Wade Davis and THE TRADE

I have been considering building a capital budgeting framework for valuing baseball trades, and Wade Davis’ start this week pushed me to starting.  Davis’ potential is actually intriguing to me, so I wanted to find a way to see how good he needs to be for me to be okay with giving up Wil Myers.  Capital budgeting is a concept from corporate finance that focuses on modeling cash flows and discounting them to present value for project selection.  Using this idea on expected wins from the players in a trade could be a great way to see what team got the better end of a trade.  I will show you a simple model that I built in about three minutes for THE TRADE and then I will discuss the assumptions and what needs to be done to make this sort of thing work properly.  If you want to forego said discussion I understand, but input from others might help me flesh this idea out more and would be appreciated.

2013

2014

2015

2016

2017

2018

Shields

4

4

0

0

0

0

Davis

2

3

3

3

3

0

TWins

6

7

3

3

3

0

Dwins

6

6.4

2.5

2.3

2.0

0

Royals 2013 Win Equivalent

19.1

Myers

1

3

4

4

4

4

Odorizzi

0.5

1.5

1.5

1.5

1.5

1.5

TWins

1.5

4.5

5.5

5.5

5.5

5.5

Dwins

1.5

4.1

4.5

4.1

3.8

3.4

Rays 2013 Win Equivalent

21.4

Discount Rate

10%

The main components to this are projecting the WAR values from each player for the contracts/player control time that was traded.  Then coming up with a discount rate, that would approximate the win inflation over the time period, to make sure wins this year are worth more than subsequent years.  The rest of it is just simple math.  As you can see I am not including Elliot Johnson, Mike Montgomery, or Patrick Leonardas I believe there value to the teams is either zero or close enough to zero that I can ignore

December 12, 2012; Kansas City, MO, USA; Kansas City Royals general manager Dayton Moore (left to right), newly acquired pitchers James Shields and Wade Davis, and manager Ned Yoast pose for photos after the press conference at Kauffman Stadium. Mandatory Credit: Denny Medley-USA TODAY Sports

them.  I put James Shields down for two seasons similar to last year at 4 WAR each, and then zero from then on out since his contract will end.  Then I projected Wil Myers and Jake Odorizzi and played with what Wade Davis would have to do to make the trade somewhat even.  You can quibble a lot with these projections, but whatever they are, Davis is going to have to be pretty good to make this trade valuable for the Royals.  TWins are a sum of all WAR for the year for that side of the trade (Shields + Davis or Myers + Odorizzi), and DWins use the discount rate to present value the wins in terms of now.  As I go on I will be discussing how to make a model like this better since this one is pretty simple.

First, the projections need to be better.  I am pretty comfortable with Shields, but the others are much harder to project.  My preference would be to crowd source these win values to get expectations from multiple fan bases both biased and unbiased, but that would require me having a significant amount more power in the blogosphere.  In lieu of that I will probably need to go look at scouting expectations and then use aging curves for the three players who are younger and going to be around for a while.  Once you have those it is just a matter of setting a discount rate.  A win this year is more valuable than a win next year, and though you could argue for specific times and specific teams this could change a lot, I think this is still a good assumption.  How to set the actual discount rate could go several ways though.

You could try and set a league wide discount rate using win inflation, meaning look at what team payrolls are per WAR year over year recently, and then project an inflation rate from there.  This could be a good way to set the rate, but it is not what I was thinking.  You could also do this on a team by team basis since the Tampa Bay way won’t allow it to grow payroll at the same rate as the Dodgers or Yankees.  Personally, I would prefer to set inflation rates separately for the players.  The model above assumes one discount rate, consistent with a league-wide win inflation rate, but I think each player’s risk profile might give better results.  That way we could use what we know about the players to set the rate.  Prospects are riskier, so their discount rate should be higher than established big leaguers (we are more confident in what James Shields will be than the other three so his discount rate should be lower).  Also, pitchers should probably have a higher discount rate than position players due to higher injury risk and year to year variation.  Once you have done all this, a good estimate of each trading team’s expected present value of wins should be formulated, and then you move on to money.

One team or the other (or for multi-team trades you could compare more) is going to have an edge in WAR expectancy, but that is not the only factor.  Payroll should be used as well.  In the case of this trade the Royals have a lower win expectancy and took on more payroll with Shields’ contract, which is why most stat heads gave the clear victory to the Rays on the trade.  Anyway, that is an introduction to the method I would like to develop.  If you like it or have any questions or modifications I should consider let me know.