Gaining a Star-Level Player

Inspired by the Yankees’ loss of Robinson Cano, I got to thinking about how teams have coped with missing stars in the past, which led to this post published earlier Tuesday. Setting an arbitrary “star” threshold of 6 WAR, the data sample wasn’t huge over 25 years, but on average teams that lost stars fared only a little bit worse than teams that kept them. And I didn’t even control for circumstances by, say, including payroll information. Basically, stars are great and more or less replaceable if you can prepare for a departure and spread resources around. The Yankees should survive a Cano-less existence.

The first comment below the post points to an obvious follow-up:

Obligatory: How teams fare gaining a star level player

So, now this is that.

The definition’s the same. For our purposes, a star-level player will be a player coming off a season worth at least 6 WAR. Again, this is arbitrary, and as recent examples, it misses the Angels signing Albert Pujols, the Angels signing Josh Hamilton, the Marlins signing Jose Reyes, and the Blue Jays trading for R.A. Dickey. As it happens, those acquisitions didn’t turn out to help a lot, but a cutoff is a cutoff and this is how I proceeded. I examined the 25-year window between 1989-2013, and I found 40 instances in which a team added a 6-WAR player over the offseason. You might notice that, earlier, there were 34 instances in which a team lost a 6-WAR player over the offseason. The numbers don’t match up because I excluded seasons in which a 6-WAR player changed teams, like Cliff Lee in 2009, or Cliff Lee in 2010. Those seasons work, for this post. They didn’t work, for that post.

The following table contains all the relevant information. There are the players, and the players’ new teams. There’s the winning percentage with the player, following the winning percentage the year before. There’s the change in winning percentage. There’s the team payroll with the player, following the team payroll the year before. There’s the percent change in payroll. Payroll information was taken from Baseball-Reference, and depending on your sources there’ll be small disagreements, but these numbers should convey the right ideas. We don’t need to be perfectly accurate to the dollar.

Everything:

Season Player Team Prev_W% W% W%Change Prev_$(m) $(m) $Change%
2013 Michael Bourn Indians 0.420 0.568 0.148 79 87 11%
2011 Cliff Lee Phillies 0.599 0.630 0.031 142 173 22%
2011 Adrian Beltre Rangers 0.556 0.593 0.037 57 94 66%
2011 Carl Crawford Red Sox 0.549 0.556 0.006 165 167 1%
2010 Chone Figgins Mariners 0.525 0.377 -0.148 100 87 -14%
2010 Cliff Lee Mariners 0.525 0.377 -0.148 100 87 -14%
2010 Roy Halladay Phillies 0.574 0.599 0.025 116 142 23%
2010 Javier Vazquez Yankees 0.636 0.586 -0.049 210 211 0%
2009 CC Sabathia Yankees 0.549 0.636 0.086 212 210 -1%
2009 Mark Teixeira Yankees 0.549 0.636 0.086 212 210 -1%
2005 J.D. Drew Dodgers 0.574 0.438 -0.136 93 83 -11%
2005 Adrian Beltre Mariners 0.389 0.426 0.037 82 88 8%
2005 Carlos Beltran Mets 0.438 0.512 0.074 102 101 -1%
2005 Randy Johnson Yankees 0.623 0.586 -0.037 184 208 13%
2004 Javy Lopez Orioles 0.438 0.481 0.043 74 52 -30%
2004 Gary Sheffield Yankees 0.623 0.623 0.000 153 184 21%
2004 Alex Rodriguez Yankees 0.623 0.623 0.000 153 184 21%
2003 Jeff Kent Astros 0.519 0.537 0.019 63 71 12%
2003 Jim Thome Phillies 0.497 0.531 0.034 58 71 22%
2002 Roberto Alomar Mets 0.506 0.466 -0.040 93 95 2%
2002 Jason Giambi Yankees 0.594 0.640 0.046 113 126 12%
2001 Alex Rodriguez Rangers 0.438 0.451 0.012 71 89 25%
2001 David Wells White Sox 0.586 0.512 -0.074 32 66 107%
2001 Mike Mussina Yankees 0.540 0.594 0.053 93 113 21%
1999 Mo Vaughn Angels 0.525 0.432 -0.093 42 56 33%
1999 Brian Jordan Braves 0.654 0.636 -0.019 61 73 20%
1999 Randy Johnson Diamondbacks 0.401 0.617 0.216 32 69 113%
1999 Kevin Brown Dodgers 0.512 0.475 -0.037 49 81 66%
1999 Albert Belle Orioles 0.488 0.481 -0.006 73 81 11%
1999 Rafael Palmeiro Rangers 0.543 0.586 0.043 57 77 35%
1999 Roger Clemens Yankees 0.704 0.605 -0.099 67 87 30%
1998 Kevin Brown Padres 0.469 0.605 0.136 37 47 25%
1998 Pedro Martinez Red Sox 0.481 0.568 0.086 44 57 31%
1998 Chuck Knoblauch Yankees 0.593 0.704 0.111 62 67 7%
1997 Roger Clemens Blue Jays 0.457 0.469 0.012 31 47 54%
1994 Rafael Palmeiro Orioles 0.525 0.563 0.038 29 39 33%
1993 Greg Maddux Braves 0.605 0.642 0.037 35 42 20%
1993 Barry Bonds Giants 0.444 0.636 0.191 33 35 6%
1992 Greg Swindell Reds 0.457 0.556 0.099 26 36 37%
1991 Darryl Strawberry Dodgers 0.531 0.574 0.043 22 33 49%

