Macrosabermetrics: Evaluating MLB Team Resources to Produce Wins (1998-2013)

There are many models generated in the field of sabermetrics. From the Pythagorean Run expectation (exponent 1.83ish), OBP/SLG to get runs scored; to linear weights, wOBA, position weight, and FIP to find WAR. And more models are focused on the microsabermetrics: swing%, balls outside the zone, pitch f/x, field f/x, event sequencing, and so on. I tend to the macro side, because things are clear enough to see for the average fan.

For this study, I referenced and used the following sites:

  • Baseball Prospectus – player compensation; draft model for value Andrecheck
  • Baseball America – top 100 prospects by year
  • Baseball Reference – salary, WAR and trade study (nixed)
  • Beyond the Box Score – a study on $/WAR by year back to 1996
  • Cot’s Contracts – salary comparison
  • Economics Website ( – inflation index
  • Fangraphs – Rookie WAR by year and type from 1998 to 2013; WAR by teams.
  • Sean Lahman Database (2013 version) – salaries, wins, etc.
  • USA Today – salary comparison

As you can tell, I didn’t just reference one or two sites for salary. In fact, by my account, they all differ in one way, or another. Bonuses applied differently; incentive payouts unadjusted; deferred payments not considered in the salary basis. But for the most part, the differences were small in comparison to Lahman, maybe 500K to 2M per year, about 30% spot on compared to Cot’s. So instead of looking into 480 teams and 25-30 players and seeing where the problem lay, I went with GED (Good Enough Data). When its totaling to 1.5B or 3.1B(Yankees), what’s 10M among baseball billionaires?

Salaries Son

You can’t win in the regular season without money in baseball unless you are Billy Beane. That would be the first substantial conclusion to draw from this analysis. Oakland is the best run, most efficient organization over the last 16 seasons. And it is not really all that close. But let’s get to the meat of what was done.

  1. Compiled Salaries
  2. Inflation Adjustment – it resulted in nearly identical R-squared to basic salaries
  3. Rookie Salaries by Year set baseline for salary on a team
  4. Baseball America Top 100 prospects from 1990-2014 generates model for WAR similar to a draft study performed
  5. Fangraphs Rookie WAR relates to prospects success rate (assumption); though it would not be that hard to do the finite analysis
  6. Average WAR for period is set at $4.8M/WAR
  7. Minimal Salaries for having a team are calculated for the duration
  8. Wins for 1998-2013; WAR level at .294 times 162 games for 16 seasons.
  9. Calculate Win Expectations for further analysis
  10. Test a Bullpen theory
  11. Do Math on outlier teams
  12. Linear Analysis and Graphs
  13. Note: I compiled the salaries together because it shows closely the total investment in ballplayers over a span. If you buy FAs just for a season, to bump payroll (Miami), but sell them for prospects, at some point, the investment is basically the dollars shelled out and the prospects received. If they are rated well, and turn out for you, you gain value (thus the prospect value).

These are the charts:


As you can see rookies have been producing more as of late, particularly pitchers. Rookies are accounting for 8-12% since 2005, around the time Moneyball and sabermetrics started influencing FOs. Or as salaries rose, finding cheaper talent (relievers, for one) made sense. Lots to explore, but just one piece.


Next BATTER WAR as it relates to salary. Chicago, Washington, and Pittsburgh were woeful compared to what others did for the same price or less. Even Oakland, which will see pitches even better, did fairly well. Boston, Cleveland, Atlanta and St. Louis get bang for the bucks.


Pitching was the key to Oakland’s success over the last 16 seasons. Huge differential from the MLB trend line. San Diego and Milwaukee show their weakness here for the dollar spent. Minnesota is also hugely effective. As you will notice, wherever the Yankees are on the left, Boston is about 20-30% behind on cash, but “in the ballpark” on results.

Baseball America and Fangraphs information

Baseball America and Fangraphs information

Prospect equation graph for perusal again. I used this as a basis to formulate what the theoretical prospect WAR should be. As guys are on the list 2-3 seasons, it’s very crude. Yet, it’s a start to seeing how far it falls off.

Macrosaberfinal  this gives a summary of what is in the next graph.

I did not include everything in this analysis for show.  But the graphics and the excel file show a tendency that the best FO team management over the past 16 years were:


Best Teams

  1. Oakland A’s
  2. Atlanta Braves
  3. St. Louis Cardinals

While, the Worst Teams were:

  1. Baltimore Orioles
  2. Chicago Cubs
  3. Detroit Tigers
  4. Kansas City Royals
  5. Pittsburgh Pirates

Now, recently history (2011-2013) can’t undo bad franchise operations of the prior decade; neither can a few bad seasons undo the great run of the Yankees – who are decidedly average – and drove the outlier too. (More money always helps.)

Oakland developed their rookies compared to their prospect rankings at a 16% clip. The Cubs, at 3.77%. By prospect rankings done by Baseball America, the Cubs had greater potential in their minors, yet, their rookies failed at nearly a 4x clip to the A’s. (Again: this is not a perfect analysis – rookie players tend to be highly thought of prospects, else they would not receive the MLB promotion…not all, but enough to draw a link.) This shows why the Cubs are hard at the minor prospecting and development academy.  Tampa Bay and St. Louis did fairly well; though Tampa had a streak of a highly-rated system.

Yet, based on the analysis, prospects development rate showed no correlation to winning. The adjusted R-squared was minutely negative. In short, it might not give all the advantages some say it does.

On the final graphic, if you REMOVE Oakland from the analysis, the R-squared jumps to 68.9%. Meaning more than 7% of the variation is due to Oakland’s performance. I’d say that suggests just how different they are from the rest.

This entry was posted in baseball america, Chicago Cubs, mlb salaries, New York Yankees, Oakland A's, prospects, St. Louis Cardinals, WAR analysis and tagged , , , , , , , , , , , , . Bookmark the permalink.

3 Responses to Macrosabermetrics: Evaluating MLB Team Resources to Produce Wins (1998-2013)

  1. Brian Cartwright says:

    Not a lot of explanatory text beyond the bullet points. It is implied in 3 and 7 that you are using marginal $ (salary above 25 minimum salaried players) but not stated explicitly. Is the last graph team marginal wins per marginal salary? You also need to remember there are three types of salary – pre-arb (slave), arb (discounted) and free agent (full market). Given the weighted mean of the three types of contracts, did each team spend more or less than average per marginal win? and for individual players, one must use only positive war, as in <= 0 WAR = min salary and positive WAR = salary above min.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s