Trading Cole Hamels: Pt 2 – Valuing Prospects

To read part 1 on Surplus Value go here

Based on the set up of this site and everything I have written before, you can tell I like prospects, and at times I have immersed myself too much into prospects.  However, the access to minor league information and the number of people and sites covering it has led us to have a very twisted view of the world of prospects.  We now have entered a world where we simultaneously over and under value prospects based entirely on hype that we generate and demand.  We expect great things immediately from prospects because we have been spoiled by recent successes, and we call players busts well before they have had time to adjust to the majors.

Adjusting Baselines:

The following is a graphic from FiveThirtyEight of the average bWAR of players in the seven years following them being ranked on a Baseball America Top 100 list.

As you can see we have some real impact at the top, but even the #1 prospect in baseball averages only 20 WAR over 7 years which is good for above average regular at just about 2.8 WAR a year (in theory this would be an increasing amount over the course of the 7 years).  But once you get to about the 20th prospect on the list you are at a peak of 10 WAR over 7 years, which is not going to make anyone happy with their prospect’s outcome.  A similar study was done back in 2012 by Kevin Creagh at Pirates Prospects, he looked at BA’s Top 100 from 1994 to 2003 and he broke it out by Hitters and Pitchers.  I have condensed his results based on fWAR here.

When these metrics are presented the numbers quoted are the average values.  Over a large sample size of prospects we see the overall value come to these averages, but each prospect is unique.  Some prospect’s skill sets lend themselves to a safe floor due to positional flexibility or current arsenal, others like Texas 3B Joey Gallo have huge swings in outcomes (often around either the hit tool or pitcher control) where they are likely to be either at the top or bottom of the spectrum.  This is important when looking at a singular prospect, the range is very large, it is best to move towards the center but relying on a straight average simplifies the factors.  For example here are the Baseball America Top 10 prospects from 2008-2012:

We can see that even in this small sample size at the top we have huge successes and real failures.  There are some players who hit the middle but have still been considered disappointments.  The larger point being, there is no easy safe player to take and bank the average.  Teams want the peak, but sometimes you get Jesus Montero instead, and get the real downside.  It is important we don’t walk in absolutes.

Simultaneously Over and Undervaluing Prospects:

It doesn’t matter whether you look at the 538 or Pirates Prospects study the tiers of outcome become more apparent.  There are of course exceptions and outliers in both studies, but we can see there are significant drop offs once we move outside the top 10 and then very dramatic after 25.  This leads to undervaluing the top prospects if we pull them down into the generic bucket of Top 100 prospect.  If we look at the 2014 Baseball America Top 100 this would mean that Carlos Correa (#7) is worth nearly 4 times as much as Kolten Wong (#58).  As we talked about yesterday the diversification makes sense if you have the roster spots for all of the players, but if you are looking for an upgrade we see that Correa is the much better bet for a team.  The study does not exist right now because there is no good way of determining prospects 101 to 200, but given the trends we can see the gap from a back of the Top 100 prospect to a player off the list is not as large as we see from top to bottom of the list.

In this way we greatly overvalue the label of Top 100 prospect.  Yes the #90 prospect is better than the #190 prospect, but we often view #190 as not a prospect in our views on a farm system.  To correct this we need to understand the gap that exists in how that kind of prospect is valued.  It shows us that just adding players to a trade won’t bring you to the value of a single stud prospect, but that the difference in players on the lower end of the spectrum might be more interchangeable.

Surplus Value:

So back to the buzz words of “surplus value”, as we saw yesterday surplus value is both linear and non-linear.  Either way taking the $/fWAR calculations we can make a translation of the fWAR above to a surplus value figure based on estimated arbitration.  Here is that translation from the same Pirates Prospects’ analysis, and looping in a study by Victor Wang on Hardball Times in 2008.

As expected we see the same gradations in the values with real distinct steps down in the more recent analysis.  This set of surplus values is what we will use in our next part as we bring this into line with Hamels and how we expect him to be valued in a trade.  Before we go into that remember these are total surplus values over a 7 year period and are averages over a range.  At the bottom you are looking at savings close to $1 million a year which is not a ton of difference to major league club.  At the top you are seeing 6+M of value for a hitter, which is a sizeable amount.

The big difference I do want to look at is the hitter vs pitcher value scale.  The Pirates Prospects analysis used to fWAR which is based on FIP and in general has less movement than an ERA based WAR like bWAR.  Fangraphs has a 43-57 split on how they allocate WAR to hitters and pitchers.  In general pitchers are more volatile due to injury risks than hitters.  It is not shown as much here because just have high and low, but there is a bit more room for pitchers to carve out a career at the floor of WAR, especially in the bullpen where there are many debate on how WAR figures in.

Up next, what is Hamels’ value