So if we can use this database as an indicator of someone skill level(Kels proves this). then what's the point of it?
I'm assuming that's supposed to say can't. And to answer the question:
1) The DB was down for two months, has NO results since 4/7, and something is wrong with the site and will not accepted new tourney uploads.
2) The DB is a work in progress, and once enough results are compiled, you WILL be able to use it as an indicator of skill level.
3) Like ALL indicators of skill level, it will never be 100% perfect. It's like trying to decide who is the better player: Jman or Lovage. S2J or Hax. random noob 1 or random noob 2. There is absolutely NO POSSIBLE WAY TO EVER KNOW 100% OF THE TIME. There are infinite variables that no system could possibly compute. So don't be a ****** and take it with a grain of salt.
4) Kels's elo is not overrated. There are a lot of players that are lower than him that are underrated. With isolated regions like the Midwest, or some areas overseas, it's impossible to determine relative skill levels without competition between other regions. Elo can correct itself rather easy, however when those regions mingle. But honestly, if everything was updated prior to Rule 6, you would see Kels rated around 1850 lose to Mango rated around 2050 and S2J rated around 1850, and he beat everybody else he played. So his Elo score is actually pretty accurate. You're just looking at the fact that he's the 8th highest rated player. That's more of an issue with missing a lot of recent results, and missing a ton of data for some very strong regions like Florida, East Coast, Canada, and Mexico. If we ever get enough data, Kels will fall to the more likely range of 30-50 (in the world) where he should be.