As pitchers and catchers have finally reported, the starting pitching competition for the Cardinals has begun. If the season started now, Joe Kelly appears to be the odd man out. An injury or ineffectiveness could create a position in the rotation for Kelly, or, as happened last year when Jaime Garcia was injured, a spot may open at some point during the season. When given the opportunity last year, Kelly performed well. Many point to his sterling 2.28 ERA as a starter, as well as it's unsustainability. The narrative is clear. Kelly pitched over his head as a starter, and is a candidate for regression due to a .271 BABIP, and a 4.76 K/9 rate to go along with a 3.52 BB/9 rate. Kelly might be able to keep up his impressive results as a starter, but in order to do so he will have to improve his strikeout rate.
Out of 432 qualified seasons in the past five years, only 21 pitchers had K/9 rates lower than Kelly's 4.76 as a starter and only six of those pitchers had ERA's under four. Just three had a FIP under four (Mark Buehrle 2010, Joel Pineiro 2009, and Scott Diamond 2012). While Joe Kelly did not qualify last season, his FIP of 3.98 would have ranked 57th out of 81 pitchers last season. Considering Kelly's poor walk and strikeout numbers, it is pretty amazing that Kelly could approach the middle of the pack in FIP. He was able to do so because he limited home runs, allowing just five as a starter. Only four qualified pitchers had a ground ball percentage higher than Kelly's 53.2% last season. While ground balls become hits more often than fly balls, the damage tends to be limited. This, along with a likely unsustainable 6.6 HR/FB rate, allowed Kelly to keep his FIP low and his ERA even lower.
If Kelly pitches exactly like he did last year as a starter, he is likely to be below average. An uptick in BABIP and a few more homers could cause a highly inflated ERA and FIP. If Kelly is to progress as a starter and use his high ground ball rate effectively, he will need to improve somewhere else. He will need to strike out more hitters. He pitches 95 miles per hour even as a starter. While he does not have a clear out pitch, he has struck out more than eight hitters per nine innings in his career as a reliever. Getting his strikeout rate up, could make him at least an average starter if he is called upon during 2014.
Over the last ten years, there have been 28 players with at least fifteen appearances as a starter and 15 appearances as a reliever through their age-25 season like Joe Kelly. Of those pitchers, there are eleven with at least 7 strikeouts per nine innings as a reliever with at least one fewer strikeouts per nine innings as a starter. Here is the list of players along with their ERA and FIP in their different roles.
Of those players, only seven were given an opportunity to start during their age-26 to age-28 seasons. Here are their statistics as starters through age 25 and from age 26-28.
Results are mixed. Mitre, Bergmann, and Rogers all maintained their strikeout rates and remained mediocre. McClung improved his strikeout rate and became slightly less mediocre than he had been. Lynn maintained his rate, but he was already striking out enough hitters to post solid results. Ross, in sixteen starts last year, switched leagues and used his excellent slider to strike hitters out. Pre-injury Marcum is probably Kelly's best bet for a comparison. They have different pitching profiles. Kelly pitches at ten miles per hour faster than Marcum, but he does provide one example of a pitcher who was able to increase his strikeout rate as a starter closer to his reliever levels.
Figuring out who Joe Kelly is as a starter now is almost as difficult as determining who he can be. His high ground ball rate represses fly balls and home runs and could help make him an effective starter. He pitches with fantastic velocity, but has been unable to maintain a high strikeout rate as a starter. His strikeout rate went down as a starter, but not because he was pitching to contact, as his walk rate went up. Kelly has proven he can strike hitters out at the major league level. If he can merely split the difference between his starter and reliever strikeout numbers, he can be a very effective starter at the major league level. If he is unable to do so, he may be better off in the bullpen where his peripherals point to Kelly remaining an integral part of the Cardinals relief corps.