1 / 7
Mar 2009

We plan to extend the scoring formula for the main SPOJ ranking.

Currently it looks like this:

user's_points = classical + challenge
classical:
= 80/(40+number_of_people_who_have_solved_the_problem) points
challenge:
for the top score in the challenge:
= 3 points
for any lower score in the challenge:
= (user's_score relative_to_the top_score) points

Our suggestion is:

user's_points = classical + challenge + tutorial + partial + publ_contests(?)
classical: (the same as before)
= 80/(40+number_of_people_who_have_solved_the_problem) points
challenge: (the same as before)
for the top score in the challenge:
= 3 points
for any lower score in the challenge:
= (user's_score relative_to_the top_score) points
tutorial:
= 40/(40+number_of_people_who_have_solved_the_problem) points
partial:
= 80/(40+number_of_people_who_have_solved_the_problem) * (percentage_of_solved_test_cases)
publ_contests:
We would like to take public contests into consideration in main ranking. How do you like this idea?
These contests would be Open Contests XXXX, High School Programming League, Olympiad in Informatics, etc.

All suggestions and new ideas are welcomed smile

  • created

    Mar '09
  • last reply

    Dec '20
  • 6

    replies

  • 1.6k

    views

  • 7

    users

  • 1

    link

I'm not sure including public contests would be a good idea. You definitely wouldn't want something that only gave points to people who participated at the time, otherwise the rank list would become unfair (some people not able to get points that others can). If you were planning to leave the public contests open after they are completed so that it would remain fair, I guess that could work, but most of the time the problems are copied into the main list anyway.

12 days later

Yeah, contests shouldn't count. And yeah, it would be nice for tutorial problems to be worth some points.

1 month later

we mean public contests (or better problems archives, without "end-time") like: CH24 (spoj.pl/OI3)

instead of giving many kind of problems (like classical, challenge, etc), we introduce new contests (or better problems archive) as a part of SPOJ system, and we plan to include it into global ranking system,

is it ok?

btw: it would be nice to split OI-contest into national categories (as problemsets),

The inclusion of public contests may be a good thing, the contests usualy contain well-thought and well-tested problems (I guess). But the "copies" of the public-contest problems that ended up in the public SPOJ lists will have to be merged with the contest versions.

I would like to know more about the scoring system you are planning to use for the contest archives. It may be a little complicated with contest like SPOJ's 2008 open contest, which had classical problems mixed with optimisation problems.

11 years later

Suggested Topics

Want to read more? Browse other topics in Online Judge System or view latest topics.