Allow user to calculate "what if" rating changes using fake entry

[Donavern]

[Enhancement]

What if users are able to "provide" a fake entry to the leaderboard just for this calculation to take place, you need to provide a real id for the rating algorithm to get data to calculate with, you can select which questions you have completed and the time you took. The user can then confirm and the system will predict what the rating change for this user would have been if he was placed at that position.

This does not affect the existing leaderboard in anyway.

Example:
Case 1: I did the contest, solved 3/4 with 2 WA using 1 hour. I want to know what if I did not have 2 WA? I use this feature to calculate my position and rating change if I solved 3/4 with 0 WA using 50 minutes. (1 WA == 5 minute penalty)

Case 2: I did not do the contest. I want to see if I solved 3/4 with 45 minutes, what is my ranking and rating change?

P.S. Without this feature, for case 2, I have been doing the contest after, manually finding people with similar rating at the supposed ranking to estimate what my rating change would have been, but I am not sure if this is accurate.

[baoliay2008]

@Donavern, thank you very much for opening this issue.

This is a great idea, and I have thought about it before. In my opinion, this is similar to the virtual contest feature on LeetCode. Basically, we can estimate someone's rating based on their ranking in a virtual contest, taking into account how many problems they solved and how many WAs they made.

Regarding your P.S., this is not entirely accurate because rating changes depend on your ranking, current rating, and the number of contests you have participated in. Without this feature, you would need to find someone with a similar rating and number of contests participated in at the supposed ranking, which is not easy to do now.

[Donavern]

Hi, thanks for informing me regarding the rating algorithm! I think this is a cool feature but it probably won't be popular, so hopefully if this is implemented, the load it will cause should be on the low side.

For case 1, if user wanted to see how much their rating would change if they did better, user could just add to the given estimated. So I don't think they would really use this feature.

For case 2, there shouldn't be too many people utilizing this at the same time.