I'll defend uniftw a bit in that I really do respect Koch's offensive skills. And a big guy who shoots 57%/76% fg/fts is giving good production. I do think the turnover to assist numbers are pretty bad and along with his foul troubles keeping him to 20 mpg are both things that need to improve if he's going to be a true impact guy.
But here's the thing and not just about Koch; it's one of my issues with sabermetrics and stat projections. Saying a guy's stats who averages 11ppg project to 20 ppg if he played x minutes, isn't the same as the guy who is doing it. They're fictitious numbers that you're making too many assumptions about. I have no problem saying you got really good production for the minutes he played, but making the assumption a guy who averaged x ppg or rbs in x minutes would average x more if he played whatever additional minutes and is therefore on the level with a guy who actually produced real stats is folly. Real stats win games , projections don't. You can make stats say whatever you want, but sometimes they make you think things that common sense and your vision tell you are not accurate. I know it's the trend in sports, and I do think there is value in looking at things from different angles, but projected stats just aren't the same as real statistics.