Making sense of Matt Carpenter's projection

Over the last three years, Matt Carpenter has been the most valuable Cardinal, accumulating just over 16 wins in that time frame, and never less than four wins. Which is why I've continued to be a little bothered by his projections. Steamer has him at 3.7 with Zips at 3.8, both lower than his worst performance over those three years. Either projection would of course represent a great season, but I couldn't shake the idea that he's better than that.

Usually, I trust the projections. Of course, regression is inevitable. Projections build that in. And age-related decline is a factor with Carpenter entering his age 30. Logically, I know all these things, but emotionally, Matt Carpenter is my favorite Cardinal. The projections are probably right, so I figured I'd look at the numbers to see how they see Marp going forward and probably be convinced that their right.

First, I wanted to look at Steamer and ZiPS projections compared to Matt's 2015 campaign and see what type of regression they expected:

Neither projection system believes in Carpenter's power outbreak last year. That makes sense, as the projections aren't only considering the previous year but his whole career. However, Carpenter's surge in power correlated with a conscious decision to hit more fly balls and be more aggressive with pitches he could drive. The projections might not pick up on that. The same can be said about his strikeouts. Carpenter struck out at the highest rate of his career last year, and both projection systems see him striking out less in 2016. However, if Carpenter continues the approach he used in 2015, he'll probably be closer to last year's than his rates prior to 2015. This is a fine trade off; Strikeouts are the least important part of a hitter's profile.

ZiPS doesn't directly project wRC+ or base-running value, so for overall hitting performance we have to use wOBA. So, entirely as a result of regression in power, both projections see Matt being more than a win worse than his performance last year. What the projections are seeing, at least is terms of how he produces value, much closer reflects Matt's career from 2013 on, when he became a full-time player:

The base-running and defense columns are pro-rated to 600 PA, so that all these stats have been adjusted to a rate basis. I think this chart shows us where the projections are coming from. The projections don't quite believe Marp's BABIP over the last three years, shaving 15 and 17 points off that. Considering he's entering his age 30 season, and the fact that you still have to regress BABIP a little even over three years, and the fact that the three year average is buoyed by a .359 BABIP back in 2013, that's not too disagreeable. But this is another spot where Carpenter could very well be more valuable than the projections let on.

Despite never posting a below average season in base running, Steamer sees him doing just that in 2016, and while Zips doesn't directly post base-running value, from eyeballing the stats it looks like Zips agrees, if a little less bearish than Steamer. Marp is entering his age 30 season, but that seems a bit harsh. It's just a matter of a couple runs, but every run counts.

Same goes for his defensive value. Over the last three years Marp has been just below an average defender, but both projections see him as a couple runs worse than that in 2016. It's understandable though, as he trended downward last year and is entering his age 30 season this year

After investigating the projections, I see where they're coming from. On offense, its BABIP regression from his three year average (but not all that different from his two year average) and ISO regression from 2015 (but similar to his three year average). Other than that they see age-related decline in his base-running and defense value.

However, I think he definitely has more chance of out-performing his projection than under-performing. I'd take the over on Carpenter's projection and feel pretty good about it.