Was thinking about this on my way to work, using round numbers and assumptions... For every good player (let's say with 2.0), there may be 4 less than average players (let's say 0.5). To average that K/D would a total of 4/5 which equals 0.8. Works for me.