### Supermagzzz's blog

By Supermagzzz, history, 3 months ago, translation,

All problems were invented MikeMirzayanov and developed by me Supermagzzz and Stepavly.

1472A - Cards for Friends

Editorial
Solution

1472B - Fair Division

Editorial
Solution

1472C - Long Jumps

Editorial
Solution

1472D - Even-Odd Game

Editorial
Solution

1472E - Correct Placement

Editorial
Solution

1472F - New Year's Puzzle

Editorial
Solution

1472G - Moving to the Capital

Editorial
Solution

• +147

 » 3 months ago, # |   -65 How the editorial posted 2 years ago? If the contest is running.
•  » » 3 months ago, # ^ |   +15 2 "Hours" ago.
•  » » 3 months ago, # ^ |   +7 You can write blogs and not publish them. In this case, it is so the author can get a fast editorial out — by pre-writing the editorial but not making it public yet.
•  » » 3 months ago, # ^ |   +3 You wrote "years" instead of "hours":)
•  » » 3 months ago, # ^ |   0 problem D was almost same as this one.
 » 3 months ago, # |   0 Can someone please explain why in 1472G - Moving to the Capital you can't go in first sample: 6->2->5->1 corresponding d[i] is [2]->[1]->[2]->[1] so there is exactly one step of kind 2.
•  » » 3 months ago, # ^ |   0 To calc the distances you need to consider the direction of the edges.
•  » » 3 months ago, # ^ |   +16 You use road with d[2] -> d[1] twice. It's not allowed
•  » » » 3 months ago, # ^ |   0 I use different roads. Are you talking about same distances change?
•  » » » » 3 months ago, # ^ |   +5 Yes, you use road with this distances twice
•  » » » » » 3 months ago, # ^ |   +3 Oh, I got where is my mistake. I mixed up two kinds. Actually [2]->[1] is second kind of move. I thought it is first kind. shame.
•  » » 3 months ago, # ^ |   +5 Your move from 6->2 is a move of the second kind (dist 2 to dist 1) The move from 5->1 is also a move of the second kind (dist 2 to dist 0).Therefore, you should stop at city 2 — a distance of 1 from the capital.
•  » » 3 months ago, # ^ |   0 yeah,I have the same question.
 » 3 months ago, # |   -37 DIV 3 PROBLEM C VIDEO TUTORIAL LINK https://www.youtube.com/watch?v=7FFzXk-brRU I have tried my best to give you guys proper intuition of this dp problem. Hope you guys will enjoy this video !!!
 » 3 months ago, # |   0 how can editorial be posted if hacking phase is still ongoing
•  » » 3 months ago, # ^ |   +9 That's ok.
 » 3 months ago, # |   0 anyone please explain d
•  » » 3 months ago, # ^ |   +3 For D you only have to justify to yourself why taking the biggest number is always the best choice to make on your turn. The editorial actually does a pretty job of this, but think of it like this — either you take the biggest item left, and it matches your parity (i.e. add it to your score), or it doesn't (your opponent doesn't gain from that item on their turn — also good for you). It's simple to see that therefore a player should ALWAYS take the biggest item (i.e. sort the array and check each move).
•  » » » 3 months ago, # ^ |   0 thanks got it : )
•  » » » 3 months ago, # ^ |   0 Thank you, really simple explanation, even i wasn't able to understand editorial's solution. Thanks again !!
•  » » » 3 months ago, # ^ | ← Rev. 2 →   0 In the code though I saw that the if condition simply added or minused from result based on it's eveness or oddness. Shouldn't it be based on whose turn it was instead? Cause Alice can take an odd number and result wouldn't change.
•  » » » » 3 months ago, # ^ |   0 the index takes care of alice/bob turn...after sorting it in reverse order...the even index corresponds to alice and the odd index corresponds to bob
•  » » » » » 3 months ago, # ^ |   0 Oh yeah, i misread it
•  » » » 3 months ago, # ^ |   0 Actually the solution was so simple that I spent 20 mins trying to prove it but still submitted this method without proving it because I couldn't find any other better method
•  » » 3 months ago, # ^ |   0 Another way to look at it:At each point the player wants to maximise the distance between the scores of one another.You may be tempted to look for parity while picking the numbers, but if the goal is to simply maximise the distance then you increasing your score is equal to hindering the others progress. therefore at each point you pick The largest number which has a potential to reduce the difference between two scores.I hope this makes sense.
•  » » » 3 months ago, # ^ |   0 Thank you much
 » 3 months ago, # |   0 Can someone please explain the approach for B (non DP approach)?
•  » » 3 months ago, # ^ | ← Rev. 2 →   -46 h
•  » » 3 months ago, # ^ |   0 There are two cases sum is odd(definetly no) sum is even if sum is even there can be two cases even= even+even (can always be divided equally) even= odd+odd.in second case if there is no zeroes then and only then it cant be divided other wise its always YES
•  » » » 3 months ago, # ^ |   0 Can you explain the odd+odd part?
•  » » » » 3 months ago, # ^ |   0 if both parts are odd then there must be atleast two onez their to make both parts equal.For ex: 6=3+3 if it doesn't consist of any both part can't be equal (it will be like (2,4) ) because by only 2's we can't make a odd number.
•  » » » » 3 months ago, # ^ |   0 Odd means even+odd So, dividing it by odd+odd It means if sum/2 is odd, there has to be atleast 1 ones in the array. As, 2 is even
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 Simple solution — let's iterate through how many candies of weight 2 we will give to Alice, then the remaining weight should be filled by candies of weight 1. If there are enough of them, then we have found a way of division. How is this true ? If we take the case of 1 1 2 2 2, then Alice will get 2 2, while Bob will get 2 ,1 and 1.
•  » » » 3 months ago, # ^ |   0 Hello jiggly_puff Lets consider the test case 1 1 2 2 2 then As you have writted by yourself Alice will get {2,2} and bob will get {1,1,2} now total weight of alice candies = 4 toatl weight of bob candies = 4 , hence it will return YES as alice_weight = Bob_weight