40 instances in which a team added a 6-WAR player. Out of those 40:

  • 26 instances in which the winning percentage went up
  • 2 instances in which the winning percentage didn’t change
  • 12 instances in which the winning percentage went down

Between 2003 and 2004, the Yankees added both Gary Sheffield and Alex Rodriguez, and both years they finished 101-61. The biggest gain belongs to the 1999 Diamondbacks, who went from expansion to contender overnight. The biggest loser is the 2010 Mariners, who added both Chone Figgins and Cliff Lee and then wound up one of the worst teams in baseball. Lee didn’t survive the season; Figgins didn’t last his whole contract.

But, okay, we care about averages. On average, in the years before, these teams won 53.2% of their games. In the years with the new star players, these teams won 55.3% of their games, an increase of about 3.5 wins over a full season. And remember, that’s not just a 3.5-game improvement — the teams were also fighting regression to the mean, which would’ve taken them closer to .500 as a group. Immediately, the new players helped, just as you’d expect them to. When you get a star, you figure his best season under your control will be his first.

But the table doesn’t just have winning-percentage data. On average, payroll went up about 22%. Which is exactly what you’d expect, since a team that acquires a star player is usually going for it and going for it implies some extra spending. There were actually a few instances in which a team acquired a star and trimmed payroll, as the 2010 Mariners did, or as the 2004 Orioles did. But these are exceptions, and in a dozen cases, payroll increased by 30% or more.

When you increase payroll, you increase your win expectations, because money buys players and players are wins in short-sleeve costumes. So you’d expect the teams to get better anyway, just from spending more, even if they didn’t get stars. I can’t say by how much you’d expect them to get better, but here are some interesting points:

  • in these 40 instances, teams increased payroll by an average of 22%, and they improved by 3.5 wins
  • in 32 instances, teams increased payroll, coming out to an average gain of 30%, and they improved by 5.3 wins
  • in 27 instances, teams increased payroll by at least 10%, coming out to an average gain of 35%, and they improved by 4.5 wins
  • In 14 instances, teams increased payroll by at least 25%, coming out to an average gain of 50%, and they improved by 4.9 wins

Take those 32 instances in which teams increased payroll. The 16 teams that increased payroll by the greatest percentage added about five wins. The 16 teams that increased payroll by the lowest percentage added about six wins. The first group had an average payroll gain of 42%. The second group had an average payroll gain of 13%.

For fun, let’s further mix some signals. Between 1998-1999, the Diamondbacks more than doubled payroll, adding Randy Johnson, and they went straight from 65 to 100 wins and a berth in the playoffs. Between 2000-2001, the White Sox more than doubled payroll, adding David Wells, and they went from 95 to 83 wins and a third-place finish. The next-biggest payroll hike belongs to the Dodgers between 1998-1999, when they added Kevin Brown, and they went from 83 to 77 wins. There’s a correlation between money and success. The coefficient isn’t 1.

