Here's my first attempt at determining the relative value of each first round pick of the MLB draft. Using Erik's data from the 1990 through 1999 drafts, I totaled the WAR over the first six years of a player's MLB career (his years guaranteed to be under team control) for each draft slot.
I also applied two adjustments. One, I modified each pick's value based on whether it was a hitter or pitcher, giving more credit to pitchers and less to hitters based on Erik's research as to their expected MLB WAR contributions. Why? I wanted to create the ideal value chart, not one based on the mis-valuing of hitters and pitchers from the nineties. Two, I actually used a weighted average for each pick number, counting the three picks on either side of it to varying degrees in order to smooth out random fluctuations from pick to pick. To be honest, I looked at about ten versions of this graph with and without each adjustment and with some other trial adjustments and none looked significantly different -- there were just varying degrees of prettiness. There appear to be unconquerable limitations given just ten data points per slot.
Anyway, the graph is pretty interesting. The really high draft picks are worth a ton, gradually dropping off in value through number seven. At that point, however, value heads back up before decreasing more or less linearly through the end of the first round. Why? I don't know. Might there be issues with signability? Teams picking earlier are often poorer (because poorer teams don't win as many games and teams who win fewer games pick earlier) and poorer teams are more likely to let high-bonus picks slide to the teams willing to spend more on signing bonuses? Might there be issues with certain teams picking in the 5th to 9th range consistently and not being able to identify good talent (which is both a cause and effect)? Might it just be random fluctuation?
I'd eventually like to include draft data up through 2004ish, but is there anything else that should be considered in this type of analysis? If the trend at the tail end of the first round continues through the second and later rounds, there's really no point to considering them. Those later picks, on average, provide insignificant value as compared to early- and mid-first round picks.
Also, total WAR shouldn't be the only goal for drafting player -- money is also a consideration. Take the NFL for example. The first overall pick is actually less helpful for winning than the 32nd pick, because of the monetary requirements. MLB signing bonuses are nowhere near as large as NFL bonuses (Strasburg aside), but the expected return on MLB draft picks is much lower, too, perhaps creating the same problem as the NFL, just on a smaller scale. Another issue is MLB salary. In arbitration, successful players are likely to make serious money, lowering their net value as compared to failed picks. I'm not sure how to account for the signing bonus and arbitration money issues, but it should be done.