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 Not the best solution, but probably might be easier to understand:https://codeforces.com/contest/1472/submission/103314992So you can use it as a base for a further optimization.
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 Approach for problem B. Let the number of 1's is one & the number of 2's is two . Let total sum = one + 2 x two. Casework: Case 1 one = 2m & two = 2n or one & two, both are even. Then, give Alice = (m one's & n two's) candies and same for Bob. So, print "YES". Case 2 one = 2m + 1 & two = 2n or one is odd & two is even. Then, sum of candies is odd and hence you can't split it into two equal parts. So, print "NO". Case 3 one = 2m + 1 & two = 2n + 1 or one is odd & two is odd. Then, sum of candies is odd and hence you can't split it into two equal parts. So, print "NO". Case 4.1 one = 0 & two = 2n + 1 or one is 0 & two is odd. Then print "NO" because two is odd, it will be impossible to divide two into 2 equal parts. Case 4.2 one = 2m (m > 0) & two = 2n + 1 or one is even & two is odd. Then print "YES" because Alice can get (n two's + (m + 1) one's) and Bob can get ((n + 1) two's and (m — 1) one's) You can also see my solution here : 103195164
•  » » 3 months ago, # ^ |   0 Lets say s represents the sum of all elements, If s is odd then its totally impossible to split it into two equal halvesOtherwise if s / 2 odd then you need to have atleast one 1 in the array to make the odd halfso the answer is no if (you have a odd total) or (a even total but odd half and no ones with you), else answer is yes
•  » » 3 months ago, # ^ |   0 I used a method based on some observations. You can have a look at my code here. If the sum is odd, obviously there's no way of distributing them in equal halves. Also, if you have no 1's and an odd number of 2's, you again can't do it. In all other cases, you can always come up with a solution, just try out some test cases yourself!
 » 3 months ago, # | ← Rev. 2 →   +4 Problem B is really similar to Div-2A Kitahara Haruki's Gift (https://codeforces.com/problemset/problem/433/A)Great Contest Overall
 » 3 months ago, # |   0 Can someone explain last testcase of Problem G. For node 3 , shouldn't the ans be 1 we can go 3 (2) --> 4 (3) ---> 2 (1)value in brackets is the corresponding d values3--> 4 is of type 1and 4--> 2 is of type 2
•  » » 3 months ago, # ^ | ← Rev. 3 →   +1 Distance to 4 is equal to 2 by 1>5>4, not 3.
•  » » » 3 months ago, # ^ |   0 thanks
 » 3 months ago, # |   0 Thank you very much, now I can solve E, F and G tasks!
 » 3 months ago, # |   +20 A way to avoid repeating the algorithm in E, is to always have H < W for every i. So just swap H and W for some person, if H > W.
•  » » 3 months ago, # ^ |   0 Could you explain a little bit more ? thanks in advance.
•  » » » 3 months ago, # ^ | ← Rev. 2 →   +4 Now I got it. the condition for the person i to be placed in front of the person j can be rewritten like this : min(h_i, w_i) < min(h_j, w_j) and max(h_i, w_i) < max(h_j, w_j) This means you can compare like that.
 » 3 months ago, # |   0 Simple solution — let's iterate through how many candies of weight 2 we will give to Alice, then the remaining weight should be filled by candies of weight 1. If there are enough of them, then we have found a way of division. How is this true ? If we take the case of 1 1 2 2 2, then Alice will get 2 2, while Bob will get 2 ,1 and 1.
•  » » 3 months ago, # ^ |   0 How is this not true?
 » 3 months ago, # |   +15 Here are my attempts to explain solutions in video form after messing almost everything up (this seems to happen more and more often recently...)
 » 3 months ago, # |   +3 on G,can someone explain why (dist[u] must
•  » » 3 months ago, # ^ |   +3 Because you can use only the first option from the statement as much times, as you want. We build the new graph, with edges that satisfy this condition. The second option we check by DP.
•  » » » 3 months ago, # ^ |   0 Hi, This is my Dfs function. Where did I perform the second operation twice? mark : stores the final answer. d : stores the distance from root. void Dfs (int t){ vis[t] = true; for(auto u: g[t]){ if(!vis[u]) Dfs(u); if(mark[u] < d[u]){ if(d[u] > d[t]) mark[t] = min(mark[t], mark[u]); else mark[t] = min(mark[t], d[u]); } else mark[t] = min(mark[t], d[u]); } } Adding that conditon you mentioned, gives me an AC. Can you please tell, where am I going wrong? Thanks
•  » » » 3 months ago, # ^ | ← Rev. 2 →   +3 I am stuck in the same situation. Even without dist[u] < dist[v] condition, we seem to be ensuring that 2nd condition is only counted once because if (dist[u] < dist[v]) { // Only count dp[v] if it's not bridge dp[u] = min(dp[u], dp[v]); } else { dp[u] = min(dp[u], dist[v]); }
•  » » » 3 months ago, # ^ | ← Rev. 2 →   0 Thumbs up to the guys above. I also got stuck over this condition. For anyone interested, here's a graph exposing this bug: 5 7 1 2 1 5 2 3 3 4 4 5 4 1 5 3 In short, without the dist[u] < dist[v] condition, your DFS may update answer for some vertex x (dp[x]) by using dp[y] at a moment when y is still being processed by DFS somewhere up the tree. In this case, dp[y] isn't finalized, it's just not ready. In the graph above, x is 5 and y is 3.Takeaway: DFS doesn't like when you close loops; better let the old DFS finish and only then spawn the new DFS tree to process x.
•  » » » » 13 days ago, # ^ |   0 Thanks man!
 » 3 months ago, # |   0 The editorial for G is very smooth. Great work Supermagzzz and Stepavly.
 » 3 months ago, # |   0 I don't know how people solve problem like today's D so fast !! sed lyf :((
•  » » 3 months ago, # ^ |   +1 Donald trump complaining even on someone solving a problem faster. Not surprising.
