We do need help for a lot of things actually. We are now working on getting the ratings and volatility of users faster. Then we need to work on getting the entire system hosted. Then I guess we have to work on scheduling.
Also there are lot of inaccuracies in the predictions. I think this is due to users with zero accepted solutions who are not in rank-list but are accounted for during rating calculations.
We need a fast method to get such users who are not in the rank-list.
For Example the first problem had a accuracy of 20% I guess for COOK144A. So many users had zero accepted solutions. And therefore the error become significant for users who had ranks above 40-50.
Getting this error on running main.py
Traceback (most recent call last):
File “D:\codechef_rating_v0-master\main.py”, line 11, in
R = RankListParser(contest)
File “D:\codechef_rating_v0-master\ranklist_parser.py”, line 17, in init
page_bar = center.find(‘div’, class_=“paginationbox”)
AttributeError: ‘NoneType’ object has no attribute ‘find’
For Those Interested I Generated A Predicted Rating Sheet just now (based on the predictor above), There might be some errors (due to users with zero accepted solution) but you guys can get an idea.
Yeah, probably. There’s a known issue when people who have submitted solution but none got AC are not detected by parser as they don’t show up on leaderboard. But i guessed that in long challenges almost everyone can get some solution AC (even partial) and show up on leaderboard.