Your 2009 Shadow Twins - Catcher

A few years back, John Bonnes (aka Twins Geek) was blogging for the Minneapolis Star/Tribune and posted a look-ahead to a season where a number of Twins from the 2002-2004 pennant era would likely end up changing. Inspired, I responded with my first-ever list of Shadow Twins -- a roster list derived from taking each expected regular and replacing him with a highly similar or better (sim score 900+) player at the same age from that player's baseball-reference.com page. Part of the reason I did this was to give people a chance to evaluate the roster in terms of players who were, at least in theory, better known than the players the Twins were about to go with. This, I hoped, would provide some insight as to how to expect that Twins roster to perform.

The Shadow Twins have become something of an annual tradition now, and though many of the folks on the roster are much better-known quantities these days, I still find occasional surprises while working through the comparison process; who a player is comparable to can tell you a lot about who he is.

Now I'm not claiming that similarity scores (originally developed by Bill James and refined slightly for baseball-reference's use) are the best way to predict how a player will perform in a given season -- just because Carlos Gomez's top of-age comp is Paul 'Motormouth' Blair doesn't mean that Gomez will finish his age-23 season hitting .293/799 with 12 triples and 11 homers as Blair did in 1967. But in my years of doing this analysis, it does appear to be true that, once a player establishes a given level of big-league production, he tends to remain in a certain 'cluster' of comparable players season after season, which provides valuable information about his relative worth as an offensive player.

Let's take Kent Hrbek as an example. Hrbek, the Twins first-baseman during their 1987-1991 championship era, came up with the ballclub for a September cup of coffee, then exploded onto the scene in 1982, finishing second in the Rookie of the Year award voting behind Cal Ripken. As a young player, Hrbek hit well, and by the end of his second full season in the major leagues, contemporary first-baseman Will Clark was one of his top comps: Clark was Hrbek's third-bese of-age comp at age 23, moved up to second at ages 24 and 25, and was Hrbek's best comp from ages 26 through 29, the prime of Hrbek's career. Hrbek began to decline after he turned 30, but even so Clark was Hrbek's fourth-best comp at ages 31 and 32 (with a score above 900 both seasons), before finally falling to eighth at age 33 and dropping off Hrbek's list entirely after Hrbek's final season at age 34.

Now it's possible to read too much into this as well. While Hrbek's raw offensive numbers were extremely similar to those of Will Clark in their primes, it also needs to be noted that similarity scores don't make any adjustments for era, league, or park. In this case, there's no need to make an era adjustment, since Hrbek and Clark were contemporaries, but Hrbek was hitting in a more favorable league for offense as well as in a significantly more favorable ballpark, and this can be seen in their relative adjusted OPS+ numbers compared to their raw numbers:

Career through age 29 --
Hrbek: .290/864, 201 HR, 724 RBI, OPS+ 132
Clark: .299/872, 176 HR, 709 RBI, OPS+ 145

Had Will Clark played for the Twins while Kent Hrbek played for the San Francisco Giants, the two men's hitting numbers wouldn't look nearly as similar as they do. Clearly, Clark was the better hitter, but this shouldn't distract us from recognizing that for Hrbek to even be in Clark's ZIP code in his prime, even with help from the Metrodome, meant that Hrbek was himself a heck of a hitter -- his OPS+ over this era is far closer to contemporaries John Olerud (135) and Jack Clark (136), who werw also known as outstanding hitters.

So while similarity scores don't give you a sense of how a player will perform in any given year (that's what systems like PECOTA and the Marcels are for), they can give us a ballpark estimation of the company a player is keeping with respect to his talent level; it's in this way that the Shadow Twins exercise is valuable, in my mind.

One last aside before we begin: I've been taken with Stick & Ball Guy's 'six word scouting reports' of late, so I've provided my own such reports for each of the players I list as a Shadow Twin. If you're not familiar with the concept, let me demonstrate using a six-word version of my previous essay about Kirby Puckett:

Worst writer-elected Hall-of-Famer

So, without any further ado, let's get started!

Catcher - Jason Kendall (945)

Kendall isn't Joe Mauer's closest comp -- that would be lively-ball era catcher Shanty Hogan, who spend most of his career playing for various New York ballclubs. Nor is Kendall Mauer's best modern-day comp -- that would be Yankee second-baseman Robinson Cano. I pick Kendall over Hogan based on era, and Kendall over Cano based on position and movement between seasons -- Mauer and Cano were much more similar when both were younger, but that similarity has been slowly declining (just as Kent Hrbek's early similarity to Jim Rice slowly declined). Also, if the Twins keep Mauer behind the plate, and there's no reason at this point to think they won't, his career will likely have a shape much more like fellow catcher Kendall's rather than second-baseman Cano's.

(Aside: All you need to know about the difference between the Twins offense, fundamentally, and that of the Yankees is that Mauer spent pretty much the season [after April, at least] hitting third in the Twins lineup, while Cano spent nearly the whole season hitting anywhere from 6th to 8th in the Yankee lineup.)

The striking thing in looking at Kendall's future, is that Kendall never did learn to hit for power, though based on his minor-league career, that shouldn't have been surprising: Kendall hit just 16 homers in over 1300 minor league at-bats, and has never hit more than 14 in a season in the majors. Expect Mauer's quick bat to slow down as he spends more and more time behind the plate, so that by age 30, his power will actually likely decline as he surrenders power to maintain contact with the ball, just as is evident in Kendall's numbers from age 29 onward. If you don't believe me, then consider that Kendall's minor league home run total and rate are both superior to Mauer's (9 HR in over 1000 ABs).

Six-word scouting report: Excellent hitter, but not for power