•  » » » 3 months ago, # ^ |   +4 I am not Donald Trump..see again
 » 3 months ago, # |   0 Problem B Video Editorial link :https://www.youtube.com/watch?v=ZVLnPd0_3BU I have tried my best to explain with intuition and corner case.I hope you will enjoy this video !!!
 » 3 months ago, # |   0 What a fast editorial! Really helpful; Thank you!
 » 3 months ago, # |   0 I wonder with B they did not include the alternate solution with DP
 » 3 months ago, # |   0 Could anyone please help me to resolve MLE in this solution for Problem C
•  » » 3 months ago, # ^ |   0 Your dp size is 20000 instead of 200000
•  » » 3 months ago, # ^ |   0 you're passing the vector by value, so each time you call solve() a new copy is created. Pass it by reference or make it global.
•  » » » 3 months ago, # ^ |   0 Thanks :-)
 » 3 months ago, # |   0 Why this dp solution gives WA on B.I thought it is equal sum partition. https://codeforces.com/contest/1472/submission/103310451
•  » » 3 months ago, # ^ |   0 I had two use greedy approach for this but I am willing to know it's do solution and fault in my code. Thanks in advance.
•  » » 3 months ago, # ^ |   0
•  » » » 3 months ago, # ^ |   0 Thanks bro.
 » 3 months ago, # |   +3 I came across an ID in this round which was submitting solution with wrong answer for a particular corner case. FOR eg if n==123234 then print "NO" , this was done so that it could be hacked by another IDs. :p
•  » » 3 months ago, # ^ |   0 Name of the ID please.
•  » » » 3 months ago, # ^ |   0 Charon_souls_carrier
 » 3 months ago, # |   0 Can someone please help me why my code gets a WA(I get why it might get TLE but no WA): 103300499
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 wrong answer jury found the answer, but participant didn't (test case 11)This is a comment of your code, so you can look for test case 11(it is visible) and see why it doesn't work. EDIT: you're only looking for elements with smaller X but a correct answer could also be element whose X is larger Y is smaller and so when it is laid sideways than it could fit in the element that you are checking. So your for loop J should go from 0 to n (i.e. check all elements) P.S. even if you did this you would get TLE
 » 3 months ago, # |   0 Can anyone please explain why my solution is showing WA. My logic is obvious. One can easily get that after seeing my code. SubmissionThanks in advance for helping :)
•  » » 3 months ago, # ^ |   0 I think its because you didn't sort the arrays. You should always take largest number, and even if you divide numbers in odd and even you should sort the array and look for larges numbers. I used a similar approach 103227720. Hope this helps.
•  » » » 3 months ago, # ^ |   0 Yeah.. Thanks I get it. I didn't sort the array (T_T) Literally if you see my commented solution below it i had sorted the array. But i missed while writing new one. I thought i have sorted the arrays. Why these things happen to me (T_T)
•  » » 3 months ago, # ^ | ← Rev. 3 →   0 I didn't actually understand your code, but I think the approach is similar. So hopefully this will be helpful! Explanation for DLet's observe first, when it's Alice's turn, she either blocks(selects the biggest odd number so the maximum point that Bob can take decreases) or takes the biggest even number.It's also similar for Bob, he either tries to block Alice by selecting the biggest even number(so the maximum point that Alice can take decreases) or takes the biggest odd number.Why this is optimal?Let's take this case as an example: $\text {N = 4, } \$ $\text{A: {5 3 4 6}}$1- Alice starts, now there are 2 options, blocking 5 or taking the 6.1.A- If she blocks the Bob, then Bob will also block her and she'll lost 1 points because of herself. And it'll stay at "Tie" when it's Alice's turn(3rd turn) again, thing is, if Alice had choosen to take 6, she would be on the lead now(check 1.B)1.B- If she takes the 6, then Bob will take the 5, and now it's Alice's turn(3rd) again, and now she is on the lead because of her decision.2- It's the 3rd turn, Alice 1 — Bob 0, and the array is {3,4}2.A- Alice can choose taking the 4 or blocking the 3, if she blocks the Bob, then Bob will also block her in the next round, so the final points would be (1 — 0) (It's true but not optimal, check 2.B)2.B- Now if Alice chooses taking the 4, Bob will take the three in the next round and the final points would be (5 — 3)Because the difference is higher, 2.B is more optimal for Alice. But, because of Bob has no action over the 1st turn of Alice, it's also optimal for him too.I hope this helps. Now for the coding phase, Code partMy code for DBasically what this code does is, if $\text{i is even}$ it means it's Alice's turn, and if the current biggest value is even, she takes it, if it's odd, she blocks it.Opposite case for Bob.Note: I wrote this 5 minutes before I went to sleep, so if there are mistakes, forgive me and correct me please.
•  » » » 3 months ago, # ^ |   0 Thanks for the help. I get it now :)
•  » » » 3 months ago, # ^ |   0 Man thanks it really helped !
 » 3 months ago, # |   0 Can anyone explain problem G? Why do we have to change the graph? Why can't we apply dp on the graph with cycles?
•  » » 3 months ago, # ^ |   0 I have applied dp only my friend. You can see my submission. Reply here if you didn't get the logic.
 » 3 months ago, # |   0 I enjoyed the last dp problems of this contest :)
 » 3 months ago, # |   0 Does anyone else feel that E was too much implementation heavy?I thought about exactly this solution but with 30 mins remaining. Then dropped the idea thinking that it might have some better segment tree solution instead of so much implementation. lol
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 Actually, it is not (at least for me).It could be implemented using a segment tree in about 30 ~ 40 lines of code.
