Friday, September 16, 2011
A second look at Ian KinslerPosted by Jeffrey Gross at 2:51am
Statistics current through Sept. 13.
People all around the world of fantasy "hate" Ian Kinsler. It's time for that to end.
In the preseason, I received a lot of criticism for ranking Ian Kinsler as my number two fantasy second baseman for 2011 behind Chase Utley, who was still healthy at the time. Some critics legitimately focused on the level of risk imposed by Kinsler's health; I thought he was a top five or 10 second baseman even if inundated with injuries. For what it's worth, Kinsler has appeared in 142 of the Rangers' 148 games this season.
Other critics focused on his raw stats, claiming that Kinsler had never proven himself on par with the likes of Robinson Cano. In reality, as I pointed out, the two players rated very similarly over the past four seasons:
Despite a clear differential in games played, they have comparable absolute home run and runs scored production, with Cano having a noticeable edge in batting average and RBI. Kinsler owned a huge advantage in stolen base totals. Cano had a substantial lead in games played over Kinsler.
Health is obviously a factor of value, but when you consider how close the two are in value even when one is constantly making trips to the disabled list, comparing their rate stats is not irrational. If we prorate Kinsler's numbers to the same number of games played as Cano, Kinsler produced the better rate stats overall. Hence, my claim that Kinsler's ceiling and floor were high enough, despite an ankle-injury-derailed 2010, to warrant top three consideration.
Eight months later, my widely panned ranking has turned out to be more or less correct. Kinsler has been the third best second baseman this year per Yahoo (second best, per my Z-Score sum calculations), behind Cano and Dustin Pedroia. Over the past 30 days, Kinsler has been the No. 1 fantasy second baseman. He has also been an elite player overall. So far through 2011, Kinsler's season qualifies him as a top 30 overall fantasy player (No. 28, by Yahoo's calculations) and top 10 fantasy hitter (via my Z-Score sums). Over the past 30 days he's been even better, and Yahoo values him as a top 10 overall player (No. 7) over that period.
Cano (Yahoo's No. 14 ranked overall, my No. 9 overall hitter in baseball via Z-Score sums), who I ranked as my No. 3 second baseman, has been the better fantasy baseball player. Thus, my argument of Kinsler over Cano was incorrect...right?
In terms of their results, yes. Objectively speaking, no matter how you slice their numbers on the season (which is honestly all that really matters), Cano has been the better fantasy player. Yahoo has their values pegged much farther apart than my Z-Score sums (below), but Cano's produced more in the relevant fantasy categories:
Z-Score sums are determined by taking the difference between a player's value in a given stat from the league's mean production in that stat, and then dividing by the standard deviation. This gives you a strong measure of relatively in evaluating how good players are in a given statistical category. Z-Scores, for example, are a useful tool for answering the question of which is more "valuable" in a vacuum—20 RBI or five stolen bases. By taking the Z-Score for each Roto category and summing them, we get a Z-Score sum which tells you what each player's overall relative value is.
To get an accurate Z-Score sum, you need an accurate pool of players. The pool of useful players from which to calculate Z-Score sums will vary wildly from league to league, depending on the number of teams, the number of players per position per team, the format of play, etc. To simplify an election bias, I simply included all players with 100 or more plate appearances to determine the relative value of hitters for this exercise.
On one hand, this inflates the value of players who have more counting stats, as the players near the bottom of the plate appearance threshold —who probably have little fantasy value—drag down the "mean" and increase the standard deviation. At the same time, this effect occurs equally on all players in the pool. Hence, while the means and standard deviations may not be accurate representations of the true fantasy means and standard deviations of any given league, they have a relatively similar effect for rankings purposes. Alternatively explained, though the absolute Z-Scores numbers may be skewed, their relativity should be reliable for our purposes.
Mathematics aside, the point remains salient. Cano has been better, but their value has been pretty close and both have been pretty excellent.
But Kinsler should be doing a lot better—at least if you buy into BABIP-luck theories. So let's look at BABIP luck-neutralized production of Cano and Kinsler and see who should be faring better.
To account for BABIP luck, I calculated the xBABIP of Kinsler and Cano using THT's own xBABIP tool. Using each player's xBABIP, I then calculated the number of hits each player should have expected to produce irrespective of luck. Using this expected hits total, I recalculated the player's expected batting average.
Then, using the differential between each player's expected hits (xH) total and his actual hits (aH) and holding each player's production rates constant, I calculated an expected stolen base, runs scored and RBI differential to be added to/subtractedfrom each player's actual 2011 numbers. I ignored the BABIP effect on home runs, as those are not balls in play and because I did not want to have to go through each player's detailed game log to find "robbed home runs" and then also account for "lucky" home runs.
I also performed the same analysis and "number crunching" using each player's career BABIP for comparison.
Here is each player's relevant BABIP data:
|Name||2011 BABIP||2011 xBABIP||Career BABIP|
Using some mathematical "reverse engineering," here are how Cano and Kinsler's BABIP-luck neutralized 2011 stats stack up using their career BABIPs:
And using their 2011 xBABIPs:
As you might notice, when you strip out Kinsler's poor BABIP luck, whether you use his career BABIP or 2011 xBABIP, he should have been a superior fantasy producer. In fact, Kinsler's xBABIP adjusted stats would place him as the fifth overall player, behind Matt Kemp (13.5 Z-Score sum), Curtis Granderson (12.9), Jacoby Ellsbury (12.6) and Ryan Braun (11.6). That's a hair ahead of Jose Bautista (10.2). Even if we hold Kinsler's counting stats constant and adjust only his batting average to reflect his career BABIP, Kinsler's aggregate production this season outweighs Cano's. As it should be evident, Kinsler is an elite player of the highest order when healthy, even if he did not put it all together in 2012.
Of course, results are all that matter in fantasy, and in that regard, Cano has had the better 2011 season. However, for the future, it is clear that my initial hypothesis—that a healthy Ian Kinsler has top-tier player upside—rings true. For 2012, neither Kinsler or Cano will come cheap. However, Cano and Pedroia will likely cost more, meaning Kinsler, if healthy, could be a top 10 player at a relative discount
Jeffrey Gross is an attorney (and die-hard Cubs fan) who currently resides in the suburbs of Chicago. In addition to writing for The Hardball Times, he also writes about craft beer as part of a side project blog titled "saBEERmetrics." He previously worked for The Daily Illini and Northern Star newspapers as a film critic and sportswriter (respectively). You can reach him by email at saBEERmetrics AT gmail DOT com.