A rating system is just a function $r(u,v,w)$ where $u,v$ are the ratings of the competitors and $w$ is the result of a match. In the ELO system, it is assumed that $u,v$ are scalars. But if you look at the math, there is no barrier from generalizing to vectors. For instance, if you assume $u,v\in\mathbb{R}^5$ (which is your case) then just replace scalar multiplication with the dot product and other similar changes to obtain the analogue for vector rankings.
Actually, performing vector rankings has been studied extensively and there is really nothing new over the scalar case.