March 13, 2012; Goodyear, AZ, USA; Cleveland Indians second baseman Jason Kipnis, easier to project as a singer than a hitter. Mandatory Credit: Rick Scuteri-US PRESSWIRE
The Indians, as I have said many times during the offseason, are a difficult team to project. Most of the team's key players, particularly on offense, are either young or returning from some kind of injury (e.g. Sizemore) or setback (e.g. Choo). This makes the business of projecting how they will all fit together difficult. This is clearly demonstrable by looking at the systematic projections of individual players on the team.
Tom Tango's Marcel projections, the smartest "dumb" projection system (or dumbest "smart" system), can be used to show this more specifically for Cleveland. In addition to forecasting the usual rate and counting metrics, Tom's projections (found here) also include a measure of reliability. The reliability measure is a product of how long a track record a player has, t
he volatility of that track record, and their forecasted playing time for 2012. The less of a track record, the less forecasted playing time, and the more volatility, the less reliability in the forecast.
Among hitters with a median plate appearance projection of 400+ (i.e. projected starters), the average reliability for those projections is 0.82. The Indians only have three regulars who clear that mark, Asdrubal Cabrera (0.85), Shin-Soo Choo (0.84) and Casey Kotchman (0.83). The Indians have a number of guys, instead, who have a fair amount of uncertainty in their projection. Travis Hafner (0.80), Michael Brantley (0.78), Carlos Santana (0.77), Grady Sizemore (0.75) and Lou Marson (0.70) all fall in the 0.7-0.8 range. You could also add Matt LaPorta, Fred Lewis, Ryan Spilborghs and Felix Pie to this group if you want to. Even more significant are the guys in line to get significant playing time who really can't be projected at all. Jack Hannahan (0.69), Shelley Duncan (0.66), Jason Donald (0.63), Aaron Cunningham (0.51), Lonnie Chisenhall (0.48) and Jason Kipnis (0.38) all fall into this group. Compare this with the Detroit Tigers, for example, who don't have a projected starter or significant part with a reliability below 0.78.
Pitching is more of a crapshoot to project in general and less of an uncertainty for the Indians relative to the league as a whole. The exception is the back of the rotation. Among pitchers projected to go at least 120 innings, the average Marcel reliability is 0.77. For the Indians starters, Jimenez (0.82), Masterson (0.81) and Lowe (0.81) all cross the average reliability threshold. The remainder of the starting options fall short in the following order; Tomlin (0.71), Slowey (0.69), Huff (0.62), Gomez (0.52) and McAllister (0.17).
For pitchers projected to see between 40-80 innings, the average reliability is 0.42. Given the lower reliability standard for projected bullpen arms, the Indians actually look like a forecastable group, with essentially all the major likely figures exceeding this value.
Another way of visualizing the uncertainty of the Indians is to simply spend a little time browsing the projections available via fangraphs, which include in addition to Marcel the Zips and Steamer projection systems, as well as a Bill James forecast, a Rotochamp forecast, and a community-based player forecast. You'll see, relatively to most other teams, the Indians are a high-risk/high-reward group. It remains to be seen whether the Indians achieve that high-reward, but the Indians are possibly exploiting the system in the most cost-effective way (outside of having a big group of pre-FA stars) available to them in the hopes of getting a big return.
UPDATE: Tango links here from his place with a few additional comments. I've corrected the text above to reflect how he measures "reliability."