Keeping or Losing a Star-Level Player by Jeff Sullivan December 17, 2013 Ordinarily I don’t presume to know your business, but something you probably didn’t miss was Robinson Cano leaving the New York Yankees to sign a 10-year contract with the Seattle Mariners. Something you more probably did miss was the following comment, left below a FanGraphs post on the subject: I would love to see a full article on how teams do the year after losing a superstar. History could teach us a few lessons, I’m sure. Within the post, I included a few examples off the top of my head, but it was hardly anything rigorous. The idea seemed worthy of something more rigorous, so, here it is. For these purposes, I always love recalling the Mariners. The team won three more games the season after trading Randy Johnson. Then they traded Ken Griffey Jr. and improved by another 12 wins. Then they lost Alex Rodriguez to free agency and got another 25 wins better. In 1998, the team was below .500. Three missing superstars later, they were arguably the best regular-season team ever. As much as anecdotal evidence can prove a point, those Mariners teams did it. If you prefer the Oakland A’s, they got a win better after losing Jason Giambi. They actually scored more runs after losing Miguel Tejada. They allowed far fewer runs after losing both Tim Hudson and Mark Mulder. That was the story with the A’s: They wouldn’t be able to hang onto their stars, but somehow they’d be able to remain competitive. I guess this is still more anecdotal evidence, and I did promise something more rigorous. To get right into the details, I based my study around numbers available on the FanGraphs leaderboards. The first step was figuring out how to define a star. No matter what, this was going to have to be arbitrary, so I settled on a 6 WAR season. Six WAR is a lot of WAR, and those are seasons put up by good players. As a happy coincidence, this was inspired by Cano’s free agency, and last year Cano was worth exactly 6 WAR. Between 1988 and 2012, there were 399 individual player seasons worth at least 6 WAR, combining position players and pitchers. The next thing to do was figure out which of those players remained with the same team, and which of those players left, either by trades or as free agents. Of those 399 individual player seasons, 365 stuck around. That means, in 34 instances over 25 years, the offseason saw the movement of a six-win player. The only difficult step remaining then was linking team performance, and team performance in the subsequent year. I decided winning percentage would do, because the whole point is to accumulate wins. Over bigger samples, the noise should more or less go away. Let’s now examine those 34 instances. The most recent is Michael Bourn, who left the Atlanta Braves. People don’t really think of Bourn as a star-level player, but he certainly played like one, and he’s just one example of many. In their final years, players who left averaged 7.4 WAR. Their teams averaged a winning percentage of 52.9%. The next year, without the stars, the same teams averaged a winning percentage of 50.7%, which represented a drop of about 3.5 wins over a full season. In 38% of the instances, the team that lost the player went on to post a higher winning percentage. That seems like a pretty sizable drop, even if we’re barely talking about an extra loss every two months. But what we also have is something of a control group — 365 instances in which a six-win player didn’t go away. It’s worth looking at those numbers, as well. The players who didn’t leave averaged 7.2 WAR. Their teams averaged a winning percentage of 53.9%. The next year, those same teams averaged a winning percentage of 52.5%, which represented a drop of about 2.3 wins over a full season. In 40% of the instances, the team that kept the player went on to post a higher winning percentage. In short, based on these samples: Teams losing players coming off 6+ WAR seasons lost about 3.5 wins, on average Teams keeping players coming off 6+ WAR seasons lost about 2.3 wins, on average Average, of course, ignores that every circumstance is different. Sometimes teams lose star players because they’re about to begin rebuilding. Teams will behave differently if they’re still looking to win right away, or if they’re setting their sights down the road. But on average, you’re looking at a difference of about a win, as regression happens to everybody. Teams that lose star players have historically regressed from being so far above .500, but the same goes for teams that have kept their star players, because it’s an unavoidable principle. The mean is a powerful magnet. The only team in the window to lose two star-level players in the same offseason is the Atlanta Braves between 2003 and 2004. Combined, in 2003, Javy Lopez and Gary Sheffield were worth 14.3 WAR. The next year, the Braves dropped from 101 wins to 96 wins, and their Pythagorean record hardly budged. Lopez and Sheffield turned into Johnny Estrada and J.D. Drew, and Drew was worth 8.6 WAR alone. Then he left and the Braves got worse, but they still won 90 games. And their Pythagorean record was comparable. A fun fact of certain interest, even though it doesn’t have to do so much with this particular topic: Those 365 players who stuck around averaged 5.1 WAR the next season. The 34 players who departed averaged 4.1 WAR the next season. This even though, before, both groups averaged a little over 7 WAR. It could be nothing, or it could be teams making the right decisions on whom to keep and whom to lose. Or maybe players end up missing where they used to play. Or maybe more players in the “departed” group are flukes. It’s hard not to notice 2009 Chone Figgins, although the Angels definitely missed him the next year. Teams with star-level players in one year tend to perform a little worse the next year. If they lose a star-level player, they tend to perform worse than if they had kept the star-level player, but the difference is small, and might be entirely attributable to different circumstances. Certainly, losing a player doesn’t cripple a team, and losing a player in the offseason means the loss can be prepared for. Teams are always in some way prepared to lose their free-agents, and of course, trades of star players don’t just happen at random. Teams that lose stars can respond proactively or reactively, and we can see this with the Yankees, as they made up for losing Cano by signing both Jacoby Ellsbury and Brian McCann. Money can be spread around, or it can be put toward new stars. A void in one place can often mean upgrades in other places. What this captures, naturally, is that teams that lose stars work to make up for the loss. If you lose a star and do nothing about it, you’ll project to be worse by quite a bit. But that generally isn’t how teams behave, unless they’re in the process of tearing everything down. And maybe the ultimate point here is to just issue a reminder that baseball teams are made up of a whole bunch of parts, and one player can mean only so much. Baseball isn’t basketball or football, and it’s a lot more like hockey. A star player on a 90-win team contributes plenty of wins. His contribution is drowned by the combined contributions from everyone else, though. At the end of the day, losing a star is entirely survivable, because stars mean more in hearts than they do on the field.