Since goals are very important in any individual's life, I decided to set a goal for myself. The target is to set 100 problems until I quit cp.
Current progress: 40
| # | Date | Comments | Feedback | |||
|---|---|---|---|---|---|---|
| 1 | July 2021 | Summer 2021 Round 1 | First problem I ever set. It asks you to find the number of bubble sort iterations to sort a given array | Good Bad | ||
| 2 | July 2021 | Summer 2021 Round 1 | This problem is a very standard problem which asks to find sum of gcds across all subarrays | Good Bad | ||
| 3 | July 2021 | Summer 2021 Round 1 | This is a quite ok digit dp problem which asks finding the number of numbers s.t. the maximum frequency of some digit — minimum frequency of some digit = $$$k$$$, where $$$k$$$ is given | Good Bad | ||
| 4 | July 2021 | Very standard problem that asks to find how many subsets of $$$a$$$ are anagrams of $$$b$$$ | Good Bad | |||
| 5 | July 2021 | Summer 2021 Training round | Very standard problem about stirling numbers of the second kind | Good Bad | ||
| 6 | July 2021 | Summer 2021 Training round | Very standard problem asking for sum of gcds across all arrays of length $$$n$$$ and values in range $$$[1,m]$$$ | Good Bad | ||
| 7 | August 2021 | Summer 2021 Round 2 | Problem asking for number of subarrays with given median | Good Bad | ||
| 8 | August 2021 | Summer 2021 Round 2 | Problem asking for number of subarrays such that if it contains a color, it must contain every color of that kind | Good Bad | ||
| 9 | November 2021 | Autumn 2021 Round 2 Div 2 | A simple greedy problem | Good Bad | ||
| 10 | November 2021 | Autumn 2021 Round 2 Div 1 | I am only a coauthor in this task. The initial problem idea was by Gheal. A quite standard problem on PIE | Good Bad | ||
| 11 | December 2021 | Autumn 2021 Training round | This problem is a math problem that turned into a troll problem with partial scoring. I don't take credit for its idea, only for making it appear in a contest :clown: | Good Bad | ||
| 12 | December 2021 | Autumn 2021 Training round | Educational problem on d&c dp optimization | Good Bad | ||
| 13 | January 2022 | Winter 2022 Round 1 Div 1 | A nice problem, but my sqrt optimizations were just overkill since very simple solutions were using just PIE. I felt bad after finding out all my work optimizing the dp calculations was useless :( | Good Bad | ||
| 14 | February 2022 | CookOff | You set the cancer dp I couldn't solve. First and hardest problem I set on codechef | Good Bad | ||
| 15 | March 2022 | Lunchtime | A quite simple problem about both FFT and PIE. Since I didn't know FFT when I submited this problem on codechef, I'd like to mention Um_nik for improving my initial solution | Good Bad | ||
| 16 | March 2022 | Spring 2022 Round 1 Div 1 | It's a quite simple dp problem with an optimization | Good Bad | ||
| 17 | March 2022 | Spring 2022 Round 1 Div 2 | Quite straightforward couting problem with sieve optimization | Good Bad | ||
| 18 | March 2022 | Spring 2022 Round 1 Div 2 | Quite straightforward problem on monotonic stack | Good Bad | ||
| 19 | April 2022 | Starters 35 | Very cancer dp... It's good to make your enemies hate you more tho... | Good Bad | ||
| 20 | April 2022 | Starters 34 | I think it's one of my best easy problems | Good Bad | ||
| 21 | July 2022 | Lunchtime | Very similar statement to the one in Mex Pairs. Here we write an obvious dp and optimize it in a very interesting way | Good Bad | ||
| 22 | July 2022 | Round #804 | First problem I set on codeforces. My initial solution was $$$O(vmax \cdot log^2)$$$, but thanks to valeriu, it was reduced to $$$O(vmax \cdot log)$$$. I think it's one of my best problems | Good Bad | ||
| 23 | November 2022 | Round 3 | In my opinion, it's a very good D2B problem, solution combines a simple observation with a not so trivial optimization to keep the balance | Good Bad | ||
| 24 | November 2022 | Round 3 | I don't know why this problem was hated in the rate problem section, I think it's a good task that requires both some thinking and implementation | Good Bad | ||
| 25 | November 2022 | Round 3 | This task was added one day before the competition. It looked a good difficulty jump from C, but the distance to E was a bit too large. I apologize for this | Good Bad | ||
| 26 | November 2022 | Round 3 | After realizing how to compute the cost in a smart and flexible way, you can easily use a BIT to solve it | Good Bad | ||
| 27 | November 2022 | Round 3 | I think it's one of the best tasks in the contest, there are lots of solutions to it, mostly all at the same difficulty level. There also exists a $$$O(nlog^2n)$$$ solution, which I have implemented, but we used the $$$O(n^2)$$$ version to fit for $$$F$$$ position | Good Bad | ||
| 28 | November 2022 | Round 3 | It has a hidden refference, but leaving that behind, I think it's a quite nice counting task, and the solution is very unexpected | Good Bad | ||
| 29 | November 2022 | Round 3 | First problem I set on codeforces that didn't have any solves during the official contest. It is supossed to be a hard problem with a quite long implementation, but looking at upsolving submissions, implementation can actually be quite short :) | Good Bad | ||
| 30 | March 2023 | Finally setting on codechef again :) Yet another counting problem that almost went unsolved ( which is very surprising ). Am I too evil to ask variable mod FFT? Do people template variable mod FFT? I don't know. Orz jeroenodb first and only solve to this problem xD | Good Bad | |||
| 31 | May 2023 | A problem I created because we desperately needed a problem for D2B position. I don't think it's great but not bad either. | Good Bad | |||
| 32 | May 2023 | A problem that was intended to be in CodeTON Round 3, but we didn't need it there so we used it here. I came up with this problem while I was in car xD | Good Bad | |||
| 33 | August 2023 | Junior challenge 2023 | Best statement of 2023. Find sum of $$$choose(n,i) \cdot i^k$$$, $$$n \le 10^9, k \le 5000$$$ | Good Bad | ||
