MikeMirzayanov's blog

By MikeMirzayanov, 9 years ago, translation, In English,
I propose to discuss here the issues and troubles associated with Codeforces Beta Round #4.  On our side, we identified three difficulties: 
  • Slow judging - apologize for the large queue, to the next round, I assume significantly increase the speed of judging;
  • Runtime errors on Python solutions - apparently the decision they want to load some library, which is prohibited by security reasons;
  • Sometimes incorrect runtime errors - apparently the result of the antivirus on testing servers, they will be turned off on the next rounds
Also express your impressions about the round.

P.S. And by the way, the round tutorial is waiting for a volunteer. It is desirable that it will be one of the leaders of today's competition. The tutorial should be in Russian and in English. It will be published on the homepage and later will be available via special link from the contest page.

UPD. The ratings have been updated.
UPD 2. Thanks to KudryashovIA for his help in making contest.
 
 
 
 
  • Vote: I like it  
  • +6
  • Vote: I do not like it  

9 years ago, # |
  Vote: I like it 0 Vote: I do not like it
Why there is no possibility to paste code instead of sending file? Probably I sent wrong file and I got WA. Also I would like to see my code that I sent.
  • 9 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it
    Ok, I noticed now submit button :) However I dont agree to get -49 points to rating. I solved problem D very quickly in practice mode. I had correct solution, but small mistake and I would fix it in time, but I didn't have verdict about any of my submissions for problem D. Also I had to wait a lot for verdicts of problems A and C (I forgot about 2 in A and chose wrong compiler in C). Could you cancel me from standings or will you clean the whole rating after beta mode?
    • 9 years ago, # ^ |
        Vote: I like it +1 Vote: I do not like it
      I agree that grading was slow, but everyone had the same problem.
      • 9 years ago, # ^ |
          Vote: I like it 0 Vote: I do not like it
        Yes, but for example I don't take part in competition like top coder where you have to wait for the verdict till the end of the round. I am used to get verdict immediately. I don't think very carefully about my solution, maybe in future I will do something abut it, but now I just write and send and if the judging was correct I bet I would be higher in standings, because I think that  some of the contestants didn't have idea how to solve the last task. If this rating is only temporary that's not a problem, but if I had to start with mine after beta mode I don't think it's ok.
  • 9 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it
    You can paste your code into the online editor in the "submit" section.
9 years ago, # |
  Vote: I like it 0 Vote: I do not like it
Ok, to sum up my previous comments.
- I think it will be good to see code that I sent to judge;
- In my opinion slow judging was more harmful for some contestants (including me) than for the others;
- I would like to know if after beta mode everyone will start again with gray color if not I would be grateful if I could be removed from standings of beta round 4;
  • 9 years ago, # ^ |
      Vote: I like it +2 Vote: I do not like it

    -100 is not so bad. The rating system has such organization, that it is OK to loose some rounds, everyone will get his own rating after awhile. Read about rating system on TopCoder, i think this system is something like topcoder's.

    • 9 years ago, # ^ |
        Vote: I like it 0 Vote: I do not like it
      Supposing I deserved it, that would be ok. I don't mind -52 in round 3.
      That's not true everyone had the same problem. For example contestants 1 and 2 sent solution to problem A at time x. After 20 minutes contestant 1 was informed his solution was incorrect and resubmit it and got acc at time x+20+1. Contestant 1 sent correct solution for problem B at time y whilst contestant 2 send correct solution for this problem at time y+30. It was all that these 2 contestants sent and in final standings contestant 1 has time x+40+1+y, and contestant 2 has time x+y+30 and is higher in standings than contestant 1, whilst in normal situation contestant 1 should have time x+20+1+y and he should be before contestant 2. It was one example, second could be when one contestant has completely no idea how to solve the problem, whilst the other knows the solution, but has a bug and wasn't informed about it during the contest. It is ok with top coder rules, but not with acm icpc rules. I think this round misrepresents the rating.
      If I knew about this I wouldn't take part, but I found out after I sent solution to problem A.
      I am not upset, because server is in beta now, I just would like to know if rating based on such rounds is beta too, and will be cleaned after beta mode.
9 years ago, # |
  Vote: I like it 0 Vote: I do not like it
About RTEs in Python, as I've already pointed out in a comment to another post, one of the modules restricted/not supported appears to be collections. This is the home of the deque and defaultdict containers in Python and is therefore important in solving a lot of problems. Of course, one can work around using it; but I don't see anything in the module that would cause security problems (there's really nothing there that does any I/O or allows system calls).

This also raises the question as to what other modules are similarly restricted, and makes submitting in Python during an actual contest a risky option.
9 years ago, # |
  Vote: I like it 0 Vote: I do not like it
I had one CE in problem B which has not been taken as penalty in my score. I think in ACM-ICPC , CE is marked as penalty toward team scores. 
So, I thought you might have a look on this, if this is any kind of bug.

  • 9 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it
    We don't count CE as +20 for penalty.
  • 9 years ago, # ^ |
      Vote: I like it 0 Vote: I do not like it
    Compilation Errors don't result in penalty at the ICPC World Finals, but I don't know if this is due to a recent rule change or if it's an old tradition.