 » 3 months ago, # |   0 G was cool, E took me a while but it was fun too. Overall good problems!
 » 3 months ago, # |   0 Is it good to write an editorial while the hacking phase is still on?? I don't think so.
 » 3 months ago, # |   0 It was a bit tough contest
 » 3 months ago, # |   +1 Are we allowed (yet) to ask inquiries regarding hacks for this round?
 » 3 months ago, # | ← Rev. 2 →   0 Can anyone tell me how ans of this test case of ques d is TIE.3 2 1I guess here bob should win. First alice chooses 2, then bob goes with 3 . Then at end alice chooses 1 and removes it. So bob wins here(since score of bob is 3 and alice score is 2).How is it draw?
•  » » 3 months ago, # ^ |   0 IF Alica pick 3 first, then if Bob pick 1 Alica wiil win because in next turn she will pick 2(Bob_score = 1, Alica_score = 2), if Bob picked 2 insted of 1, Alica will be left with 1 (Bob_score = 0, because he piced even number, Alica_score = 0, because she picked odd number). Then answer is TIE.
 » 3 months ago, # |   0 Can C be done by top-down approach?
•  » » 3 months ago, # ^ |   0 Yes , link to my submission https://codeforces.com/contest/1472/submission/103216440
•  » » » 3 months ago, # ^ |   0 Got it, thanks!!
 » 3 months ago, # |   +1 Why is this Solution got hacked? It is almost same as the Editorial. Please let me know as I don't wanna to repeat the same mistake again. Thanks in advance !
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 Yes, I was wondering the same thing as well. My solution (and many other Java solutions of a similar format) were also hacked (and ultimately received TLE), but I cannot find any apparent inefficiencies.
•  » » » 3 months ago, # ^ |   +2 I heard Java sort uses quick sort. Quick sort in a worst case takes $O(N^2)$ complexity in an anti quick sort test.
•  » » » » 3 months ago, # ^ |   +5 That explains it. Thanks!
 » 3 months ago, # | ← Rev. 3 →   -6 Is this Dp forces ?
 » 3 months ago, # | ← Rev. 2 →   0 I have two implementations for E which I think are easier than the one in the editorial.
 » 3 months ago, # |   0 Frustrating , Frustrating .... Can't believe it i wasted more than 1 hour , to solve D and never noticed variable was crossing integer range , So guys always check if range is crossed or not :(
 » 3 months ago, # |   0 In the first test case of the given example of question E, Test case:33 45 43 3Ans: -1 3 -1Why is it -1 for the first pair? 3,3 can be placed in front of 3,4 right? Am I missing something?
•  » » 3 months ago, # ^ |   0 because the height and weight have to be strictly less than the person behind. 3 is less than 4, but 3 is not less than 3.
•  » » » 3 months ago, # ^ |   0 Oh, silly mistake
 » 3 months ago, # |   0
 » 3 months ago, # |   0 DP Contest :)
•  » » 3 months ago, # ^ |   0 There is only 2 problems I solved with dp and I think this is enough. Problems which have dp tag also have some other solutions this contest
 » 3 months ago, # |   0 cf 693 1st question please help me , on what test case my code fails because I think that my code is right Its humble request thanks in advanceinclude using namespace std;long long count(int n){ long long c=0; while(n%2==0){ n=n/2; c++; } return c; }int main() {int t; cin >> t; while (t-- > 0) { int w, h; long long n; cin >> w >> h >> n; string s = ""; if (n == 1) { s = "YES"; //continue; } else if (w % 2 == 1 && h % 2 == 1) { s = "NO"; //continue; } else if ((w % 2 == 0 && h % 2 == 1)) { long long x1 = 2*(count(w)); if (x1 >= n) { s = "YES"; } else { s = "NO"; } //continue; } else if ((w % 2 == 1 && h % 2 == 0)) { long long x2 = 2*(count(h)); if (x2 >= n) { s = "YES"; } else { s = "NO"; } //continue; } else { long long a1=2*(count(w)); long long a2=2*(count(h)); if (a1*a2 >= n) { s = "YES"; } else { s = "NO"; } //continue; } cout << s << endl;}return 0; }
•  » » 3 months ago, # ^ | ← Rev. 4 →   0 U multiplying count (w/h) by 2. Rather, make it like this-long long x1/x2 = pow (2,count (w/h))And in, a1/a2 u r multiplying it.. i don't know if it works or not.. but u can give it like thislong long a1= pow (2,count (w)+count (h))then, just check if a1>=nReason- suppose u can cut 3 times, u can make 8 papers. But your answer gives 6 papers
•  » » » 3 months ago, # ^ |   0 My code give correct answer on the test cases given by codechef .In this I use 5 condition 1st if n==1 if n is one then ans is yes 2nd if w and h is odd and n is not 1 then ans is NO 3rd if w is even and h is odd then I count how many times w is divided by 2 ,and finally multiply by 2 For example w= 4 and h =7 then max=4 4th case is vice versa of 3rd 5th case if both w and h is even then I count how many times w and h is divided by 2 ,and and multiply both by 2 For example w=4 and h=8 then max=24But when I submit the code it is saying wrong output on test cases 2 Please help
•  » » » » 3 months ago, # ^ |   0 I made the changes that I told you to make and it is giving me AC.Here you go
•  » » » » » 3 months ago, # ^ |   0 Thank you so much Finnaly accepted I need one suggestion from you Suppose I gave a codeforces contest and solve 2 question and I got stuck in 3rd question Then should I only have to see the 3rd question editorial or how many questions should I have to up solve I mean that how I improve myself Any tip from your side
 » 3 months ago, # |   0 I literally did the same thing for problem D — 103294044Can anyone help me how to optimise this, because I got tle
•  » » 3 months ago, # ^ |   0 Quoting some lines in your code : while (n>0): maximum = max(arr) Finding max in an array takes O(n) time, and you are doing that operation for n times, that means your code is effectively getting to a complexity of O(n^2). But the constraints on n is (1 ≤ n ≤ 2 * 10^5), so for O(n^2) your code would have to do (2 * 10^5) * (2 * 10^5) = 4 * 10^10 operations per second. But normal CPUs can do only about 10^8 operations per second. Hence you are getting a TLE.
