There’s a lesser-known algorithm in baseball analytics that rarely shows up in the press but can give us an interesting view of how a young player might develop. That algorithm is called similarity score. Let me tell you a story to illustrate how a similarly score can help us think about players.
On June 5, 2017, I sent this tweet out to respond to the hordes of Astros fans advocating that Alex Bregman be traded for an ace so that the Astros could win the 2017 World Series:
I know it is hard to remember now how loud these fans were. To be fair, Bregman was batting 0.250/0.315/0.408 in his second season (after batting 0.264/0.313/0.478 as a rookie) as the Astros took the field in Kansas City that night. Even A.J. Hinch had Marwin starting that night to give Bregman a break. Despite all of that, I boldly predicted he would be the next Chipper Jones or Derek Jeter. Why? I knew who Alex Bregman was before the Astros drafted him. I was aware of the workout stories from his time in LSU. He was picked number two overall in the 2015 draft. He had succeeded at every place he had played baseball. I had faith that the Alex Bregman we all know now would figure out how to excel in the major leagues – and he has.
This is one of my proudest @LarrytheGM moments on Astros Twitter. I refuse to allow the same negativity that crowd had with Bregman be extended to Kyle Tucker today (see my previous comments about Tucker’s potential). Will Tucker prove me as right as Bregman did? It is too early to tell, but I have confidence. I will give you two names as my comps for Tucker:
Jack Clark (0.267/0.379/0.476/0.854- 340 HR, 77 SB, 53.1 WAR) eventually moved to 1B and DH. This would be my negative defense and speed projection for Tucker.
Andrew McCutchen (0.287/0.378/0.481/0.859- 223 HR, 185 SB, 42.0 WAR, active) would probably be my ceiling for Tucker.
How did I decide these? Baseball Reference offers similarity scores for players once they have a couple years of experience. I used these similarity scores to find a player I think Tucker could be. I paged through different similarity scores for those players and looked at their career stats. Clark and McCutchen represent the range I expect of Tucker. This was the exercise I used to decide Bregman really could be Chipper Jones as well.
This leads me to a broader question. Who are the three most similar players for the 2019 Astros at their current age and what does that indicate about each of them? It can be an interesting exercise, if you have any baseball historical reference about who these players are.
For each of these similar players I included their Baseball Reference career WARs. I averaged these three career WARs and sorted them based on the average. See below:
Players shaded in yellow are active players. Players shaded in orange are Hall of Famers. The last column is the needed additional WAR if the Astros players are to reach the average career WAR of the three most similar players. You can consider it a projection of the possible remaining WAR for each Astros player.
A few highlights from this table:
- Notice who Correa, Altuve, and Bregman are most similar to at this age point in their careers. Each of these players is most similar to at least one Hall of Famer. At best, we are watching an infield of Ripken, Sandberg, and Jones. Think about how awesome it is to be an Astros fan today. How many fans get to watch three potential Hall of Fame position players simultaneously?
- Springer and Brantley are similar to very good but not historically dominant players. Both have amassed career WARs already close to the career WAR of the average of the three most similar players. I believe they will outplay their peers.
- Reddick is already beyond the career WAR for his most similar players. This is likely due to his superior defense. It may pose a warning sign that he may not have many productive years left.
- The other position players are fairly pedestrian in how they project.
- It is too early for similarity projections for Stassi, Kemp, or Tucker.
- Verlander has similar players who are Hall of Famers or near-Hall of Famers. He will be too.
- Cole and McCullers compare to very good players.
- Miley and Rondon surprisingly project to have some good years ahead.
- Osuna projects to be a career dominant reliever. This is an example of why the Astros traded for him.
- It is too early for similarity projections for James, Valdez, or Perez.
If you know the individual historical players, this can be an interesting way to look at your current Astros. If you don’t know them, let this be your spring training homework assignment to get excited about who your Astros could be.
Stats per Baseball Reference