randomguy999's blog

By randomguy999, history, 2 days ago, In English

Nowadays I got confused what kind of problemset should I practice, for example sorting the difficulty rating of the problemset, or solving the latest problems, etc. Can someone give me an advice for solving what kind of rated problems, what kind of problemset?

  • Vote: I like it
  • +6
  • Vote: I do not like it

»
2 days ago, # |
  Vote: I like it 0 Vote: I do not like it

Solve more 800 problems and you will be 1000 soon.

  • »
    »
    2 days ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    but my rating is 1077, i think a better problemset can be advised. thank you though.

    • »
      »
      »
      2 days ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      It was a sarcastic comment. Obviously, there is no need to solve 800. Think about it a little bit deeper.

      • »
        »
        »
        »
        2 days ago, # ^ |
          Vote: I like it 0 Vote: I do not like it

        okay

»
46 hours ago, # |
  Vote: I like it +1 Vote: I do not like it

Hi! I think it is hard to find a perfect problemset without knowing your background and previous knowledge, but I can share what I suggest to beginners:

  1. If there is an Informatics Olympiad course in your country/language, just follow it. Or use USACO Training course. Main idea here, is that there is a lot of effort into teaching the basics in Informatics Olympiad, so it is always good.

  2. A personal recommendation of mine: https://mirror.codeforces.com/blog/entry/65133. I like this roadmap since it is divided by levels and topics, so it is pretty easy to follow. Also I like a lot of the training ideas of the author.

  3. Follow the CPBook3 (which is free online) problems. You can start from chapter 2. and try to finish the problems there. I personally like Uva problems, since they are very similar to the icpc problems in my region and they have helped me a lot. Main downside, is that you have to use UVA xd

  4. Just do problem A, B, C from the last contests. Idea is that you try them, then if you fail, you read the editorial and implement them. If it is a new topic, you learn the topic. After some weeks repeating you should see some progress in real contests.

  • »
    »
    36 hours ago, # ^ |
      Vote: I like it +5 Vote: I do not like it

    Thank you sir! Your advices are perfect! tiojuan orz!

  • »
    »
    35 hours ago, # ^ |
      Vote: I like it +5 Vote: I do not like it

    point 3, are problems similar to icpc have impact on your cf rate?

    • »
      »
      »
      23 hours ago, # ^ |
      Rev. 2   Vote: I like it 0 Vote: I do not like it

      It has helped a lot for sure. For me, ICPC training is a lot more of standarized problems/ classic problems, that do sometimes appear in codeforces. On the other side, cf contests require usually more intiution and proofs that you I had to learn through cf problems.

      The most optimal thing to train for cf rating, is to do cf problems. Also Codeforces platform is super good and has editorial for every problem!! So another point for cf problems.

      But if you get bored of cf problems, icpc problems are also fun and will improve your ability. And what matters at the end is that you do problems and that you don't get bored xd

  • »
    »
    4 hours ago, # ^ |
    Rev. 2   Vote: I like it 0 Vote: I do not like it

    Can u please provide a link to the book u mention?

»
35 hours ago, # |
Rev. 2   Vote: I like it 0 Vote: I do not like it

I think it is hard to find a perfect problemset without knowing your algoritham knowledge and your problem solving skill.But i can give you some suggestion

Practice strategy

  • Two types of practice are needed
  1. to build critical thinking / problem solving / application of concepts known.
  2. to learn new concepts, tricks and standard problems.
  • For the first kind of practice you have to solve 40 each in the rating range 800-1400. You can skip a level if you are way too comfortable with it .So do 40 each 800->900->1000->1100->1200->1300->1400.
  • Now at this point you would be able to do A,B and C sometimes.
  • Now do 10-15 more of each if you want .

Topics needed to cover(>=1400)

  1. stl
  2. Basic maths , basic logic
  3. Searching,Sorting , greedy , two pointers ,sliding window
  4. Implementation , stl based
  5. Binary search ( very important and a saviour)
  6. Number theory ( sieve , gcd , factorisation , prime numbers etc )
  7. Basic bit manipulation and basic combinatorics. (very basic)
  • »
    »
    34 hours ago, # ^ |
      Vote: I like it 0 Vote: I do not like it

    Thanks sir! I would also want to ask which problems should I solve. From recent contests or from old contests?

    • »
      »
      »
      23 hours ago, # ^ |
        Vote: I like it 0 Vote: I do not like it

      Recent problems are better than old ones. With some exceptions of classic problems, the default should be to pic recent contests/editorials

      • »
        »
        »
        »
        14 hours ago, # ^ |
          Vote: I like it 0 Vote: I do not like it

        I have started doing 800 rated problems from recent contests, after doing 40-50 problems from every level I will increase the rate of problems. Is it a good strategy, sir?

        • »
          »
          »
          »
          »
          2 hours ago, # ^ |
            Vote: I like it 0 Vote: I do not like it

          Go for A and B Div 2, 800 might be too low for you.

          • »
            »
            »
            »
            »
            »
            2 hours ago, # ^ |
              Vote: I like it 0 Vote: I do not like it

            what's the secret of Marckess ?

          • »
            »
            »
            »
            »
            »
            110 minutes ago, # ^ |
              Vote: I like it 0 Vote: I do not like it

            Thanks sir! I will just grind some 10-15 problems from 800 rated problems (div. 2 a and div. 3 a, b problems sometimes), then i will grind the next levels!