I'm not sure if this is what you're looking for, but I project my guys similarly to J, although I tend to average my projections with others in order to eliminate my own bias to a certain extent.
When it comes to ranking players, I calculate the mean and standard deviation value of every category, then determine the number of standard deviations from the mean (above or below) in each category. I add up total sd's and that's a player's score. The higher the score, the higher the rank.
I like this way because it allows me to compare not only 2 hitters or 2 pitchers, but one of each.
I know it's nerdy, but it really gives a clear picture of value in 5x5 leagues.