The gist: of course, on average, teams who add star-level players tend to improve, at least in year one. They improve, on average, by a handful of wins, but they also tend to increase payroll, sometimes by kind of a lot, and any increase in payroll should lead to greater success on the field because money buys numbers and numbers are wins. I don’t know enough to say whether things are changing, but it’s worth keeping in mind additional recent cautionary examples like the Angels, Marlins, and Blue Jays, who have gone for it and gotten burned. They added stars that didn’t quite meet the threshold above, and they didn’t succeed. Yet, the Phillies were thrilled to have added Roy Halladay and Cliff Lee. There’s a time and a place, and it’s always important to do things wisely, even if you feel like one player can put your team over the top.

The long and short of it is that the Yankees should be fine without Robinson Cano, relative to the Yankees with him. And the Mariners should be better with Robinson Cano, at least right away, but they probably won’t be better by leaps and bounds unless they do plenty more or end up getting lucky. Stars are the best players in baseball, but the best players in baseball are a relatively small part of baseball.




Print This Post



Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.


35 Responses to “Gaining a Star-Level Player”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. Dan Gladden says:

    what is a Greg Swindell?

    Vote -1 Vote +1

  2. Boris Chinchilla says:

    Screw you figgins, wherever you are

    Vote -1 Vote +1

  3. Tom Steele says:

    Obligatory: How teams fare neither gaining nor losing a star-level player.

    +12 Vote -1 Vote +1

  4. Jack Z says:

    Mmmm… Chone Figgins pudding…

    Vote -1 Vote +1

  5. Pat G says:

    Jeff, you’re an excellent writer, and often pick some really interesting (read unique) perspectives to write about. But what percentage of season after season change could acquiring one player account for? I mean, this study, at best, has to be WROUGHT with statistical noise.

    The gist of this article seems to be, teams can get better, stay the same or get worse if they add a star. They get better more often than they get worse, and rarely stay EXACTLY the same.

    Or it was a trip down memory lane for the last times that stars changed hands.

    Vote -1 Vote +1

  6. The Foils says:

    Appreciate the quick turnaround.

    Vote -1 Vote +1

  7. adohaj says:

    Jeff you are awesome, thank you. Readers suggested valuable things in the comments of the previous article and you delivered. This is how the internet is supposed to work! Great stuff.

    Vote -1 Vote +1

  8. isavage30 says:

    Hmm, that Indians’ 2013 payroll appears to be at least $5 million too high

    Vote -1 Vote +1

  9. Jay29 says:

    Damn, Jeff was productive as hell today.

    Vote -1 Vote +1

  10. Dave says:

    This may not be the correct forum to ask this question, but I was wondering if someone could help explain some elements of wOBA, wRC+, and WAR to me in the following context: Josh Donaldson had the highest WAR among 3B last year at 7.7. Meanwhile, Evan Longoria was 3rd at 6.8 WAR and Adrian Beltre was 5th at 5.2 WAR.
    Player: WAR wRC+ wOBA Avg/OBP/SLUG 2B/HR UZR
    Josh Donaldson 7.7 148 .384 .301/.384/.499 37/24 9.9
    Adrian Beltre 5.2 135 .379 .315/.371/.509 32/30 -1.2
    Evan Longoria 6.8 133 .360 .269/.343/.498 39/32 14.6
    A. Beltre (2012) 6.3 142 .388 .321/.359/.561 33/36 10.0

    When I look at those numbers, I fail to see how Josh Donaldson is worth over 2 more wins than Adrian Beltre and nearly 1 win over Evan Longoria. Moreover, Beltre had a higher average, slugging, more HRs, and fewer K’s, while Donaldson had more walks and doubles. As far as the offensive production, one could argue that Beltre was just as productive if not more, as the higher walk rate for Donaldson led to just 26 more walks than Beltre for the season, and Beltre made up for those walks by having 25 more hits on the season than Donaldson. I understand that Donaldson was better on defense than Beltre, that accounts for part of the discrepancy in WAR, but why does Donaldson have a much higher wRC+ and wOBA than Beltre as well? Aren’t those stats (wRC+ & wOBA) independent of defensive play, and if so, what exactly is the reason for the large gap in these stats as well between the two players?

