I thought I'd look at all Top 5 picks over the 2006-2010 period to assess trends and risks. It is not a huge sample, so if someone has the time and inclination to look at years further past it would be a great idea.
The 25 players in my sample broke down as follows: 9 college pitchers, 7 college hitters, 2 high school pitchers, 7 high school hitters. That in and of itself is interesting. It shows that teams do not shy away from college players when making very early picks. If there is a group that tends to be underrepresented within very early picks, it is high school pitchers.
Now lets look at the individual groups to see if we can say something about risk. I'll list the players by group.
College pitchers: Hochevar, Reynolds, Lincoln, Morrow, Price, Moskos, Matusz, Strasburg, Pomeranz
College hitters: Longoria, Wieters, Alvarez, Posey, Ackley, Tony Sanchez, Christian Colon
HS pitchers: Hobgood, Taillon
HS hitters: Moustakas, Vitters, Tim Beckham, Hosmer, Tate, Harper, Machado
Among college pitchers, I'd say five out of nine have been busts (Lincoln, Moskos, Reynolds, Pomeranz and Matusz)
Among college hitters, I would classify four out of seven as busts (Colon, Sanchez, Alvarez and Ackley).
Among HS pitchers, Hobgood is a bust, but Taillon is still regarded as a very good prospect even after TJ surgery.
Among HS hitters, three out of seven are busts (Beckham, Tate and Vitters).
So with just about every group, about half the picks are busts. Given the small sample size, that's about the most reasonable conclusion that can be reached. There is risk with every group when it comes to top 5 picks, and the risks do not seem to vary much by group.
In terms of upside, there have been stud college pitchers (Price, Strasburg), stud college hitters (Longoria, Posey) and stud HS hitters (Harper, Machado) in the sample. So both the upside and downside risk seem to be fairly evenly distributed.