Refining WARP and OFP Pricing

Throughout the Brewers rebuilding effort, a difficult question occupied my mind: how does one assess a successful rebuild while it is occurring? There is obviously plenty of room for post hoc analysis, and perhaps room to create expectations about a rebuild (for example, one could say, “the Brewers rebuild is successful if they win 450 games between 2018 and 2022,” or “the Brewers rebuild is successful if they reach the playoffs once,” etc.). What is much more difficult, and more abstract, is to keep a tally on the development of talent over time, and the value of accumulating talent (as well as the effectiveness in cashing that talent into MLB wins either through trade or successful development). There are countless methodological issues with pricing this type of concern, and attempting to answer the question, not the least of which include the following biases (biases which need not even be consistent, but pull in several different directions):

  • MLB production by known assets will frequently be valued more so than unknown production by future assets (even where numerous scouting accounts and talent assessments are available).
  • Trades involving MLB players and prospects will be difficult to assess because immediate production at the MLB level requires a different scale for assessment than numerous contract reserve years for a prospect.
  • The existence of six-to-seven years of contract reserve rights for prospects will tempt analysts to value length of control over quality of production (ex., the Brewers traded Martin Maldonado, a superior player, for Jett Bandy, a player with a potentially similar profile but more risk as well as more “control years”).
  • Analysts, fanbases, and probably even teams will be biased in favor of the value of the players that they currently reserve within their system and undervalue the players that are outside of their system.

Throughout the 2016 season and 2016-2017 offseason, I worked on a Benefit-Cost Analysis style pricing system that sought to place MLB Wins Above Replacement Player (WARP), prospect/role Overall Future Potential (OFP), and cash ($$$) on an interactive scale. The culmination in this system was my “Translating OFP” article, which hypothesized that the potential value of a prospect could be calculated by categorizing historical baseball performances. By taking the broadest picture possible (indeed, calculating the WAR/WARP surplus for 18,000 batters and pitchers), I ascribed monetary value to specific prospect classes. Since these classes are interactive, they can effectively be used to test various assumptions (for example, if one wishes to calculate a player’s full range of 40 OFP-floor-to-60-OFP-ceiling, like the case of Lucas Erceg, one can average that full range of values; or, one can simply focus on highest possible ceiling, likely middle ground, etc., for prospects).

The point is not that there is one way to assess prospects, nor that player value is solely transactional. Instead, the point is that since MLB teams do in fact engage in rebuilding or win-now strategies at varying points in their franchise tenures, and do in fact trade minor league players for MLB players (and vice versa), these player types are indeed interactive even where their value is placed on entirely different scales. Taking inspiration from Benefit-Cost Analysis is important, then, as one can understand that the economic value of logging can indeed be assessed against the value of endangered species that reside within a particular logging habitat, or that the cost of a highway can be assessed in terms of fiscal impact, shipping improvements, commute time improvements, safety for drivers (expressed reducing preventable deaths), etc. Once again, these types of metrics are not absolute, which is why reasonable Benefit-Cost Analysis procedures include post hoc assessments to review assumptions and actual asset performance (compared to expectations). The same can be done for MLB players and minor league prospects insofar as front offices do indeed transact for such players.

Series & Revisions:
Translating OFP
OFP and Minor League Pay
Revisiting the CC Sabathia Trade
Cashing Out OFP 2
Organizational Logic and Playoff Trades
Historical WARP and OFP

In a forthcoming post, I will present an update on the 2013 Baseball Prospectus Organizational Top 10 lists, which I began tracking this season in the hopes of designing a realistic time series assessment of prospect value. The ideal in developing this approach is to answer the question, “How do prospect classes progress over time?” A related question can then be answered: “How do prospects of a particular OFP perform over time?” By the time 2017 started, clear delineations between 50 OFP (average), 60 OFP (first division), and 70 OFP (superstar!) prospects were already forming at the MLB level from the 2013 Organizational Top 10 class. These classifications further solidified during the 2017 season, and the benefit is that there are now five full seasons from the publication of this list through which player value can be assessed.

