made his major-league debut for the Chicago Cubs
on May 7, 2010
, less than two months after his 20th birthday. The next year, Castro led the National League in hits with 207 at the age of 21.
The first table below presents Starlin Castro's
career through his age-22 season (2012) as measured by Player won-lost records.
Basic Player Won-Lost Records
| ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ || ------ |
|CAREER (reg. season)|| || |
Players Comparable to Starlin Castro
So how good could Starlin Castro turn out to be? One way to try to answer this question would be to look at players who had similar starts to their career to Castro and see how they turned out.
Starlin Castro earned more than 20 eWins (and pWins) at shortstop in both 2011 and 2012 at the ages of 21 and 22. The total number of players since 1947 who have earned at least 40 eWins at shortstop through their age 22 season is 12. So there aren't a lot of close comps to Starlin Castro. Here is the list of players with >40 eWins at shortstop through their age 22 season and their final career records (through 2013).
Excluding Andrus, who is only one year older than Castro, the other six shortstops on the list here all had generally long, productive careers. They all played at least 1,900 games, amassed at least 200 pWins and eWins, and finished with a Player won-lost record that was above positional average
for their careers. Of the four players who have been retired long enough to have appeared on a Hall-of-Fame ballot, one
is in the Hall of Fame and another
is in the Hall of Merit
. Obviously, a 50% chance of putting together a Hall-of-Fame worthy career is pretty good odds.
But the above list of players is "comparable" to Starlin Castro merely in terms of playing time, not necessarily in terms of quality, value, or the shape of their production. For example, Alex Rodriguez
accumulated more Player wins over replacement level
(either eWORL or pWORL) in his age-21 season (1996) than Starlin Castro has accumulated in his career.
Let's compare three of the eight players from this list: Castro, Garry Templeton
and Robin Yount
When Castro was younger, I saw Garry Templeton
mentioned somewhat frequently as something of a "worst-case" scenario for what kind of career Starlin Castro could end up having. Templeton's career is seen as being somewhat disappointing in part because his best seasons were at ages 23 and 24 and also because he was traded straight up for a first-ballot Hall-of-Famer
in what ended up being a very lopsided trade.
, on the other hand, represents something like a best-case scenario. Yount played until he was 37 years old, won two MVP awards, accumulated over 3,000 career hits (all with the team
that drafted him) and was elected to the Baseball Hall of Fame in his first year of eligibility. Needless to say, Cubs fans everywhere would be thrilled if Starlin Castro's career even came close to approaching Yount's.
Through his age 22 season (1978), Garry Templeton had played in 361 games with a career batting line of .299/.319/.405 with 162-game averages of 4 home runs, 24 stolen bases, 20 walks, and 43 errors (he committed 24 errors in 53 games in 1976). That translated into a (context-neutral) Player won-lost record of 46.5 - 45.5, 2.2 eWOPA, 5.8 eWORL.
Robin Yount became the Brewers' regular shortstop at age 18. From age 20 - 22 (the same ages as Templeton's and Castro's first three major-league seasons), Yount batted .276/.315/.364 in 442 games with 162-game averages of 6 home runs, 18 stolen bases, 38 walks, and 32 errors.
That translated into a (context-neutral) Player won-lost record of 53.5 - 57.0, 1.0 eWOPA, 5.3 eWORL.
Starlin Castro, through 2012, had a career batting line of .297/.336/.425 in 445 games with 162-game averages of 10 home runs, 21 stolen bases, 37 walks, and 30 errors. As shown above, that translated into a (context-neutral) Player won-lost record of 58.4 - 59.8, 0.1 eWOPA, 4.4 eWORL.
Those three records seem fairly similar. But are Yount and Templeton the most similar players to Starlin Castro through age 22? To answer that question, we need to set up a systematic way of comparing Player won-lost records over a given age range.
Most Similar Player Won-Lost Records
There are four basic factors for which players can accumulate Player wins and losses: batting
, and fielding
. There are two dimensions across which player similarity can be measured: quality and quantity.
