is less than 1600, then the round will be rated for you.
Problems have been created and written by our team: myav, Gornak40, senjougaharin and Vladosiya.
We would like to thank:
MikeMirzayanov for Polygon and Codeforces platforms.
SomethingNew, rniya, zwezdinv for red testing.
makrav, snowysecret, AlexanderL, KseniaShk, pavlekn for yellow testing.
gigabuffoon, EgorUlin, KerakTelor, Pa_sha, I_love_geom, petyb, MinaRagy06, toniskrijelj, FynjyBath for purple testing.
Abo_Samrah, dan_dolmatov, pedrosorio, ezdp, azureglow, Chrisedyong, Apachee, arseny2606 for blue testing.
t0rtik, Sergey140146659, hader239, Modern for cyan testing.
- mkshh, petertromso for green testing.
Good luck!
UPD: Editorial
Waiting for that round. Hopefully, It will be a great round for me. Best of luck to all contestants :)
+2 nice
I hope This contest to be my promotion contest, I wish (:
Hey man, good luck and stay safe!
I wish you all the best, son of Palestine, home of brave heroes
Love you Brother. Salam.❣
How are you bro!?
As a green tester I can ensure green participants that they will enjoy this round!
loved this round (happy emoji) (As cf don't allow to add emojis)
😄
This is how
damm boi 😁
Why's this not on the home page yet?
As a tester I hope, that contest will be very interesting for all participants!
As a tester, I hope that this contest will be enjoyed by all participants. Read all problems and good luck! ฅ^•ﻌ•^ฅ
expectantly ^^
Deleted
sory i can`t get a think in line 11 you sed"do not have a point of 1900 or higher in the rating." but in line below it you say "Regardless of whether you are a trusted participant of the third division or not, if your rating is less than 1600, then the round will be rated for you." so is that contest rated for experts or not ?
It's not, it's just people who at least once reached 1900 will not get in the official standings(though if their current rating is below 1600, the round still will be rated for them).
thank you
Ez Expert for me
At least we know that there will be no bitset problems :)
Waiting eagerly. Best of luck to everyone :)
I'm not welcome for Div.4 anymore so I must bring out my best in Div.3!
Hoping to become an expert this round.
Your graph looks exactly like mine!!!! OMG :)
Do you guys give contests together? /s
what do you mean? I am from lebanon, bro is from india...
Need more div.2 rounds
As a tester, I hope the problems will be enjoyable for you!
For the life of me, I will never understand why a Div3 contest has 4 times the Expert+ testers as cyan/green testers. Who are these contests even aimed for?
Anyone else who just saw a message that the contest has started right now?
As a late "as a tester comment", I beg for contribution
W tester = W contest
Going to become Expert.
hack like 20 people and it is very possible predictor says you are 3 points short
make it 2000 people cause your solution to A got hacked
We don't get points for hacking in div 3.
hack people above you they get below you
Never felt dumber with a div3C.
First time to solve till F, my best performance in div3 :)
Problem C
did a problem similar to C yesterday, (a div4E), so i was able to solve it kinda quickly
can you still remember the name of that problem ?
https://mirror.codeforces.com/problemset/problem/1703/E
Same here, i will never take that much time again, constructed an algorith to traverse clockwise lol,
How to solve D ?
Prime factorize every number and count all the primes. if all the prime count is divisible by n then yes
How did u come up with this? Is there any reference?
Note that over any operation, product of all numbers is invariant. Thus, if the final state of all numbers is $$$g$$$, then $$$\prod_{i=1}^na_i = g^n$$$. Thus, for the final state to be achieved, product of all numbers must be a perfect $$$n$$$-th power. Turns out this is sufficient as well. (Why? Hint: Think of how primes will be distributed)
would clarify the relation between N and number of all primes ?
You can find out all primes no more than N in O(N) using Euler sieve.
Find the prime factors of each element and count the total number of occurrences of each prime factor. If all the total number can be divisible by n, then output YES.
Anyone found the A,B,C much annoying problems.. this div3 contest must have focussed on the problems D,E,F ... rather than keeping the participants to got stuck at A,B,C these days...!
Ironically, it was quite the opposite
C was so annoying , D < C
D, E, F were interesting but i found A a little difficult for div3 A unless there is a trick that i couldn't see. Edit: Thanks got it.
A was just bruteforce, that is usually the case, you can just add s to itself a few times until you get to a fairly big size where you are sure the substr won't appear again and use .find() to check if m is a substring. The strings can only get to like 100 so checking till 200 does the trick.
