I don’t think i should mention the name, but i saw a guy who got 680 long and 321 short global rank by just giving 37 solved problems. Whereas i’ve also seen coders that have solved about 100 problems and have rank either around this or beyond this one (mentioning only the contests and cookoffs).
Please tell me how this system of ranking works ? Is there also a smarter way out in this ?
The rating points and the ranks in CodeChef is calculated for the problems solved in the live contest. CodeChef organizes a long contest, a short contest known as a COOK-OFF and a lunch time contest (primarily for school students) each month.
And there is no change in ratings if you solve a problem from the practice section.
The guy with the rank you mentioned may have solved all those 37 problems in the contests only. On the other hand people with more questions solved must have solved fewer problems in the contests.
The particular user you are referring to is @kgcstar. He is the winner of AUG13. He is somewhere around the top of the rankings table for the contests he has participated in.
If you have read and understood hoe the ratings are calculated, then you can now understand why his ratings and/or ranks are good.
This also shows something: the more skilled you are, the more problems you solve and the higher you will be ranked and faster your ratings will improve. And improvement in our ratings can only mean improvement in our skills!!
@tijoforyou : yeah i read it little bit. Means the essence is that he must have (and we must) solved the problems efficiently and at the same time beating others. am i right ?
Then, you must see the rank of ACrush. He, only participated in two cook-offs one in way bay March 2011 and the last cook-off. And solved only 9 problems and he got 32 rank in short contest.
So, moral of the story ratings and ranking don’t depend on the number of the problems you solved. It depends of many factors as stated by others.
PS: It depends on how well you performed in a particular contest.
You need to be placed higher in the rankings for a single contest, so that there is a massive increase in the rankings.
Solving ten problems in a single contest is good. Solving ten contest problems, one each in ten contests is real bad (given that, MANY have solved more than one problem per contest on average)! The ratings WILL reflect this idea!!
Yep. Solve more problems in a single contest. Get placed higher in the rankings table for that contest. Then, you will have great increase in you rating for that month.
Like I said, solving 10 problems in one contest is not similar to solving 10 problems in 10 contests.
It is not the overall total number of solved problems that determine the ratings. It is the performance, relative to others in each of the contests that determine the ratings.