•  » » » 3 months ago, # ^ | ← Rev. 2 →   0 Oh ryt! So would it be better if I first sort and then use them? or is there any other way? which is even better than sorting and using
•  » » » » 3 months ago, # ^ |   0 First Sorting and then using would work fine. I am also linking my solution to the problem so that you may take reference (Although it's pretty much the same what editorial says) — 103483268. If my answer helped you in any way, please do vote a like!
•  » » » » » 3 months ago, # ^ |   0 Thanks a lot! I liked it too XD
 » 3 months ago, # |   0 I took a different approach for A, first I reduced the width and height to minimum it could be. Thus I have the width and height of the smallest card. I can use it find the area. Now I divide the area of the original card with the area of the smallest card. Now I have the number of cards I can create.
 » 3 months ago, # |   0 For D first test case my code is giving right output in my pc whereas on codeforces it gives wrong output 103274282
•  » » 3 months ago, # ^ |   0 else if (o[j]>=e[i] && !fl) { sumb+=**o[i];** --j; fl=!fl; }Maybe this is causing some problem
 » 3 months ago, # |   0 I was given similar problem as D in one of my interviews and I couldn't think of greedy approach, could only come with DP approach and was rejected. Now I solved it with so much ease in comfortable environment. Hate interview pressure.
 » 3 months ago, # |   0 Can someone please explain E?
•  » » 3 months ago, # ^ | ← Rev. 2 →   +1 Ask everyone to stand (make $h_i \leq w_i$ for every $i$). Sort them based on $h_i$ then $w_i$. Now we know that the $i$-th person is at least as tall as all of the people who come before him ($h_i \geq h_j$ for every $j < i$). Iterate through everyone and find the largest $j$ such that $j$-th person is strictly shorter than the $i$-th person ($h_i > h_j$). As they are already sorted, now we know that the $i$-th person is also taller than all of the people who stand in front of $j$-th person. Now we have $j$ possible candidates that can be placed in front of $i$-th person. We no longer have to check whether they are shorter than $i$-th person or not, because we already know that from the previous step. We need to find among the $j$ possible candidates, is there a person who is thinner than the $i$-th person. There could be more than one person who fits the criteria, so we will just greedily pick the thinnest person among those $j$ people and check if the thinnest person is thinner than $i$-th person.
•  » » 3 months ago, # ^ |   0 Have a look at the video by galen_colin: https://www.youtube.com/watch?v=p5FNuXJoLDk&t=1054s
 » 3 months ago, # |   0 In problem D (language used : JAVA 11) , I am getting TLE in testcase 10 with this snippet of codewhile(t-->0){ int n = ni(); int arr[] = new int[n]; for(int i=0; i=0; i--){ k++; if(k%2 !=0 ){ if(arr[i]%2==0) a+=arr[i]; } else{ if(arr[i]%2!=0) b+=arr[i]; } } if(a==b) pn("Tie"); else if(a>b) pn("Alice"); else pn("Bob"); } and getting AC with this snippet of codewhile(t-->0){ int n = ni(); Integer arr[] = new Integer[n]; for(int i=0; ib) pn("Alice"); else pn("Bob"); } Can anyone help me to know what is the cause?
•  » » 3 months ago, # ^ |   0 Exactly what you are not able to understand?
•  » » » 3 months ago, # ^ |   0 In first code I sort the array in non decreasing order and I traverse the array from last for chosing the number for alice and boband in second I sort the array in non increasing order and I traverse the array from starting for chosing the number for alice and bobBoth are about same approaches nearly while first one is getting TLE and second one is AC
•  » » » » 3 months ago, # ^ |   0 Ok, I found the reason: It's because you're using Wrapper class (Integer class) in the second case. When we use wrapper classes, we are basically sorting "Objects" instead of "Values", and Arrays.sort() handles them differently. It implements "Merge Sort" while sorting Objects (Worst case time complexity: O(N*log(N))) but implements "Quick sort" while sorting Values ((Worst case time complexity: O(N^2))). So, if you use Wrapper class in the first case instead of primitive type, it would work fine as well. Accepted Submission: https://codeforces.com/contest/1472/submission/103375034For detailed explanation, check out the reference: https://www.geeksforgeeks.org/java-handling-tle-while-using-arrays-sort-function/
•  » » » » » 3 months ago, # ^ |   +1 Thank you so much
•  » » » » » 3 months ago, # ^ |   0 Learn a lot from above
•  » » 3 months ago, # ^ |   0 Please use ruffle sort (Randomized quick sort to avoid quick sort worst case)-: My Code in case it helps-: `package Div3_693;import java.io.BufferedReader; import java.io.InputStreamReader; import java.util.Arrays; import java.util.Random;public class D { public static void main(String[] args) throws Exception { BufferedReader br=new BufferedReader(new InputStreamReader(System.in)); int tc=Integer.parseInt(br.readLine()); for(int tt=0;tt=0;i--){ if(turn){ if(arr[i]%2==0){ ascore+=arr[i]; } }else{ if(arr[i]%2==1){ bscore+=arr[i]; } } turn=!turn; } if(ascore>bscore) System.out.println("Alice"); else if(bscore>ascore) System.out.println("Bob"); else System.out.println("Tie"); } } static void ruffleSort(int[] arr){ Random r=new Random(); for(int i=0;i
 » 3 months ago, # |   0 https://codeforces.com/contest/1472/submission/103236341 this is my solution of D main idea is what person must bring max(his parity, opposite parity(it means not give this element to enemy))
 » 3 months ago, # |   0 Can problem E be solved with monotone stack? Think the array to be circular, then for each element find the next element so that (h1 < h2 and w1 < w2) or (w1 < h2 and h1 < w2)
 » 3 months ago, # |   0 anyone can provide better sol for Problem E or any video
•  » » 3 months ago, # ^ |   0 I wrote my solution here.
»
3 months ago, # |
0

