clock menu more-arrow no yes mobile

Filed under:

Shelby Miller is not a Cy Young candidate

Shelby Miller is off to a great start so far in 2015, but his peripherals suggest that he is due for regression. In addition, he is already behind NL leaders in important metrics like fWAR and FIP, which are being utilized more often now by Cy Young voters.

Brett Davis-USA TODAY Sports

We are now over a quarter of the way through the 2015 season, and for some people, this means that we are far enough into the season to start speculating on awards. One narrative that I've been seeing a lot of over the last week or so is that Shelby Miller is the early favorite for the Cy Young Award in the National League. Doing a quick Google search, I was able to find no fewer than six articles from reputable sources which made this very claim.

Most people can agree that Miller is off to a great start. He has made nine starts and pitched 60 innings with a 1.50 ERA and a 2.8 RA9-WAR, all near the top of the NL leaderboard. As you can probably guess, I picked these specific stats because they make Shelby Miller look especially good. In no way are they a complete representation of what Miller has done so far this year.

Likewise, the people who are calling Miller a Cy Young candidate are cherry-picking certain stats, mainly his 1.50 ERA and his 5-1 win-loss record. (I hope you are not scarred for life now that I have mentioned a pitcher's win-loss record on this site.) The only issue with this approach is that Cy Young voters are much smarter than they used to be; they almost never vote strictly based on win-loss record and ERA anymore.

There are numerous examples of this, such as in 2009 when Tim Lincecum won the Cy Young Award over Adam Wainwright and Chris Carpenter despite having fewer wins than either pitcher (15 compared to Carpenter's 17 and Wainwright's 19) and a higher ERA than Carpenter (2.48 vs. 2.24). What Lincecum did have, though, was the best fWAR (nearly two wins more than Carpenter and Wainwright), FIP, and strikeout rate of all pitchers in the National League.

In fact, NL Cy Young voters have not made any obvious mistakes since 2004, when Roger Clemens won the award over four other pitchers who were probably more deserving. A case can be made that Clayton Kershaw and R.A. Dickey didn't deserve their respective 2011 and 2012 Cy Young Awards, but in each case, the perhaps more deserving runner-up (Halladay in 2011, Kershaw in 2012) had won the award the previous year, meaning that voters might have simply wanted to give the award to a different pitcher. Even if we say that voters made the wrong choice in 2011 and 2012, the winners were still near the very top of the leaderboard in metrics like fWAR and FIP.

So we know that Shelby Miller has a sparkling ERA, win-loss record, and a near no-hitter so far in 2015. But where does he rank among NL pitchers in other categories?

Player IP K% BB% K-BB% HR/9 ERA FIP fWAR
Shelby Miller 60.0 20.7% 7.8% 12.9% 0.45 1.50 3.25 1.1
NL Rank 9 25 36 26 11 3 15 14

As you can see, Miller isn't in the top ten in any categories besides ERA and innings pitched. So far, he hasn't struck out a lot of batters, and his walk rate is slightly above the league average (7.7%). His home run rate is low, partly due to his 50.3% ground ball rate, the highest of his young career. However, Miller also has a 5.8% HR/FB rate, suggesting that his early home run suppression may not be sustainable.

By FIP and fWAR, Miller has been one of the better pitchers in the National League, but he is still well behind the top pitchers in the league. Since 2000, no Cy Young Award winner (in either league) has finished outside the top ten in fWAR, and only one (Eric Gagne in 2003) has finished outside the top five. Miller will have to pitch even better than he has so far this year in order to be within striking distance of the league leaders in FIP and fWAR.

The chances of that happening, though, appear to be slim. In addition to having an unsustainable HR/FB rate, Miller has the lowest BABIP (.206) and third lowest strand rate (87.1 percent) of all National League pitchers. Given these factors, it is easy to see why Miller's xFIP is around league average at 3.79 (an xFIP- of 99). Indeed, his rest of season ZiPS projections seem to reflect the fact that he is due for some regression.

Shelby Miller 23 131.0 8.15 3.23 0.98 0.288 75.8% 3.45 3.91 1.4

While it's easy to say that Shelby Miller's great start is the result of some incredible luck, it can be difficult to say precisely how much is luck and how much is skill. While DIPS theory generally holds true for most pitchers, there are a small number of pitchers who consistently outperform their FIP for a variety of reasons. For example, left-handed pitchers outperform their FIP more often than right-handed pitchers, perhaps due to their ability to control runners, as John Choiniere pointed out on this site earlier this month.

There are also pitchers, Johnny Cueto being the most notable example, who consistently outperform their FIP by maintaining an extremely low BABIP. Sometimes this can be the result of being backed up by a good defense for a number of years, but there is also some evidence that certain pitchers are more skilled than others at inducing weak contact.

I bring this up because Shelby Miller has significantly outperformed his FIP in each of his first three seasons. In 430 career innings, Miller has a 3.08 ERA and a 3.90 FIP, in large part due to a low career BABIP of .259. Defense doesn't appear to be a big factor; he has outperformed his FIP for a good defensive team (2014 Cardinals: 5th in team defense), a bad defensive team (2013 Cardinals: 26th in team defense), and an average defensive team (2015 Braves: 13th in team defense). While it is probably still too early to say that Miller is one of the rare pitchers who is skilled at outperforming his FIP, it is something to keep an eye on.

For fun, I thought I'd do a quick linear regression in Excel to predict pitcher BABIP using the batted ball type and quality of contact data found on Fangraphs. Using only 2015 data, I ended up with an R-squared of .433, which isn't too terrible given the context. Using the coefficients calculated in Excel, I plugged in Miller's batted ball data and came up with an expected BABIP of .267, which is significantly lower than the league average BABIP of .293.

Looking at Shelby Miller's batted ball numbers, it isn't too hard to see why his BABIP is below league average.

Player LD% GB% FB% IFFB% Soft% Med% Hard%
Shelby Miller 17.0% 50.3% 32.7% 9.6% 23.3% 47.9% 28.8%
MLB Average 21.1% 45.4% 35.5% 9.8% 18.8% 52.8% 28.4%

The two numbers that stand out are Miller's line drive rate and soft contact rate. Based on the data so far this season, these two variables appear to have the highest correlation with BABIP, and they are where Miller is most clearly doing better than league average.

In fact, Miller is doing drastically better than in previous years, as he has increased his soft contact rate by seven percentage points and decreased his hard contact rate by six percentage points from last season to this season. It remains to be seen how sustainable these quality of contact statistics are over the course of the season. Given Miller's new approach, especially the increased use of his cutter, there may be reason to believe that his changes in batted ball data are for real.

With that being said, Miller is still due for some regression. Even if we say that he is better than most pitchers at generating weak contact, he will not be able to maintain a .206 BABIP. Ultimately, he will never be a dominant, ace-quality pitcher unless he is able to increase his strikeout rate and decrease his walk rate. He can certainly be a quality, major-league starter for a long time if he continues his current approach, but he still has to make significant improvements before we can truly say that he is a Cy Young contender.

. . .

Nick Lampe is a contributor at Beyond the Box Score and Viva el Birdos. You can follow him on Twitter at @NickLampe1.