I think the best strategy in an autodraft is to go against the grain. Find out what type of player is popular and then find the bargains among the type which is not popular. Some of the types you might have to go with are non * starters, injury-prone players (some good values here), poor fielders at key positions (more fairly priced in '06), platoon players, or lefty-killers. All of these types of players can result in a winning team, depending on what type of league you are in. I believe I could piece together a decent staff of non * starters for US Cell from the leftovers of any autodraft. The staff durantjerry has above would suffice in US Cell.
I haven't tried a lefty-killer team this year, partly because we used to see three or four Sheas in all the autodrafts in previous years when Shea was very lop-sided. But I've had success in the past putting together teams that would play .500 against RHPs, but win 7 of ten against LHPs. If you get in a league with 40% of the starters left-handed, you can win a league with guys like Francoer and Damian Easley. These are guys who are almost always available.
The one thing you need is experience. You can't know who's popular without playing a few leagues.