The difficult part would be actually tracking who's point boosting and who's simply doing things fairly, and then the additional issue of implementation. It's actually very similar to the oft-requested demand of, "get rid of teams".
Right now, not everyone whom I've talked to in the staff even agrees what constitutes point-boosting. We agree on a generic definition of, "fallaciously inflating the number of wins by playing rigged games," but we have not agreed on how a game would be considered rigged. Does it depend on the number of wins the person gains, or the number of people in the game? What if not every game is has the same winner, is it still considered a rigged game? What if it was simply a streak of good wins? And how do we differentiate between rigging a game with friends and simple teaming?
On a technical side of things, the above issues are even further complicated. We can interpret, infer, and imply a lot of information by our social understanding and historical extrapolation of previous games. Computers, for all their computational power, cannot be easily made to understand such things. We cannot make our servers flag people for malicious stats activity because there is no distinction between "malicious" and "innocent" along the lines of code; we could code in certain conditions that the system checks for, and once those conditions are met we could make a flag appear, but how can we make a computer do it if we humans can't even agree on it ourselves? It's technical implementation issues like that that make such things problematic. I would prefer if the Devs kept their focus on their current projects than deal with this rather-rare and innocuous issue.
The way I personally see it, stats-boosting isn't really practical or effective. It requires a lot of set-up, is subject to a lot of variety, and would require a lot of time. It requires a large-enough group of associates who are coordinated enough to dominate a certain server. Even if the 12-slot server was half-full with associates, that means that six other free players could still kill the others to ruin their scheme. And even assuming that this does occur with 100% success rate, they are still subjected to the standard wait times and persistence required to rank into the top 500. Considering that this is a long, drawn-out, boring, uneventful, and immoral process involving 3+ people, I don't see anyone taking it as a long-term viable option for internet fame.