Offer and Demand of problems

The problems with higher demand in CodeChef are the hard, medium-hard and approximate problems. We have some problems in the queue with difficulty less than or equal to medium. However, if your idea is very interesting, it can be moved quickly to the top. It may sound weird, but good cakewalk problems are also on demand.

Regardless of the topic we give more attention to the originality and ideas involved in solving the problem. Of course for a given contest we try to maximise the range of topics involved.

The standard queries on a single array and trees are kind of over-exploited. If you use more fresh combinatorial structures the probability of using the problem becomes high.

The following types of problems are always welcome, I included some “model” problems

  • ad-hoc: Is nice when the problem is very original and the contestants have to create the theory. Take a look to ALICE, ADASHOP, LVMFFN, KNICOV

  • games: in addition to standard grundy games, there are many other ideas, e.g misere games, find the longest game, maximise profit, … from my own hand: PWNCHESS, TWEGRA, ADAPWN, ADAPWNS, BINIM, BINIM2

  • constructives: DFMTRX, CENTREE. We’ll have another one in March Long

  • interactives: GUESSPRM, DYNAMO, there was one from lewin last year, does anyone remember the name of the problem? btw, we’ll have another one in March Long :slight_smile:

  • graphs: connectivity, flows, matchings. So far I haven’t received a 2SAT or bridges problem! The hardest one: CUREK, BALNET is also a nice problem.

  • geometry: They tend to be hard to implement, from my own hand: ADAFARM (no one solve it during a short contest), DDART

  • advanced topics: matroids (HAMEL, CNNCT2.), matrix theory, advanced graph theory, number theory and cryptography, combinatorial optimization, …

  • strings: So far I’ve used only one in about 6 contests!

  • dp tricks and greedy: One simple problem that I liked: CCC

  • counting: I’ve seen some ffts last year, I haven’t used a groups theory problem last year though.

Data Structures problems are not my favourite, but we always include some of them in the problem set.


I really liked Circular Merging (CIRMERGE) question that came in July Long last year (I think I maybe wrong).
But I think Data Structure questions involving Segment Trees or BIT should be included as Long Challenge gives a chance to learn new topics and this is a very good initiative to start and easy-medium graph questions would be fun to solve as well like Even-Edges.

Would synthetic geometry problems be encouraged? (Those regarding circumradius, inradius, etc.)

Maybe … I always wanted to make a compass and straightedge problem (I remember there is an old tie-break problem about it). Also it can lead to some interesting constructive problems.