Hi, I am facing some problems with D. Even-Odd Game. I have read the editorial as well as the solution provided. I think I am doing the exact same thing given in the solution: always selecting the biggest available number for either player and adding or subtracting it from a global sum depending on who the player is and the parity. However, I am getting wrong answer on test case 3. ~~~~~

# include <bits/stdc++.h>

using namespace std;

int main() { int t, n; cin >> t;

for(int i = 0; i < t; ++i)
{
cin >> n;

int a[n];

for(int j = 0; j < n; ++j)
{
cin >> a[j];
}

int size = sizeof(a) / sizeof(a[0]);

//sort array in ascending order
sort(a, a + size);

int score = 0;

int num;

int turn = 0;

for(int j = n-1; j >= 0; --j)
{
num = a[j];

//Alice Turn
if(turn % 2 == 0)
{

if(num % 2 == 0)
score += num;
}
//Bob Turn
else
{
if(num % 2 == 1)
score -= num;
}

++turn;
}

if(score > 0)
cout << "Alice\n";
else  if(score < 0)
cout << "Bob\n";
else
cout << "Tie\n";

}

return 0;

} ~~~~~

•  » » 3 months ago, # ^ |   0 Make score long long
•  » » » 3 months ago, # ^ |   +1 Thanks! It worked.
•  » » » » 3 months ago, # ^ |   0 Always take care of data type you are using and what max data program can generate.
 » 3 months ago, # |   0 Thanks for this round. I like problem G very much, it is the classic Codeforces' style task. Quality of this round was much better than previous five or even more.
 » 3 months ago, # | ← Rev. 2 →   0 In problem D i have used erase for deleting the element and it has failed, but i see solution with same logic but using pop_back() instead of erase getting their solution accepted. I want to know that whether the time complexity of erase more than pop_back() or are they same in c++ ?
•  » » 3 months ago, # ^ |   0 See this and this.
 » 3 months ago, # |   0 I have a different solution for F, where I keep a offset between the top and bottom pointer, and use some case analysis. Time is O(mlogm).
 » 3 months ago, # |   +5 As the problem D has a trival solution, I have faced a false plagarism accusion on it. "Your solution 103217882 for the problem 1472D significantly coincides with solutions lit2019039/103212550, whohet/103217882."This question has a really trival solution and hence is clearly a coincidence. Please help and remove this plagarism mark. This is literally a coincidence, as the problem really have same solution. My solution link https://codeforces.com/contest/1472/submission/103217882Match solution link https://codeforces.com/contest/1472/submission/103212550Please have a look as problem really have a trival solution
 » 3 months ago, # |   0 How does the code which was giving RE on testcase 4 with a slight modification in the comparator function (used for sorting) gives AC(aka happy new year). The change simply is using "<" instead of "<=". any input in this regard is highly appreciatedcode with AC VERDICT:https://codeforces.com/contest/1472/submission/103375747code with RE on test4 :https://codeforces.com/contest/1472/submission/103375782.
•  » » 3 months ago, # ^ |   0 I bealive comparator function should be unique. Meaning if A > B then B < A. In your case with <= you are saying that at that if A == B then a could be before or after B which should not be possible because then sorting is not unique. I.E. if your functions returns A < B it should also return !(B < A)
•  » » » 3 months ago, # ^ |   0 I believe when you say "A<=B" that would mean that if A and B are equal then A should be before B , Although I understand what you are trying to infer, and I think that it could be right but that would depend on the sorting algorithm, furthermore if it was the case that A and B are repetitively compared with each other, shouldn't it result in a TLE instead of RE?
•  » » » » 3 months ago, # ^ | ← Rev. 2 →   0 I bealive it works that if you compare A and B, than comparator automaticly compares B with A, and it should yield same results. In your case it gives A before B in first, and B before A in second which is contradictory. And so it gives an error.
•  » » » » » 3 months ago, # ^ |   0 alright that makes sense. thanks a ton. have been struggling with it for a while now. thanks!
 » 3 months ago, # |   0 I felt E was harder than F. I spent wayy too long on E. Shouls have attempted F first.
 » 3 months ago, # |   0 Why is there a "Tie" in the second test at problem D?
•  » » 3 months ago, # ^ |   0 Each player takes the largest number so their opponent can't take it. So Alice takes 3, then Bob takes 2, than Alice takes 1, after that score is 0:0. This is optimal for both players. If Alice takes 2 in first turn then Bob will take 3 and win, so she has to take 3 herself. Same logic aplies after first move.
•  » » » 3 months ago, # ^ |   0 Thanks
 » 3 months ago, # |   0 can anybody please explain C ?
•  » » 3 months ago, # ^ |   0 could you be a bit more specific as to what you didn't understand?
 » 3 months ago, # | ← Rev. 2 →   0 Can someone please help me out for question D. My submission is showing runtime error on testcase 1 but it works fine on my ide
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 yeah, just select g++ 17 instead of whatever variant you're selecting right now, also your logic fails the further test cases. incase you need to see the correct code similar to yours, you can go through mine :https://codeforces.com/contest/1472/submission/103348877 however even this is a too slow code and can be made better by using pop_back instead of erase. but anyway it passes the test cases. :)
