Hello everyone ! some times we had seen that codeforces took much time (sometimes few days) to update the problem rating of recent contest problemset , and i personally waited for days to see the rating of problem that i had solved in the live contest as soon as possible just to see did i crossed my limits (obviously when i solved more problems than my usual limits),and i hope you might also have experienced such thing.

so recently i had created a project on supervised machine learning algorithm that predict the rating of recent codeforces problemset with the help of all the past data of codeforces using codeforces api.

i know it can't predict as accurate as codeforces but it is very close ,like the overall rms error of model is approx 100 for div1 and div2 and 60 for div3 contests.

model has some other feature too like

• accuracy of problem (accepted sol/total submission)
• submission ratio (accepted sol/ total participant)
• average time taken to solve a problem (sum of time taken by ac participant/no. of ac participant)
• no. of user who tried a problem but can't solved it

for div1 contest

for div2 contest

for div3 contest

guys please share your reviews or suggestion toward this codeforces tool and also would you like to use this tool if i host it on internet.

This is the original version of problem ,during the contest I tried to solve it using some combinatorics logic but i can't and for this version of problem we can just iterate over all possible PIN codes because constraints are small, yes i didn't able to get this logic during the contest, but i am curious to know the approach to solve this question if PIN size will be of 10 digits.

I write a nlogn solution for this problem but it is giving tle how to solve this using KMP or other string matching algorithm. Help will be really appreciated . Thanks :)