Can minor league statistics be used to project which minor league prospects will be successful in the major leagues? Obviously such a model would be extremely useful in constructing major league teams and their farm systems. It's not a new topic. Many years ago, Bill James developed major league equivalents (MLEs), which translated minor league batting stats into major league stats. MLE algorithms continue to be churned out, but they fall short of projecting which prospects will be successful.
Chris St. John at Beyond the Boxscore has developed a projection system for minor leaguers' success rates in the major leagues, which he calls Javier. The BtB author has successively revised and built up his system. I wrote about a previous version of Javier here. St. John recently published output from a revised Javier Model: "Javier Prospect Comparison System, Now With Speed."
You can proceed to the article in the link above to read more details about the model, and the revisions. One of the biggest changes is that prospect success is now measured by a composite Value Over Replacement Player (VORP), similar to the Baseball Prospectus VORP metric. As discussed in the article, the model has been revised with the addition of a speed score, age adjustments, and other algorithm changes. As in the original model, statistical z-scores surrounding ISO, BB%, K%, and speed are used compare players to their peers. The addition of a speed variable appears to have noticeably changed the model's rankings of players.
The system uses thousands of minor leaguers over 36 years of minor league play to develop profiles or comps for each current prospect. This is critical for Javier's estimation of the probable major league performance level of minor leaguers. The previous minor leaguers with the highest, elite level of VORP computed by Javier from minor league performance results in a list of 35 players who are among the best major league hitters in recent memory. The top five in Javier VORP: Ken Griffey, Jr., Adrian Beltre, Johnny Damon, Andruw Jones, Alex Rodriguez. Craig Biggio is in the middle of the list, and Vlad Guerrero is No. 35. These are the 35 previous players with a Javier VORP of 60 or more. St. John suggests that a Javier VORP of 30 or more is noteworthy.
Here are the Astros' prospects with a VORP of more than 30.
|Astros' Prospect/Javier VORP|
The surprise is how well Brett Phillips fared in the Javier projection. Only two minor league prospects not currently in the majors had a higher VORP than Phillips--Kris Bryant and Joey Gallo. Phillps VORP is within the elite range of the 35 previous major leaguers mentioned above. The depth of the Astros' hitting prospects is demonstrated by the eight prospects with a Javier VORP above 30. The Astros have the second most hitting prospects above 30, behind only the Cubs, with 9.
Back to Phillips, Javier likes Phillips better than other uber prospects like Buxton, Pederson, and Seager. I'll repeat. Surprising. Not so surprising is that Springer is ranked as a probable very good ML player. Marisnick should have a good shot at starting in the majors, if you believe Javier. Falling just below the noteworthy 30 VORP threshold is Delino DeShields with 27. The Astros' third basemen and third base prospects fared less well---Dominguez, Moran, and Davis are graded below 10 VORP, due in part to poor speed. For some reason, Rio Ruiz was excluded from the data base.
How much weight should we put on the Javier results? I don't know. I would characterize this model as a work in progress. And, as noted by St. John, all prospects have a liklihood of failing which is higher than the probability of becoming a productive major leaguer. So, don't make any bets. But I think Javier results are useful information. And that's how I would characterize it: a piece of helpful and interesting information to include in your evaluation of prospects.
(Note that the BtB article has excel spreadsheets you can download. I corrected an error in the spreadsheet data which mistakenly placed Singleton in the Phillies' organization.)