You could just naively check for at most $$$12$$$ operations or so. Time complexity would be $$$O(2^{12}nm)$$$ which is definitely not optimal, but anyways there are not too many testcases so it can't hurt to try too much
UPD: $$$k=12$$$ got hacked, maybe try something smaller like $$$k=6$$$
Any Hints for F? I was finding max and smax ( max and smax will be maximum and second maximun marked node from children ) and considering maximum distance from a node will be max( max distance from children (max) , distance from parent + 1 (max+1 or smax+1) ).
Problem F
Hint
Claim
You don't need to calculate distance from every marked vertex. Just calculate distance from the two marked vertices
x
andy
, such thatdistance(x, y) >= distance(i, j)
for all pairsi
andj
where1 <= i, j <= n' and 'i ≠ j
. This is sufficient.Proof
Let's assume that
x
andy
are the two marked vertices with the maximum distance between them.Now, suppose there exists pair of vertices
v
andz
in the tree such thatdistance(v, z) > max{distance(v, x), distance(v, y)}
.We can claim that:
max{distance(z, y), distance(z, x)} == distance(z, v) + max{distance(v, x), distance(v, y)} > distance(x, y)
.However, this statement contradicts the fact that
x
andy
are the pair of marked vertices with the maximum distance between them. That means there's no suchz
andv
.So, you only need to calculate distances from vertices
x
andy
to satisfy the given condition.Formula
$$$f(i) = max(distance(i, x), distance(i, y))$$$ where $$$(1 \le i \le n)$$$ and $$$(x, y)$$$ are the two marked vertices with the maximum distance between them.
Time complexity:
O(n)
at most 4 tree traversal. two for finding (x, y). two for calculating distances.What is the level of complexity of problem E in terms of rating? Is it like the 1600 problem?
I would say 1300
Many of F solutions have been done using some segtree, or dp....
But mine i just did tree trimming....
Can You explain what is tree trimming or can provide links of segment tree submissions.
coderdhanraj
Is there any approach to solve F using segment tree.
i feel E tests are weak (check my solution)
Hi can anyone help me check my solution for D why it WA3? I had to change the loop to a get prime factorial function but I don't see where this one goes wrong. Thanks in advance https://mirror.codeforces.com/contest/1881/submission/227896125
you should write clear code and use different variables , your loop conditions might get altered because of value of n.
You're only storing the factors which are <=1000. But there maybe factors >1000 for the given constraints. Try as an example [2182, 2186], the output should be NO, but your code gives YES.
wow what a stupid mistake thank you very much!
It is very sad to say that I have solved B and C, which is normal for a 'pupil', BUT I couldn't pass A :(
Can someone tell me which test case can spoil my code ? 227915244
most probably try larger value of i, like 5 or 10.
I have tried after the contest and got wrong answer. 227927816
Can you try this?
Also please note, it's a
logarithmic
function.Yeah! I replaced my check function with
And I got accepted. Thank you
If x="abc", s="cabcab", your code will get a wrong answer.
Isn't the answer 2 ?
Sorry, I gave a wrong example. I think you should modify your check function. Once s1[i]!=s2[j] and j!=0, maybe you should change i to i-j+1.
I tried to use this:
And it works. I am so sad for what I did. It is a silly mistake :( I hope I don't turn to grey again.
For example, if x="abcabcabd", s="abcabd", your check will fail when i=5, j=5, and then you only change j to 0, but i is still 5, which makes you can't find the right match: i=3.
Oh okay, Thank you for your illustration sir
How did you solve B >_<
Your goal is to make all values similar to the smallest value in the given array. This makes you use less steps otherwise, you need to use more than 3 steps for sure. Let's call the smallest value in the array ('mn'). Also you have to make sure that all values in the array are divisible by 'mn'
Check my code 227853520
is the contest not rated for div4? made account on cf few weeks back and still a newbie, though i solved 2 problems but my rating didn't change
ratings update is tmrw
A — Annoying.
balanced contest overall. it's disappointing i wasn't able to solve E because i haven't learned dp yet :(
Can problem F using bfs many source?. I can't implement it during the contest
Multi source bfs gives nearest red vextex not farthest
I felt that today's E was similar (in terms of statement, not solution) to 1798E.
Also, great problemset!
Is there a more elegant way to solve F than some very annoying-to-implement tree DP? (Check my submission for my details)
You need to calculate the diameter of tree (say $$$d$$$), considering the marked nodes as the ends. Now, intuitively the required node should lie in between this diameter, so distance would be $$$\lceil \frac{d}{2} \rceil$$$
I used the re-rooting technique
Not as annoying as mine XD (227924103)
How to solve C?
let ip= n-i-1 , jp = n-j-1. cells with indices (i,j) , (j,ip) , (ip,jp) , (jp,i) should be equal for each 0 <= i, j < n
Screencast+Explanation | A-E
Can anyone tell me the
371st
test for test case2
in Problem G?If someone prefers video explanations or interested in knowing how to think in a live contest or reach to a particular solution.
