porkenbeans wrote:If that was done, then couldnt an N.F. get to the top and then change his ways and play 100 games against his peers ?
Yes it does mean that. I don't agree with putting a label on someone and excluding them forever. I just wanna know who is the best.
I know for a fact that there are many very skilled players engaged in farming. They choose to farm because it is a viable method under the current scoring system. They are practical and ambitious and many of them, I'm sure, are capable of beating high ranked EQ's should they square off.
Some of these players are bored of farming and want the challenge of playing peers - they are simply waiting for a system to reward that behavior. By computing RR across all games, these players might have to play thousands of games with their peers to get their mean RR above any cutoff simply because they have been farming to some extent for a really long time.
People talk about "true rank". If you are a colonel you will eventually make it to colonel and probably not make it to brigadier, or not make it stick. The concept here is that if you are really better than your peers, than your score will increase to a certain extent after playing them. Eventually you'll reach an equilibrium, the "true rank".
Theoretically, you can achieve your "true rank" from no matter what starting point. So it doesn't matter that a guy who got to 4000 points through farming plays 100 opponents or a guy who got to 4000 points through playing peers plays 100 opponents. Their scores after playing the 100 peer opponents will readjust to the "true rank" relative to their peers. It is possible that many of these players will not even experience a score loss. These guys deserve recognition if they can play peers and earn it.
We just need to make sure that the number we use for the rolling average is great enough to ensure a thorough "trimming" in the case of a noob farmer that really sucks against their peers. Otherwise a cycle of farming -> peer playing could be used by players to make it onto the competitive scoreboard for various lengths of time.
porkenbeans wrote:But still the advice is correct about just playing some high ranks to raise your RR. right ?
Yes.
But what I'm saying is, requiring someone to play 5000 peer games (for some people) is, for the reasons above, not necessary to ensure that an "inflated" farmer gets amply "deflated". I think we are looking more at something on the order of 100-1000 opponents.