David, you’re basically stating “This is how I use the site, so I want it to reflect that interest instead of attempting to service a broad range of players.”
Now, I think you have a fundamental misunderstanding of “overall rank”. It is not an average. It is calculated as an individual time-setting rank, effectively. So, it’s accurate whether or not you play one time setting or all time settings, your overall rank is calculated against your opponent’s overall rank. It should pretty closely match the time-specific ranks. Sometimes it doesn’t, but that’s something the devs are working on, as HowToPlay mentioned.
Here’s an example. You play only live games, and you play someone in a live game who has never played one before; they’ve always played correspondence. You’ll see each other’s overall ranks, let’s say 10k, and also your time-based ranks. Let’s say yours is also 10k, but your opponent’s is still 25k. Ideally, this should have been corrected to be close to their overall rank, but that’s not the point. If you win, your overall rank will have a victory over a 10k opponent factored in. Your live rank, however, will only have a victory over a 25k opponent factored in, meaning it will rise slower than accurate. Likewise, a loss would affect your live ranking far more severely than your oveall ranking.
This is the compromise the devs have gone with, and seems to do a good job, all things considered. Sure, you’ll get some odd games every once in a while, but ranking is a hard problem, and one that’s still being worked on.
Regardless, “wrong” and “wrong for me” are different things.