- 分享
- 2021-12-27 22:00:00 @
AOCode Rating System
AOCode Rating System is a system that was created in December 2021, which is to calculate participant’s rating after a contest. AOCode Round #3 & AOSC #2 is the first contest to use it and the first \(Rated\) contest.
Now the system only supports the contests using OI mode, which means that ACM/ICPC mode contests won’t be \(Rated\) even if it is an Official Contest. And for the Running Mode, Multiple-Run Mode isn’t supported, so only Normal-Run Contests could be \(Rated\).
The Calculating Method could be described as down below:
Imagine that \(ori\) is your Original Rating, the contest has \(n\) problems, the \(i^{th}\) problem has a difficulty of \(D_i(1 \le i \le n)\), and your score on the \(i^{th}\) problem is \(S_i(1 \le i \le n)\).
First get your performance \(\displaystyle perf=\sum^{n}_{i=1}\dfrac{D_i\times0.7*S_i}{100\times\log_{10}(\dfrac{D_i}{12})}\).
Then, if the contest duration is \(t\) hr(s), we need to make \(perf=perf\times(\dfrac{3}{t})\).
Then compare your \(perf\) with your original rating \(ori\):
\(perf<ori\):
Your new rating \(newr=ori+\dfrac{{perf-ori}}{{\log_{10}(perf)}}\).
\(perf>ori\):
Your new rating \(newr=ori+\dfrac{perf-ori}{\log_{10}(perf \times 1.2)}\).
Then we calculate your other things about the \(Rating\), like the \(\rm Max.Rating\) and so on.