For this exercise, then, I looked at five breakdowns of Player won-lost records: batting, baserunning, pitching, fielding, and overall; and looked at two measures: total wins and wins over some benchmark. For batting and baserunning, the benchmark I looked at was average non-pitcher winning percentage; for fielding, the benchmark I used was replacement level - I chose this to allow for comparisons across fielding positions that controlled for the relative difficulty of different positions. For total wins, I looked at wins over both positional average as well as replacement level. In all of these cases, I used expected (context-neutral) wins
To start with, then, I compiled Starlin Castro's career totals (through 2012) for these various measures. I then normalized all of these figures by dividing by the standard deviation for these figures across all player-seasons for which I have calculated Player won-lost records. This puts everything on the same scale so that, for example, baserunning is weighted the same as batting, and wins and wins over average are given equal weight.
To find the players most similar to Starlin Castro at ages 20 - 22, then, I calculate the same figures for every other player for whom I have calculated Player won-lost records at ages 20 - 22. For a given player, then, for each measure, I calculate the difference between that player's value and Castro's value and square it. Squaring the differences has two effects: first, it treats values slightly higher than Castro's value the same as values slightly lower than Castro's value, and second, squaring the difference spreads out the scale, increasing the penalty for being very different - in this way, being a little bit different at everything will produce greater similarity than being identical at some things but very different at some other things. For every player, these squared differences are then summed up and the players are sorted based on the sum of squared differences. The player with the smallest sum of squared differences is then the most similar player to Starlin Castro.
The top 10 most similar players to Starlin Castro at ages 20 - 22, calculated in this way, are presented in the table below.
When I first did this analysis in the 2012-13 offseason, the second most-similar player to Starlin Castro from the ages of 20 - 22 was Robin Yount
. Some changes in my calculations have pushed Yount (and the aforementioned Garry Templeton
) off of Castro's most-similar list.
Castro's top 10 most similar players still does include two other Hall-of-Famers, however, Bill Mazeroski
and Ron Santo
, although they were both inducted more than twenty-five years after they retired.
As noted above, there were only a handful of players who were everyday major-league shortstops by age 21. Given that, it should not be too surprising that Castro's top 10 most similar players only includes one shorstop: Edgar Renteria
. The list also includes 4 third basemen and 1 second basemen, which seem like reasonable enough comps: Beltre
, and Mazeroski
were all well above-average defensively for their position, making them comparable in defensive value to a below-average defensive shortstop, as Castro
rates via Player won-lost records.
Probably the most surprising names on the list are the four outfielders: Ruben Sierra
, Greg Gross
, Rusty Staub
, and Chet Lemon
It seems somewhat odd to compare a shortstop to a corner outfielder. Nevertheless, the Sierra comp feels kind of right to me: crazy-good tools, kind of frustrating to watch, never quite felt like he really put it all together. That said, Sierra ended up putting up a pretty long and fairly decent career for a guy who never quite seemed to put it all together and was kind of frustrating to root for: 2,186 games, four All-Star games, MVP votes in four seasons, hung around long enough to play 100 games for a playoff team at age 38.
If one thinks the best way to project what a major-league player might do in the future is to look at what other players have done in the past - and this is the basis for pretty much all player projection systems - then it seems logical that the more similar those other players are to a player the more instructive they are likely to be in helping to make projections. That said, I have not done any research into how projectable Player won-lost records are and whether player similarity impacts projectability. And, of course, even if similarity translated very well into projectability, that still produces a range of outcomes in the case of Starlin Castro
that ranges from Ron Santo
to Bob Bailey
(a player who sufficiently frustrated two franchises
that he found himself with his third major-league team, the expansion Montreal Expos
, at the age of 26).
So, how good is Starlin Castro
going to be? We're just going have to wait and see.
All articles are written so that they pull data directly from the most recent version of the Player won-lost database. Hence, any numbers cited within these articles should automatically incorporate the most recent update to Player won-lost records. In some cases, however, the accompanying text may have been written based on previous versions of Player won-lost records. I apologize if this results in non-sensical text in any cases.
List of Articles