What exactly is a top prospect?

With all the talk about whether the Mets should trade top prospects for Johan Santana, and with names like Cameron Maybin, Andrew Miller, Phil Hughes, Jon Lester, Jacoby Ellsbury, and Lastings Milledge in the news this offseason, it's worth thinking about how good we can expect top prospects to become. When prospects are mentioned, opinions vary from "we can't let this future hall-of-famer go" to "prospects are worthless until they actually produce." So, what's a prospect really worth? How likely is it that a top prospect actually becomes a top player? Here's my extremely unscientific study to try to quickly analyze the question:

Methodology: Baseball America is the preeminent publication when it comes to baseball prospects. Every year, they publish a list of the top 100 prospects in baseball. For this "study," I looked at the top 20 from every year of the 1990s to see how they turned out. Top 20 seemed like a pretty good compromise between sample size and not including mediocre prospects. 1990 is the oldest list on their website and including more recent lists would've resulted in players for whom it's a little too soon to say how their career turned out. For each of the 146 players in the sample (the reason there aren't 200 is that some players appeared in the top 20 multiple times), I rated their career by giving them a number between 1 and 4:
1 = Dud - never made it to the majors or played very briefly (i.e. Brien Taylor, Bill Pulsipher, Mark Lewis)
2 = Average - player was a regular in the majors for at least a couple years, but wasn't really above average (i.e. Wil Cordero, Ben McDonald, Arthur Rhodes)
3 = Good - either played very well for a brief period or was consistently above average for a decent number of years (i.e. Robin Ventura, John Olerud, Jason Schmidt)
4 = Star - a great player - not necessarily a Hall-of-Famer, but at least somewhat close (i.e. Bernie Williams, Mike Mussina at the low end)

I realize this seems extremely subjective, but it was a quick easy way for me to do it, and I don't think many of my rankings would cause much disagreement. All the "duds" are pretty clear cut, and I don't think any reasonable person would disagree with more than a few players in any of the categories.
Here's the breakdown of the 146 "top 20 prospects":
Duds - 40 (27%)
Regulars - 56 (38%)
Good Players - 30 (21%)
Stars - 20 (14%)

So, a little more than 1 in 4 of these prospects were total busts. More than 1 in 3 became above average players and more than 1 in 8 became stars. So, those prospects turned out pretty well.

It becomes more interesting if you separate the pitchers from the hitters. There are 56 pitchers and 90 hitters in the sample. Here are the percentage outcomes for each group:
Pitchers: 45% Duds, 38% Regulars, 11% Good Players, 7% Stars
Hitters: 17% Duds, 39% Regulars, 27% Good Players, 18% Stars

So, for pitchers, there's a 45% chance that the prospect (who was once considered on of the top 20 prospects in all of baseball) will be a total bust and only an 18% chance that the player will be at least "Good." For hitters, those percentages are reversed. Those trends are even more apparent if I just look at players ranked in BA's top 10. Among top 10 prospects, 41% of pitchers turn out to be duds, but only 10% of hitters; only 11% of pitchers make it to either "good" or "stars," while 53% of hitters do.
That's a pretty remarkable difference. I'm not going to try to elaborate too much on what this means, but I think this once again makes it clear that it's a lot easier to predict how hitters will develop than it is for pitchers. Beyond that, I don't think this is enough to draw any other major conclusions about how much a prospect is "worth".

Here's another interesting way to look at the relationship between top prospects and top players: in reverse. I decided to look at the top hitters and pitchers in baseball today and then check how they were regarded by BA when they were prospects. I chose 25 hitters and 20 pitchers, somewhat subjectively. I glanced at the VORP leaders for the last 2 years, but I don't have any specific numerical argument for why these guys make the cut and other players don't. I didn't include any Japanese players because they didn't come up through the minor leagues and I didn't include any players who are too old (Bonds, Smoltz) to have been ranked as prospects in the 90s.
14 of the 25 hitters had been in BA's top 20, 19 of the 25 were in the top 50, 23 in the top 100. The only "elite" hitters who never appeared on BA's lists are Matt Holliday and Jorge Posada.
For pitchers, there was much less of a connection between top prospect status and future stardom. 7 of the 20 pitchers had been in the top 20, 10 in the top 50, 12 in the top 100. 8 of the 20 pitchers were never ranked by BA: Johan Santana, Brandon Webb, John Lackey, Tim Hudson, Dan Haren, Aaron Harang, Mark Buehrle, and Chien-Ming Wang.