I always thought a good way to go about rankings would be to assign a percentage of the league total for each category to each player. I'm just generating random numbers, but say if David Ortiz stole 100 bases and your entire league totaled 1000 stolen bases, Oritz would score 0.10 in SB. You could then add the totals together to get a players score.
Two problems instantly come to mind with that method, however.
One is how on Earth would you work averages like AVG and ERA. You can't devide them against league average, or have them be on the same scale as the other categories (every player's total added together wouldn't equal 1).
The second, the rankings wouldn't reflect which categories your team was already strong in. You could keep on drafting power hitters that scored well in both RBI and HR but hurt you in AVG and SB because their overall totals combine each category into one generic number.