Two years ago I devised a plan to better evaluate front offices besides the typical wins and losses used by many casual fans. I perform this evaluation by creating specific metrics using preseason projected statistics and then compare those derived metrics to the ones calculated with the final statistics at the end of the year.

My approach is driven by the simple model of the firm and the many economic principles that go along with that method. The economic fundamentals used in my research have been slightly adapted to fit the constraints of baseball. For example, instead of dollars, the WARP stat might be used to calculate a value. Also, certain economic equations break down when using them in a pure baseball sense, so alterations and in some cases new equations were needed to comply with these baseball controls. I will do my best to explain these adaptations as I go along.

In the following models I have chosen to use WARP in place of dollars. There are two main reasons I chose to use WARP. First, it can be universally applied to pitchers and position players much like a form of currency. Second, team WARP is nicely correlated with team winning percentage as you can see in the chart below.

Scatter tWAR~WPct

 

The first concept I attempted to apply in this front office evaluation project was that of economic profit. Comparing economic profit to accounting profit is the first step in understanding how this can be applied to baseball. Accounting profit refers to the actual revenue calculation of a firm. This is the hard dollars and cents. Economic profit is more theoretical and attempts to factor in opportunity costs. In my baseball model accounting profit is replaced by total WARP. Total WARP is the sum of WARP produced by each player during a season. Economic WARP (eWARP) takes the place of economic profit. Instead of looking at team WARP to evaluate a club’s performance, we can look at eWARP, which accounts for the lost opportunity cost of players traded, released, or in some way removed from the team prior to the current season.

The formula for eWARP:

eWARP = Total Team WARP – WARP of Players Exported from the Team

Last year I used Steamer for my preseason front office evaluations. Because I am now writing for BP Bronx, this year I will be using PECOTA preseason projections. When looking at the above formula, team WARP should be self-explanatory. Exported players are those players who were traded or let go by their respective team. This is where the lost opportunity cost comes into play. For example, the Braves traded Shelby Miller in the offseason. PECOTA projects that Miller will be a 1.6 win player in 2016. Based on how I am calculating economic WARP, the Braves could have a lost opportunity cost of 1.6 wins by letting him go. Obviously, the Braves received some players in return for trading Miller. This will be taken into account shortly when we look at gross domestic product and how that can be applied to evaluate front offices. For now, take a look at the chart below that shows you the general managers who did the best with regards to the measure of economic WARP.

2016_Pecota_eWARP

The next concept I used to evaluate general manager performance is gross domestic product. Once again I will use wins as the form of currency which allows me to shift the name of this metric to gross domestic wins (GDW). GDW allows us to isolate the transactions made by general managers even more so than eWARP. It allows us to compare the players a club acquires versus the ones they let go in the context of WARP.

I had to alter the GDP formula to fit the model of a major-league baseball organization. GDW can be determined by adding the sum of the wins above replacement for the team’s call-ups and the wins above replacement of net imports.

The formula for GDW:

GDW = WARPcallups + (WARPimports – WARPexports)

The “call-ups” in the above formula could be rookies or guys who may have spent some time in the big leagues, but had more recently been in the minors. “Imports” are players acquired by the team through trade or free agency (domestic or international) and “exports,” as explained before can be players released, traded, or in some way removed from the team prior to the beginning of the current season.

Take a look at the chart below showing each organizations preseason GDW value and their rank in this category. As before, WARP from the 2016 PECOTA preseason projections were used to calculate gross domestic wins.

2016_Pecota_GDW

This year I am adding an additional statistic that I had previously not covered. It falls along the same line as GDW, but uses mean values for WARP of call-ups, imports, and exports. By using the average of these stats we can better understand the performance or – in the case of the preseason evaluations – projected performance of each pool of players. This eliminates the bias of one pool containing more players than the others. For example and a complete hypothetical, let’s say the Pirates may have removed or exported 15 players who are projected to receive big league playing time this season with their new teams. However, they may have only acquired or imported three players expected to play for the major-league club. This difference in the number of imported and exported players could skew the figures so that it looks like they let go of more value than they took in. Mean stats eliminates this potential counting problem.

The formula for AvgGDW:

AvgGDW = Average WARPcallups + (Average WARPimports – Average WARPexports)

2016_Pecota_Avg_GDW

This last chart is the one you should reference as the final result of evaluating general managers and their front offices for the 2016 preseason. In my opinion, it is the most accurate of the three I listed. It must be stated that evaluating the general managers’ decision based on preseason projections is not without its flaws. By doing this I am assuming that teams are themselves evaluating players using similar methods to PECOTA. Also, there could be other technical constraints skewing a team’s numbers. For example, an organization may contain several underperforming players with large contracts. This may not allow them to easily move the player thus making it difficult to improve their eWARP or GDW. It should be noted that minor league players who have been traded or released who do not have PECOTA projections are not considered. I am still trying to find a feasible way of accounting for future value of minor league players in these transactions.

For a more in-depth analysis of where the Yankees – and possibly a few other teams – fall on this list and why, you can watch for my next post at the Baseball Prospectus Locals site BP Bronx.

Stephen writes about Major League Baseball at BP Bronx and Banished To The Pen. He also informs readers about college baseball at the blog Underground Baseball. Follow Stephen on Twitter at @steve21shaw

Next post:
Previous post:

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.