•  » » » 3 months ago, # ^ |   0 My logic did not fail! Thank you so much for the help though :)
•  » » 3 months ago, # ^ |   0 You pass odd.size() - 1 as a parameter — size() (in many implementations) returns an unsigned 64-bit value, so if it's 0 the result after subtracting will be the highest representable value. However, that is converted to a signed long long, and since the value is greater than can be represented in a long long, the value is implementation-defined.You can cast the result from size() to a signed type first.
•  » » » 3 months ago, # ^ |   +3 Thank you so much! It has worked now!
 » 3 months ago, # | ← Rev. 3 →   +6 Is there any way that I can uphack my solution to G. It's linked here 103279910. I have tested it on custom invocation on CF and I know there is definitely a case that makes it TLE. I waited out the hack phase since I wanted to keep my rating, but now I don't see the option to hack it anymore. EDIT: I think you have to be 1900+ to uphack. Here is my generator that would cause my solution to TLE. https://pastebin.com/uYxwMnYzAny1 who is eligible to uphack, go for it :)
 » 3 months ago, # | ← Rev. 2 →   0 Problem E.Link1 solution is accepted wherease link2 is not. The only diff between the 2 solutions is that for pair(hi,wi) whose hi is least is being calculated differently:-In link1: let the main loop take care of everythingIn link2: comparing all (hj,wj) with (hi,wi) for second condition [wj < hi && hj < wi = > ans[i] = j] [Since first condition can never be satisfied]Why does link2 solution fail?
 » 3 months ago, # |   0 can anyone link a good video solution for E? thanks
•  » » 3 months ago, # ^ |   0 https://codeforces.com/contest/1472/submission/103375747 this happens to be a self explanatory code.
•  » » » 3 months ago, # ^ |   0 for me it is always hard to understand other people's code, can you please elaborate how did you solve it? did you follow editorial's method or your own different method?
•  » » » » 3 months ago, # ^ | ← Rev. 2 →   0 same here, anyway I guess the code they've provided is too complicated and involves a lot of things like binary search etc. i have rather written a much simple code. here is how I did it make a vector having 3 fields -> height, width and index before populating the vector sort the height width in decreasing order OR just swap height with width if width>height. now sort the vector populated above, such that the first field (height) is the dominating factor, in case 2 elements have same height then compare width. After step 3 you'd be having a pretty much sorted vector. Now the only thing that remains is going through the sorted vector(left to right), and keeping a check of a the smallest width encountered till now (as heights are already sorted they need not be checked), however one edge case needs to be verified, the minimum value of width could be of a element having same height as the current element, hence i've used another variable for it, for every element having index(3rd field) 'i'if above condition is met by a previously encountered element having index 'j' then save ans[i]=j; else ans[i]=-1; I guess I might have missed a few/many things (lemme know if you know any further explanation).
 » 3 months ago, # | ← Rev. 2 →   0 What's the DP state/solution for D no. problem? I can just feel the greedy approach. But problem tag shows DP too.
 » 3 months ago, # |   +1 Another way to think about E:We can think about the pairs $(h_i, w_i)$ and $(w_i, h_i)$ as points on a 2D plane.Then the problem becomes: for each point $P(h_i, w_i)$, find if there is any point that lies completely inside the rectangle that is formed by the origin $O(0, 0)$ as its bottom left corner and $P$ as its upper right corner.
 » 3 months ago, # |   +4 I've uploaded my post-contest stream to YouTube: https://youtu.be/nCSLlTCDnDs
 » 3 months ago, # |   0 Can someone please point out what is the mistake in this solution 103260792 for problem D? I hope this is the correct place to ask... Thanks!
•  » » 3 months ago, # ^ |   0 Use long long instead of int for storing the sum as the range of int is small for storing sum in worst case.
 » 3 months ago, # |   0 Can someone please explain why my code is giving TLE for 1472E — Correct Placement.I have swapped w and h whenever w>h while storing the inputs in a vector. I have then sorted the vector, found index of hj for every hi such that hj
 » 3 months ago, # |   0 For D how can this input be a tie? 3 2 1
•  » » 3 months ago, # ^ |   0 Alice will pick 3 in his first move(because he picks 2, then bob will pick 3 and wins thereafter), Then bob will pick 2(same reason for 1). Then alice will pick 1 at end. Both will have 0pts at the end.
•  » » » 3 months ago, # ^ |   +1 Ok, got it, basically, it means when the game ends, both of their scores is 0.
•  » » » » 3 months ago, # ^ |   0 Yes
•  » » » » » 3 months ago, # ^ |   +1 Thanks!
 » 3 months ago, # |   0 https://codeforces.com/contest/1472/submission/103513913 Can anyone tell me what is wrong with this? It is for the Problem G.
 » 3 months ago, # | ← Rev. 2 →   0 Could someone who solved D explain thought process/mental steps leading to solution?Is there any general approach to similar problems?
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 I found pretty nice explanation here https://youtu.be/p5FNuXJoLDk?t=665
 » 3 months ago, # |   0 Could anyone please help me,I Am getting different answer in different online compilers. (Problem B) (link to my solution: https://codeforces.com/contest/1472/my) Thanks.