From the 2013 prospect class, I constructed the following table comparing each prospect category to the average MLB performance from that class, and used those figures to determine a monetary “price” for those prospects and MLB production. The difference between each OFP category and the average production from the full prospect class can create a WARP figure that can then be “monetized,” thereby translating OFP into a WARP category. “Risk” (based on each OFP category’s frequency of reaching the MLB) can also adjust these WARP and monetary figures. In this case, however, OFP production value is concrete instead of speculative; this is what certain OFP classes actually did at the MLB level (the same goes for these risk classifications, which are also “actual”):

2013 Prospect Org Top 10 MLB AvgWARP AvgValue 70Context 70Value 60Context 60Value 50Context 50Value
2011 3 0.2 $1.2 -0.1 $0.7 0.2 $2.8 -0.2 $0.0
2012 35 0.0 $0.3 -0.1 -$0.7 0.1 $1.0 0.0 $0.1
2013 87 0.6 $4.3 0.6 $8.5 0.1 $5.0 -0.2 $2.9
2014 123 0.8 $5.5 0.2 $7.2 0.3 $7.7 -0.3 $3.5
2015 157 1.1 $7.4 0.9 $13.6 0.0 $7.1 -0.2 $6.0
2016 178 1.3 $8.9 1.1 $16.5 -0.2 $7.6 -0.1 $7.9
2017 173 1.2 $8.7 0.7 $13.3 0.0 $8.5 -0.2 $7.3
Summary 756 0.7 $5.2 3.3 $82.1 0.5 $43.3 -1.2 $19.3
Risk 77.9% 0.6 $4.02 96.6% $45.7 77.2% $25.0 74.7% $18.1

Assessing 756 MLB seasons, played from 2011-2017 (with most prospects playing between 2014-2017) by the 2013 BP Organizational Top 10, allows for a broad range of data to be analyzed. This produced time series for each player that can be analyzed, as well as annual summaries and prospect class summaries. From this table, one can compare initial historical OFP estimates with risk-oriented “bare minimum” prices and “top ceiling” prices:

Prospect Class Historical Model (Risk) Historical Model (Ceiling) 2013 Prospect Model (Risk) 2013 Prospect Model (Ceiling)
50 OFP $7.0M (40-50) $19.5M $18.1M $19.3M
60 OFP $20.8M (40-60) $48.9M $25.0M $43.3M
70 OFP $45.8M (50-75) $100.0M $45.7M $82.1M

The same criticisms of using WARP and OFP to assess players obviously apply: MLB players and minor league prospects are dynamic, and at different points of time their WARP or OFP may indeed mean different things; a player can add a new pitch or make a delivery adjustment and improve upon their previous performance (like Jimmy Nelson), or answer a specific set of approach and mechanics questions to achieve a higher grade (compare Josh Hader’s 2013 Top 10 appearance as a 50 OFP Orioles prospect with his 2017 Top 10 appearance as a near-surefire impact MLB player between 55-60 OFP depending on high-leverage relief or Number 3 starter risk assessments). For 2018, Brewers CF prospect Monte Harrison will be a great example of a prospect whose OFP role has evolved over time, due in part to injuries and also in part to on-field adjustments and advancements; Corey Ray will almost certainly demonstrate this fact in the opposite direction, due to mechanical question marks and growing pains.

So, indeed, WARP and OFP should not be viewed as fixed, or definitive metrics. They should be viewed as reasonable snapshots of a player’s value or potential at a given time, which MLB teams can indeed use in order to assess transaction value. A common complaint about my method is that MLB teams do not use WARP and OFP to assess players, but I find that this complaint misses the point that teams do boil players down to a “number” when they decide to trade, or keep, a player; their “number” might be predicated on a blend of proprietary scouting insights, years of mechanical data, and other performative metrics. That MLB teams do not use WARP and OFP to assess trades does not mean that fans and analysts should not use these metrics where other proprietary information is not available; for failing to capitalize on the interaction between Future Potential, Replacement Value, and cash means that the MLB analyst and fan missing one specific constellation with which transactions can be assessed.


Photo Credit: Steve Mitchell, USAToday Sports Images

Related Articles

Leave a comment