Wouldn't it be better if the score estimation function takes into account that it is difficult to improve when you have a higher rating, also some weight should be given to the highest rating the programmer has attained in the period, etc?
Nothing fancy, but there can be better functions. For example, ${y} = ((cur - ini)^2 / (max - ini)) * 100$, takes into account that it is difficult to improve on higher ratings than on lower ones. We can improve this further (or come up with something different) to account for 2/3 other factors.
Wouldn't it be better if the score estimation function takes into account that it is difficult to improve when you have a higher rating, also some weight should be given to the highest rating the programmer has attained in the period, etc?
Nothing fancy, but there can be better functions. For example,${y} = ((cur - ini)^2 / (max - ini)) * 100$ , takes into account that it is difficult to improve on higher ratings than on lower ones. We can improve this further (or come up with something different) to account for 2/3 other factors.