Translating OFP

After seeing the scouting strengths and weaknesses, as well as the Overall Future Potential (OFP) and realistic floors, for the Brewers system, one can begin the New Year with a great sense of the “what if…?” that is the current state of the rebuilt farm system. There is no question that the Brewers front office has assembled, and now will try to develop, quite a deep array of talent. Yet, there is a serious question about how this talent will translate into the MLB: what is the likelihood of Brewers prospects reaching their ceilings? How many will surpass their respective ceilings? Are the floors good enough to built a contending team (or trade for players who can contend)?

In framing expectations, it is worth investigating the strengths and weaknesses of MLB players as summarized in their overall replacement value. For this exercise, I will use Baseball Reference Wins Above Replacement (WAR) instead of BaseballProspectus WARP, since I used the Play Index to construct sets of 18,848 total MLB careers evident in batting and positional searches of the Play Index (this is remarkably close to the 18,918 players that have worked in MLB since 1871). The benefit of taking this longview is to assess what it actually means to accumulate wins above replacement players — that is, what it means to be a better MLB player than the readily available person that a club could simply land off of the waiver wires or recall from the minor leagues. Using “replacement theory” to judge MLB players has many flaws that can be criticized at length, but in this case one of the strengths of WAR will be seen in the ability to categorize extremely large numbers of players that have played across nearly 150 seasons.

My goal in presenting these charts is to problematize OFP. Earlier this week, I discussed the difficulties in assessing prospect value in order to judge transactions, since there are so many vantage points from which to assess prospect value. Yet, one will notice that in that assessment, there is quite a close range of value for prospects from Grades 40-45 to Grades 55-60. By clustering MLB players according to WAR, I hypothesize that one can translate OFP onto a wider landscape of value, which will allow one to truly capture the distance between different groups of prospects. This exercise is purely abstract insofar as I will link potential OFP ranks to WAR figures, rather than judging whether (or how) players from specific OFP classes surpassed, met, or failed to met their respective ceilings.

Starting with position players (I excluded pitchers from batting WAR conversations), expectations for assessing a 50 OFP, or potential league average player, are deflated: nearly 58 percent of MLB position players earned between -1.0 and 1.0 WAR in their respective careers, and a player that has assessed 2.0 WAR is easily better than 66 percent of all MLB position players in history. But is this what we mean when we describe a 50 OFP prospect, someone who will play for any given amount of time and assess two wins above replacement? I suspect not, and have tried this experiment:

I ascribed 50 OFP to the 1.1 to 13.9 WAR group because this group may most effectively encompass the ideal of serving as an above average player for a short period of time (say, Scooter Gennett, my recent favorite example) or a long period of time (if a player averages 2.0 WARP for five or six seasons, that’s quite a good career, but probably someone who is generally average in any given season). I also favor using this method because it captures the fact that the 300 ranked organizational prospects, or even the 300 ranked plus 150 “just interesting” prospects, comprise between four and six percent of all minor league players to begin with; in 2016 BaseballProspectus lists, for example, the 300 top organizational prospects included approximately 100 55+ OFP prospects and only 51 60+ OFP prospects. It stands to reason that although 60+ OFP prospects are severely rare in the minor leagues (0.7 percent of all minor leaguers), those ranks would occupy more MLB rosters over an extended period of time because those players would stick as solid-to-great players while MLB teams shuffled through 40-55 OFP prospects to find roster value. So, I don’t think it’s ridiculous to suggest that “only” 12 percent of MLB players in history should be assessed as 55+ OFP in terms of translating WAR, and I certainly believe that 70+ OFP players should only occupy the top percentile of MLB position players.

Judging pitchers, the results are rather similar. You will undoubtedly notice, by the way, that pitchers and position players add to more MLB careers than total MLB players from 1871-present. This is undoubtedly due to position players that have pitched. However, attempting to remove these players is quite difficult from this survey, since more than seven percent (!!!) of all MLB pitchers have worked fewer than two games in their respective careers. Even double-checking this type of list would be a monumental task for this type of survey. So, the position players that pitched stay, for now…

While I was constructing the pitching list, I realized that a category discrepancy emerged between 42.0 and 65.9 WAR for position players, and 42.0 and 55.9 WAR for pitchers. This affected 47 position players in that survey, and probably muddied up the 65 OFP and 70 OFP grades to some extent. But, by that point it’s splitting hairs among the top three percent of position players in history, so I left my mistake unharmed (if you protest about the range between Johnny Damon and Willie Randolph, I understand).

Coupled together, pitching and positional WAR provide a much wider range between 40 OFP and 80 OFP. While the total monetized value of these WAR ranges appear problematic for judging prospects, one can simply correct for the fact that approximately 20 percent of prospects make the MLB (or so). Using this type of harsh depreciation, one can construct a relatively strong transactional value marker for MLB prospects:

Does it work? Let’s use an eyeball test with Ryan Braun’s expected trade value, using both depreciated and non-depreciated versions of Braun’s value (reflecting the range of values teams may apply to Braun):

This model appears to be fairly intuitive in this case. If a team uses a depreciated version of Braun’s performance and applies it to his contractual situation, Braun’s trade value would be worth approximately two 50 OFP prospects, maybe a 50 OFP and 55 OFP package if a team really wants Braun or has some value discrepancy with one of their prospects, or one 60 OFP prospect. If teams take Braun’s production value and contractual situation at face value, the veteran is worth approximately two 60 OFP prospects, or a 55 OFP and 60 OFP prospect at the very least. This works remarkably well compared to the theoretical models discussed earlier this week. Let’s run another test, using the Carlos Gomez-Mike Fiers trade (with value assessed on the day of the trade, rather than in hindsight):

This model gets relatively close to capturing equilibrium. The trouble is assessing Fiers’s contract, of course, since four arbitration years of control means that Houston can cut Fiers without paying a dime (which vastly inflates Fiers’s value). Taking Fiers’s performance value of $19.6M and Gomez’s total surplus, their value of $87.2M was countered with a prospect value that returns a historical depreciation value of $89.3M, which is much closer to equilibrium. Obviously, in the actual circumstances of a trade, it is arguable that teams are not trying to return equilibrium value, or rather, that teams are trying to return “equilibrium value” that suits their organizational needs to the greatest extent possible. Even with this caveat, it appears that stratifying the history of MLB according to WAR performance levels, and then grading each percentile with an “OFP” actually crates quite a useful “at a glance” model for analyzing trades in Cost-Benefit Analysis style.