Prospects, what are they good for? A Quick and Dirty Historical Analysis of WAR Generated by the BA Top 100 Prospects

via img801.imageshack.us (data: fangraphs.com)

The scatterplot above represents the average seasonal fWAR generated by ballplayers appearing in Baseball America's Top 100 list for the years 1995-2009 (i.e. a fifteen-year sample).  Six years' worth of data (begining with the year of the BA ranking) is included for each ranked player (where available--obviously, the 2006-2009 prospects have yet to generate six-years' worth of data), and all of that data is averaged together to create a single per-season fWAR value for each position in the rankings (i.e. 1 through 100).  Because I am interested in assessing the value of each ranking position, rather than the individual players, I have not eliminated duplicates.  As a result, Joe Mauer (who was the #1 prospect in both 2004 and 2005) is responsible for a disproportionate share of the value associated with the number-one spot.  He also did his share to help out the #4 spot (2003) and the #7 spot (2002). 

The overall pattern seems pretty clear, and I took the liberty of adding a trend line to emphasize the point.  The number-one spot has generated an average of 3.825 fWAR per season over the course of the last 15 years, and it tails off somewhat logarithmically after that, with the trend levelling out around 1 fWAR somewhere near the 60th ranked spot. 

Also, when looking at this data, you should keep in mind that WAR is a counting stat, so it is heavily influenced by playing time.  Because BA is ranking prospects, rather than established players, there are a disproportionately large number of partial seasons (including, for instance, September call-ups and other "cup of coffee" situations) in the data.  With this in mind, consider the following graph, which compares the average number of "Top 100" prospects who log time in any given year after being ranked, with the average seasonal fWAR value generated by the entire group in each year.

via img31.imageshack.us (data: fangraphs.com)

As you can see from the vertical bars on this graph, roughly half of the players appearing in the BA list will actually see some time in the Majors in the same year that the list is published (i.e. "Year 1" in this graph).  That fact surprised me a little bit, since BA does rank prospects in the low minors.  It seems clear that the BA lists have an overall bias toward guys who are pretty close to ML-ready.  You may also note from this graph that something like 20 guys on any given list are never going to see the Show, and that roughly one third will not be in the league long enough to reach free agency (i.e. six years). 

Turning to the WAR values at the top of the graph, you can see that prospects generate much less average per-season WAR in "Year 1" than in the succeeding years.  Some of that is undoubtedly due to the inevitable "adjustment period" associated with the jump to the Majors, but it is probably fair to say that the lion's share of that difference is attributable to playing time (i.e. many of the "seasons" included in that first-year average are actually just a handful of September games, etc.).  Because the number of ranked prospects in the League levels off so much between years 3 and 4, I suspect that the playing-time distortion falls out of the data around that point.  After that, the headcount starts to drop, presumably as the maginal guys wash out of the League, and I suspect the continued rise in seasonal fWAR can be attributed, mostly, to the "washouts" no longer dragging down the overall average.  As a result of this general line of reasoning, I suspect the year 4 average seasonal fWAR of 1.598 comes much closer to the "true value" of these ballplayers, on the whole, than the total value of 1.388 fWAR/season that I calculate from the complete dataset.

I am hoping to ultimately add to this dataset, so that it will eventually include all of the BA top-100 lists going back to 1990, but with this year's list already on the way, I thought someone might be interested in seeing some of the info I have gleaned so far.  Once I see the actually BA 2011 list, I'll use the numbers to put a value on the Royal's farm system based upon the number and position of our newly ranked prospects.  I'll probably compare that to some of the other well-regarded farm systems of the past decade or so.

In the meantime, I made a couple of pretty graphs.  I hope you like them.