JRS is a library that ranks players based on their relative performances and uses these rankings for matchmaking and building leaderboards.
The current implementation is based on The Glicko-2 algorthim developed by Professor Mark Glickman at Boston University. (http://math.bu.edu/people/mg/)
Glicko-2 is differentiated from many previous ad-hoc systems in that it is a mathematically rigorous system that treats rankings as hidden statistical variables and attempts to deduce them based on statistical analysis of wins and losses over time.
It has been accepted by the United States Chess Federation and a number of other ranked head-to-head sports as their official ranking system.
Glico-2 in its current form only formally supports 1:1 games. This project will include some experimental ad-hoc n:n support. Professor Glickman has expressed some interest in properly extending Glico-2 to n:n. If/when that happens we will likely move to his solution.
This open source code project has been pursued with the explicit approval of Professor Mark Glickman.