    If you say the large, nearly 3 win gap in WAR between the two players is solely due to the defensive contributions by each, then it raises the following question for me, where in 2012, Beltre had a higher wOBA, similar wRC+, and a higher UZR than Donaldson did in 2013, but his WAR was still only 6.2 in 2012. If I go back and look at Beltre’s 2012 numbers, his offensive production was actually very similar to 2013, if not slightly better, due to a higher batting average and slugging pct, by having more extra base hits. During the 2012 season Beltre had a wOBA of .388, higher than Donaldson’s wOBA in 2013, and a wRC+ of 142, which again was lower than that of Donaldson’s for some reason, but I understand that from year to year there is a variation in the weights used for calculating those stats. Beltre’s UZR in 2012 was 10.0, which is also higher than Donaldson’s UZR in 2013, albeit by 0.1. Why then is there such a large gap between the amount of WAR accumulated by Donaldson in 2013 (7.7) and Beltre in 2012 (6.3)? I guess a simpler question is, why exactly is Donaldson’s WAR so high in 2013 (what component or components make his WAR total go up to 7.7 when similar offensive production was worth only 5.2 WAR for Beltre, and similar offensive and defensive production was worth only 6.2 WAR in 2012 for Beltre?) Longoria had a slightly worse offensive season than Donaldson, but he was better on defense, yet he was also worth 0.9 WAR less, so what exactly makes Donaldson so much more valuable than these other players?

    Vote -1 Vote +1

    • Travis L says:

      is wOBA park-adjusted?

      Vote -1 Vote +1

    • Luke says:

      Dave:

      Scrolling down to the “Value” section of Donaldson’s and Beltre’s stat pages will help to answer a lot of your questions.

      Your statement about Beltre getting 26 fewer walks, but making up for it with 25 more hits, is a little misleading because Beltre also had 22 more plate appearances. With those 22 more plate appearances, he made 22 more outs (works out nicely, doesn’t it?). Those 22 outs count for a lot. XBHs are important, and singles are better than walks, but not by as much as you’d think since the most important thing to do at the plate is to not make an out. In other words, Donaldson’s advantage in OBP is worth significantly more than Beltre’s advantage in AVG and SLG, and accounts for Donaldson’s wOBA being 5 points higher.

      wRC+ then takes wOBA and adjusts it for park effects and scales it so 100 is league average. The difference in park factors between Texas and Oakland is huge. As a result, that 5 points wOBA difference corresponds to a 13 point difference in wRC+, which over a full season of plate appearances ends up being worth 8.6 batting runs.

      The end result is that Donaldson was 8.6 runs better with the bat, 2.9 runs better on the base paths, and 12.8 runs better with the glove, adding up to 24.3 runs (or 2.4 wins). Basically, he made fewer outs in a much worse ballpark, played better defense, and ran the bases better.

      Vote -1 Vote +1

    • Luke says:

      The difference between 2012 Beltre and 2013 Donaldson also has a lot to do with park factors. Basically, a .384 in Oakland is just awesome.

      Vote -1 Vote +1

    • ZB says:

      For wOBA: Look at the triple slash numbers. Beltre had a 10 point advantage in SLG, but Donaldson had a 13 point advantage in OBP. each point of OBP is more valuable, according to the weights, than each point of SLG. Why? OBP essentially measures the number of times a player doesn’t make an out. Outs are the currency you spend to earn runs and you only get three of them at a time so conserving them is kind of important. Basically it boils down to this: you get more production per out by making fewer outs than you do by being marginally more productive per hit.

      So that explains the wOBA diffeerence, hopefully. What about wRC+? wRC+ takes wOBA and adjusts it for park and league and then displays the number of runs created relative to the league average. Beltre played half his games in one of the best hitters parks in the League while Donaldson played half his games in a pitcher’s park. Basically that means that a run in Oakland is more valuable than a run in Texas. That is, because more runs are scored in Texas (on average) each run is worth fewer wins.

      Donaldson had a better offensive performance in a tougher offensive environment. The UZR difference accounts for about a 1 WAR difference as well. Does that answer your question?

      Vote -1 Vote +1

  11. gareth says:

    Jeff Sullivan is awesome. So is szymborski and sarris.

    Fangraphs, razzball and mlbtraderumors.com are my three go to baseball websites.

