I think there are a few flaws w your system in that draft position doesn't accurately tell how valuable a guy is due to the protection rounds skewing people's draft positions. This method would work well in GDL where drafting position is a more accurate gauge.
According to your numbers, trading Chevy for Ron Francis is a decent trade (only 3.4 difference) and I'm pretty sure anyone in the league would disagree with that notion. Also, based on half of trades it thought were bad involving Luc, I gotta feeling that your system must overrate him, or perhaps a large group of guys who fall in that area of the draft where you probably wouldn't protect if r1 is protection, but you would if r2 is protection. It also doesn't adjust for how important different positions are (a starting goalie is obviously more important than 3rd winger). It is fun to play with though.