Have you noticed that in Div2-only matches there are always unrated people in top? And later most of them are abandoning that account? Even though they could participate unofficially with their main account, they are stealing rating from Div2 people.
Some of the most obvious heroes: superpears siIlycross multiset Bigsophie
Multi account isn't always explanation: When you are rated on the top 5 of Div 2, you are often too good for stay or come back on div 2, so I think it's logic that the first are often unrated. Bmerry is a excellent exemple.
But you must be right for some accounts. In fact, the statistics are in your favor (a lot of unrated who were on top 5 have only done 1 match). Nothing prove that this is due to multi account, but it can be the reason.
I wrote in the post of the last contest:
Each time we have 2 division contest simultaneously, we have less than 500 coders in 1 div. Now, we have a lot of new coders in first place each time we have only second division. Are they the same with a new user? **** top 20 ---> 13 new users. Good Luck in 1 div!!!
But I get -9 in this comment, maybe because I am green.
I tried to find out the solution to the problem :
Problem -> Div1 coders participating in div 2 rounds using multiple accounts.
Reason -> They want to taste the contest flavor and not participate in Virtual Contest or possibly spoil the ratings of Div 2 coders.
Possible solutions ->Force them to participate in Virtual Contest mode .
If there is an another rating called something like "Div 2 contests ratings " for Div 1 participant which only rates them in Div 2 rounds , that could minimize these fake accounts because of the added incentive.
One thing I cannot understand is why there are only Div 2 contests . If the author managed to prepare only easy problems , then they are easy for all the participants in Div1 and with the Elo system of ratings used it will be balanced. But what happens if a Div 1 coder only participates in Div 2 rounds ?? Will it be fair to others ?? The answer I think is simply NO , because then there is nothing like Div 2 rounds only then and he will not be able to know before hand whether the problems given are easy or tough..
Please leave your comments on these....
Why, the simplest solution is to allow them participate in div2-only contests, but rate them separately (as when contests are separated for two divs), only among them.
I.e. it will look like ordinary separated contest but with the same five problems for both divisions. Surely div1 will do it faster, but that does not matter much. Also their rating change maybe bit less accurate, but this do not matter much too.
Separate round with the same problems is not a solution.
If discussed proceder is cheating (I think YES, but I'm not be a judge) then motive is a desire for victory, and there is no solution.
My suggestion: calculate rating for other div2 participants without first-round-participants.
Another motive can be increasing of city/country-rating.
Okay, and look at their last visit. It was exactly, when the competition was.
This is mostly caused by the simplicity of a registration here.