Currently, each question in codechef long contests (except the challenge question) carries 1 point, be it easy, medium or hard. My suggestion is that that there should be different points allotted to problems of different difficulties. For example, the hard questions could be worth 3 points, medium ones worth 2 and easy ones worth 1. By the second day of of the long contest, any competitor can easily gauge the difficulty of the problems by looking at the number of submissions. Thus, assigning different points will not really reveal anything which the competitors don’t really know in the contest. Such a system of different points has many advantages.
- Fewer ties. I know there is the challenge problem. However, take the case of January Challenge, where there were some 13 ties at the 15th position.
- More motivation for competitors to do harder problems. this improves the competition quality.
I would appreciate that admins consider this suggestion.
In my view, the current system is not so bad. The idea behind such a system is that “If you can solve a hard problem, then you must be able to solve an easier problem.” Sometimes, you manage to solve a hard question. But the contest nature forces you to solve the easier ones so that you can show the difference between you and another person who has solved an easier question. Besides, all other online judges have this system of allocating points to each question. It is good to experience another system which works differently. And if you bring the points system, the setters would have to undergo the additional burden of judging how much points to give to a certain question. This again would lead to discussions after the contest where people feel one particular question should have been given more points than the other.Also, the challenge problem can no longer be given the (Your Score/ Highest Score) normalization. Well, you could always scale it up, but again, how much will you scale it? Challenge problems do not have difficulty level. The difficulty level is set by the top scorer for that problem, following which everyone tries to reach that score. This month’s long challenge had an easier question, yet the really high scores obtained by those on top made it a difficult one.
I guess the system is stable at the moment. Maybe they could experiment it for a month, but I feel they should let it be as it is
I missed this discussion before…
…my late answer - to find out before the contest which problem is easier then other is not easy to set in advance (I saw it multiple times in CodeForces, when people complain that problem with 2000 points was easier then one with 1500 points and so on).
The idea of dynamic score might be interesting. But of course with more levels. Me personally, I do not like those levels, I’d prefer more continues function.