I didn't mean to imply DPs are not important. Just that I'm not as convinced that the performance history is quite as useful as a predictive tool. I once entered two identical rosters into seperate open leagues (a roster full of 1880's A+++ range guys). I saved the league spreadsheet. While their errors were almost exactly the same (most players differed by 1 or 2), their turned DPs were vastly different. McPhee (2b) and Glasscock (ss) turned 71 and 82 DPs, respectively in one league while the same two players turned 87 and 99 DPs, respectively. That's a 33-DP difference. Again, the entire 25-man rosters were exactly the same, but the opposition was different.
That being said, if the adjusted AVGs are close, then I'll almost always lean toward the better range guys, figuring the extra DPs will be the difference.