| 34 | August 2023 | Junior challenge 2023 | This problem was intended to be in CodeTON round 3 as D, but was removed few days before the competition. This version is however buffed. | Good Bad | ||
| 35 | October 2023 | RCPCamp 2023 Iasi day | This shares a similar statement to Doping, but is much easier. | Good Bad | ||
| 36 | December 2023 | Codeforces Round 915 (Div. 2) | Initial name was Adam Lying Face :( | Good Bad | ||
| 37 | December 2023 | Codeforces Round 915 (Div. 2) | I was expecting this to be as hard as E | Good Bad | ||
| 38 | December 2023 | Codeforces Round 915 (Div. 2) | Used to have some funny refference and a nice drawing but some magician removed both of them. | Good Bad | ||
| 39 | December 2023 | Codeforces Round 915 (Div. 2) | Was intended to be D1B in round #875. We ended up removing it because it was too hard. | Good Bad | ||
| 40 | December 2023 | Stelele informaticii 2023 | Probably D1B~C | A rejected proposal from round 875. | Good Bad |
leafeon.
What is quickest and the slowest amount of time you have taken to develop a problem? And is there any correlation with effort vs quality in developing problems?
I think it took me like 3 days to code the solution for Mex Pairs. About shortest, I don't know what to say.
Of course putting much effort into developing something makes it better.
What do you do when a problem idea of yours turns out to require a much harder solution than expected? When it's the contrary, I think it's usually not so bad, as this "easier solution" in this case is usually one that is much more interesting, but when it gets harder than expected, I often put it off to when I get a better understanding on what may be a good solution. However, as every work has to be done sometime, I know it can't be delayed infinitely, or else I'll forget the idea. So what would you do in this case?
UPD: One more question. In problemsetting, do you make the solution after the problem, or the problem after the solution? This is a very constant dilemma that would be going on every problemsetter's heads, because usually the former makes fun problems that are hard to prove, and the latter makes problems that are easy to prove but is trivial/standard. Which approach do you take?
It's better to come up with the problem and then find the solution, it's more natural. Of course the reverse works, but it sometimes leads to an unnatural mess.
I don't understand anything from the first paragraph to be honest, I think you're asking what am I doing when easier solutions exist than my intended one ?
Nah, I mean when a problem idea (without a clear solution) started simple, but turned out to be much complex than what you imagined first. What would you do when this happens?
How can one write concise statement for Calorifer
One does not do that. Only God can.
How do you come up with such hard tasks with peak rating at 1600? I am 2100 rated, but the hardest task I invented was 2900.
orz, you are a really talented problemsetter.
Bcs I do be trolling contests ( I don't even do contests on codeforces, mostly olympiad training ). I am also way better at solving tasks given more time to think rather than 2 hours. I could come up with the solution on my own to all of the aforementioned tasks.
Admitting that you troll contests is a lie to anyone, you only trolled (so far) two (codeforces rounds)
I am not the OP, but I think I might be able to answer, I do be making problems above my ratings. I think problemsetting is basically a whole different skill compared to skills in contests. As tibinyte said, problemsetting is not done under a short time limit. So, they basically have a lot of time to think of new ideas and proofs for those ideas. Also important to problemsetting, would be an interest to a wide variety of algorithms/problems/concepts, sometimes ranging far above their own ratings. It would indeed be strange for a 1300-rated person to solve 2600-rated problems to practice for contests, but it is completely fine for one to think of the same problem deeply because it is interesting enough to research further, and these "interesting problems" are a lot of help later on in the problemsetting process. This help includes a whole problem idea, made from an existing one.
So, yeah. A problemsetter can make problems, and the problems' difficulty isn't always correlated to the setters' ratings.
The hardest task I invented is -inf rated :(
How can I submit a contest proposal when my rating isn't 2100+ ?
By default you cannot, but if you were author in some other round, you gain permissions.
How can I be the author in "some other round"?
Have friends that can propose contests.
Thanks!
As a problem-setter myself, I find it very impressive that you are able to make an entire D1+2 round by yourself (except A). I know someone already asked you how you create such hard tasks as a green, but I am still in disbelief... You mentioned that you do not do Codeforces much, so what do you think you rating would be if you do Codeforces contests regularly.
And how do you come up with such hard problems? For me, I find it very tough to make hard problems as when I come up with an idea, I will think that it is not solvable after some time and stop thinking about it. As the result, all the problems I set are those that I can solve in a shorter period of time, and hence easier. Can you share how you came up with the recent D1+2 H?
So what do you think you rating would be if you do Codeforces contests regularly
I've been really bored this summer so I did some virtual contests.
Can you share how you came up with the recent D1+2 H?
Well as you can tell, H is very similar to F. The main inspiration for both these task were Phoneix and computers and PWCA. H has a similar operation to the one in PWCA, but counting it for substrings is very interesting, and is much harder. Not to mention the implementation of H is a very hard task just by itself :)