Personalized rating scales fail when you base others on your own approximate skill level which is why I try to avoid rating people if they happen to ask me what I think. You really need to base it on a lot of different players whom your see play. Most players rank themselves a little bit higher than they actually are, and in general, giving an online ranking is the same thing as creating an online tier list. Neither of them is necessary because they are largely based on skewed results.
For instance, I've seen random players generally rate me around a 9 when I know they've only played me when I've used the gamepad a while back. Personally, I'd put myself about a 6.5, maybe a 7 on pad. That leads me to believe they think they are a bit better than they really are based on their predeterminations of skill of the person they are against, and do not take into account handicaps. "Wow I'm doing pretty good against Shock and he's obviously about a 9, that must mean I'm better than I thought." Netplay not withstanding, they obviously don't know that I'm considerably better on a stick than a pad, and even better on the keyboard than on a stick, so would they rank me higher after playing me on a stick, or rank themselves lower?
I've seen people rank other players whom I would put at like, a 6, as high as an 8 and change and to me that is obvious evidence that the person ranking players too high hasn't played many people better than that person, or not even many, but ANYONE better than that.
I'm curious to see how some of the newer players I haven't met are doing, so anyone who wants to get in some games who hasn't played me yet, feel free to IM me on AIM and we can set something up, I'm generally on after 10PM but that doesn't always mean I'm available but I can try to get in a few games.