    Keep up the good work!

    Vote -1 Vote +1

  12. icbeast says:

    10 out of 40 signed by the same team, blerg. A list of every team and the highest previous season WAR free agent they have signed would be interesting. At least to a Brewers fan who’s just wondering what team is last on the list. Looking through their signings it looks like it was Yuni!!! Oh wait, that’s a negative WAR.

    It’s actually Aramis, then Randy Wolf, and I think Mike Cameron. And if you add Aramis’ and Wolf’s previous season WAR together they even it make it to 6.2 WAR. At least I didn’t spend the last hour staying up way too late figuring that out.

    Vote -1 Vote +1

  13. Luke says:

    I was surprised that Curt Schilling didn’t make the cut as I remember him being elite when he moved from Arizona to Boston in 2004. Turns out that he just missed 6 WAR, only getting 5.7 in 2003 because he got injured and only made 24 starts. He was probably only 1 or 2 more starts away from 6 WAR.

    Vote -1 Vote +1

  14. Zack says:

    Interesting that only 18 different teams have added a 6-win player since 1991.

    25 of the players were acquired by AL teams, 15 by NL teams.

    9 different teams have acquired more than 1 such player, with 6 of those teams (NYY, BOS, BAL, PHI, NYM, ATL) playing in one of the East divisions and none playing in one of the Central divisions.

    Breakdown according to current division alignment:
    AL East – 16 6-win players acquired since 1991
    AL West – 8
    NL East – 7
    NL West – 6
    AL Central – 2
    NL Central – 1

    Full breakdown of teams and number of 6-win players acquired since 1991:
    NYY – 10
    TEX, SEA, BAL, PHI, LAD – 3 each
    BOS, NYM, ATL – 2 each
    CLE, CHW, LAA, TOR, HOU, ARI, SDP, SFG, CIN – 1 each

    Vote -1 Vote +1

  15. RC says:

    I HATE HATE HATE when shitty statisticians do this.

    If you’re trying to create a group that is similar to Cano, you don’t set his performance as the absolute minimum to be included in the group. You poison the entire sample doing that.

    Cano put up 6.0 WAR, so something like 5-7 WAR would be much more useful of a comparison group (or more likely 5-8 or so, to control for the fact that there are more 5 WAR players than 7 war players)

    -6 Vote -1 Vote +1

    • This isn’t specifically about Cano. These thoughts were just inspired by his movement. That he happened to put up exactly 6 WAR was a coincidence, in that I selected the threshold independent of that. But hey, thanks, you’re all right.

      Vote -1 Vote +1

      • RC says:

        It doesn’t matter what you set out to do.

        The entire thrust of your article is on how the yankees and mariners are going to do because of the adding/subtracting a star player, but the problem is that, with your definition, and the error margins in WAR, its as likely that Cano doesn’t fit into that category as that he does.

        -6 Vote -1 Vote +1

  16. Mr Hanusa says:

    You should put some kind of disclaimer for M’s fans when you mention the name Ch_ne Figg_ns

    Vote -1 Vote +1

  17. dl80 says:

    Am I the only one who doesn’t think of a star player as “a guy who once put up 6 WAR”? I think of it as a guy who was well above average for several years in a row.

    What about doing this with guys who put up multiple 4+ WAR seasons in a row? Or 15 WAR over 4 seasons or something similar? That might give us a better picture by removing guys like Figgins and Bourne who had one huge outlier year.

    Obviously not many players put up multiple 5 WAR years in a row, so the sample may get to be too small rather quickly, but I’d be interested to see.

    Vote -1 Vote +1

  18. locdog284 says:

    I appreciate all the answers to the Beltre Donaldson question.. helps clear up a lot of the confusion.. This seemed to be the case again in 2012 when Chase Headley was worth 7.2 WAR vs Beltre’s 6.3, despite Beltre bettering Headley in almost every category including UZR, except OBP and SB’s. I guess the fact that Headley plays half his games in a pitchers park and still got on base more than Beltre was worth more than the other categories in which Beltre had an advantage. I actually think this case makes for a better example of how much a park can influence WAR than in the 2013 case of Donaldson and Beltre, since in 2012, Beltre was ahead of Headley in more categories including defense, and yet still had a lower WAR.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>