Here is my live Screencast of solving problems [A -> E] (with detailed commentary in Hindi language).
Why are there so many hacks on A? What is the hack?
Also: https://mirror.codeforces.com/contest/1881/submission/227851818
This looks kinda sus, to escape plagiarism checker?
maximum case with -1 on every TC's answer, hackable when they tried too much on bruteforce
He's definitely a cheater, I see no point int having an infinite while loop with a break statement.
Is my solution for problem A hackable
Nice ProblemSet :)
Maximum possible answer in problem A can be log2(m)+1 . Isn't it?
True
A,B,C,D were tougher than usual div3.
Code for generating a testcase to check TLE for problem A
https://ideone.com/c8Rs0m
Is using unordered_map in D correct? Shouldn't it give tle? https://mirror.codeforces.com/contest/1881/submission/227902139
It's fine. Input constraints guarantee there are at most $$$10^4$$$ values, each less than or equal to $$$10^6$$$, which means up to 20 factors per value ($$$2^{20} > 10^6$$$), so a total of ~200,000 prime factors to add to the map.
I think problem F is to find the spanning tree connecting the selected vertices and then get the diameter r->(r+1)/2 as the result but I don't know how to code it. is my idea correct??
Yes, the idea is correct. Maybe you can use the fact that on the smallest tree spanning the marked vertices, every leaf node MUST be a marked vertex. So, root the tree on some arbitrary marked vertex, and trim the leaf nodes until all leaf nodes are marked vertices.
I had the idea for the problem F but ran into a bug and ended up wasting my time :3
A friendly contest to beginners! Although my B FST... :|
everyone seems to be overcomplicating B observation: we can divide a, b, c each into threads of size equal to the greatest common divisor of a, b, c. this will always be optimal therefore, the number of operations would be equal to: ceil((a + b + c) / gcd(a, gcd(b, c)) / 2) (it only takes 1 operations to split into 2 equal threads
Hello everyone, I am in bit trouble, I have tried to solve the first question of the contest but I am unable to debug it up. If anyone can help me, I will be very thankful. Thanks My submission
Actually, you're missing out on one index in this for loop,
for(int i=0;i<x.size()-s.size();i++){
The variable
i
should go uptox.size()-s.size()
, so the for loop condition should bei<=x.size()-s.size()
Corrected: https://mirror.codeforces.com/contest/1881/submission/227970028
Why this code always work for Problem D?
It's doing the same thing as other's have mentioned above. If the answer is yes, m will be an integer the way it is being calculated
Question A: What data can be used for hack
Hello, how do I solve 1881G - Anya and the Mysterious String with Segment tree using Lazy propogation? This is what I've done 227967016, I haven't used Lazy propogation and it's TLE.
Thank you in advance.
Hackforces again ;)
problem A was very bad for div3.
Can anyone check this i am getting wrong answer in test case 4 for Problem G
here's my submission : 227971449
Overflow. do x%=26 in query1
My solution was hacked yesterday, again I submitted the same solution and it got accepted. Do the test cases used in hacking not get added to the main test-case?
they will get added after the hacking phase is finished, at the time of system testing.
Can someone please explain E. Block Sequence as dp states please and also suggest similar problems
Let $$$dp[i]$$$ denote the minimum number of deletions for the suffix starting on the $$$i$$$-th element: $$$a[i..n]$$$. Base case will be $$$dp[n + 1] = 0$$$.
For $$$dp[i]$$$, we can either delete or take it. If we delete it then $$$dp[i] = dp[i + 1] + 1$$$. If we take it, then the problem reduces to the suffix starting on the $$$j = (i + a[i] + 1)$$$-th element: $$$a[j..n]$$$. So $$$dp[i] = dp[i + a[i] + 1]$$$. We take the minimum.
Can anybody help me figure out why there is TLE in this solution (Problem D)?
Ignoring TLE, your solution is wrong. For the test case $$$[7, 13]$$$, it outputs "YES" when it should be "NO".
Yes, I figured out the reason for the wrong solution. I am checking up to n(size of the input array), while I should be doing it up to 1e6. However, still, I am getting TLE solution
You're going to be doing that for every testcase though. So it'll require at least $$$2e9$$$ operations when $$$t = 2000$$$. You can use map/hashmap to keep count of the primes factors.
Ignore below, it does work.
Though even then, I'm not convinced you won't TLE in later testcases.
THIS CONTEST REALLY MADE ME TERRIFIED TO CONTINUE COMPETITVE PROGRAMMING
It was a nice contest , i have a lot of silly mistakes , but still it was a quality contest
What's up with the rating? System testing got completed hours ago?