•  » » 3 months ago, # ^ |   0 first of all you used a variable 'z' for size of array , and also you used it as variable to count number of 2 present , and also after correcting this , your code will not give right answer for test case — {2,2,2,1,1}
•  » » » 3 months ago, # ^ |   0 Thanks.
 » 3 months ago, # | ← Rev. 2 →   0 problem E VIDEO TUTORIAL LINK : https://www.youtube.com/watch?v=M8jqjbJjoxA I HAVE TRIED BY BEST TO EXPLAIN WITH PROPER INTUITION(twopointer + dp)SO THAT YOU CAN UNDERSTAND EASILY . I HOPE YOU GUYS WILL ENJOY THIS VIDEO !!!
 » 3 months ago, # |   0 DIV 3 PROBLEM C VIDEO TUTORIAL LINK https://www.youtube.com/watch?v=7FFzXk-brRU I have tried my best to give you guys proper intuition of this dp problem. Hope you guys will enjoy this video !!!
 » 3 months ago, # |   0 Can someone help me find the complexity of this code:103390550? It just barely passed the 4s time limit.
 » 3 months ago, # |   0 Can anyone explain to me why my problem C solution is having a memory exceeded limit? What should i change to make it more memory efficient? 103327344 — Here i thought i used dp to max it faster but my other solution(103317010) without dp was quicker. Can anyone explain why this is?
 » 3 months ago, # |   0 Problem E from my view: Always assume that w h, we want to swap them), sort the array of (w, h)s in ascending order, we may want to save index (for tracking indexes) Traverse the sorted array from left to right, for each i-th pair (w, h), i.e. (wi, hi), the problem becomes searching for j (j = 1..i-1) where wj < wi and hj < hi Now that we have wj <= wi for any j in 1..i-1, we only needs to binary search for upper bound of j which wj not exceed wi-1, let's call the bound is k The only problem is we want to search for some j in 1..k where hj < wi. Let's construct a min array in such a way that min[k] = min { hj, 1 <= j <= k }, we also want to maintain 'argmin' array to store index which contains min Then, result of finding j where hj < wi is j = argmin[k]
 » 3 months ago, # |   0 can someone please explain the DP approach from problem B?
•  » » 3 months ago, # ^ | ← Rev. 2 →   0 SpoilersFirst, you can sort; it does not change the result. Second, observe that for equal numbers like 1 1 1 1; you only need to increase the last number (maintain increasing property). Third, let dp[i][0] and dp[i][1] be the answer for the first i elements where we do not increase ai or we increase ai. You build the DP relations for dp[i] and dp[i-1],
 » 3 months ago, # |   +5 I cant stand the fact that the dude on problem E invite n friends 1<= n <= 2 * 10^5 to celebrate new year! Guys we are in corona virus pandemic we should be more carefull! Imagine what will happen after 14 Days if there was someone with corona in the party! Be more carefule stay home and code, stay safe :D
 » 3 months ago, # |   +8 Simpler solution for F: 103672111 Sort the blocked cells by x, y Pair the adjacent(after sort) blocked cells in disjoint pairs. Between blocked ceils within the same pair: distance has to be odd(or opposite color in chess grid as described in editorial) Between the second cell from ith pair and first ceil from i+1th pair: they cannot be on the same column. Otherwise it's going to block the flow of white cells, which is going to create an odd number of white cells in both side, which is impossible to fill.
 » 3 months ago, # |   0 for problem G: can someone explain on what basis the second kind kind of action is taken?for example :// if there are cities "A", "B" and "C" of distance 3 having a road to city "D" of distance 4, and with 2nd kind of action city "D" can become of distance 1 making all cities "A", "B", "C" and "D" of distance 1.// we have city "A" of distance 3, has a road to city "E" of distance 4 that can be reduced to distance 0 with 2nd kind of action, making "A" and "D" of distance 0.which one should be chosen??
 » 3 months ago, # |   0 for 1472D - Even-Odd Game it shows a dp tag , but I am not able to solve with dp if anyone has solved with dp please post ur solution
 » 3 months ago, # |   0 sorry I am a noob...but what will be the complexity of this algorithm??
 » 3 months ago, # |   0 Why there is a graph tag in problem C?
 » 3 months ago, # |   0 i need help..! submission to D.Please find what is wrong with this code Your text to link here.... i am not able to find any error.! it is giving runtime error.!
 » 3 months ago, # |   0 Does anyone know a possible reason for MLE in Problem G?
 » 2 months ago, # |   0 QC: Long jumps Why i starts from n — 1 in this loop? for (int i = n - 1; i >= 0; i--) { dp[i] = a[i]; int j = i + a[i]; if (j < n) { dp[i] += dp[j]; }
»
2 months ago, # |
0

# include <bits/stdc++.h>

using namespace std; typedef long long int lli; typedef long double lld;

# define FOR(i,l,u) for(lli i=l;i<=u;i++)

int main() { lli t;cin>>t; FOR(k,1,t) { lli n;cin>>n; lli A[n+1]; FOR(j,1,n) { cin>>A[j]; } lli score[n+1]={0},maxi=0; FOR(j,1,n) { if(score[j]==0) { lli i=j; score[j]+=A[i]; i+=A[i]; while(i<=n) { score[i]=-1; score[j]+=A[i]; i+=A[i]; } if(score[j]>maxi) maxi=score[j]; } } cout<<maxi<<"\n";

}

}

why am I getting time limit exceeded? In the above code no element is traversed more than twice, so time complexity is-O(n) for each test case.

•  » » 7 weeks ago, # ^ |   0 Technically ,your solution is O(n^2).
 » 4 weeks ago, # |   0 I've written a Python solution of F, with my own thinking process inside of it. Not a solution that I could expect to achieve within the time frame of the contest, but hope it helps!