With Jeremy Hellickson, R.A. Dickey and Bartolo Colon off the board this early in free agency, an already bare cupboard has become barren. If you want a starting pitcher on the market now, you're looking at about five options, and only two of them managed to be average Major League Baseball players last year.
Suddenly, the second-best starter behind Rich Hill is Ivan Nova, the same Ivan Nova who spent 2014 and 2015 alternating between hurt and terrible, the same Nova who lost his spot in the Yankees' starting rotation last year.
Despite his flaws, the FanGraphs crowd projected that Nova will sign a three-year, $43 million deal. Dave Cameron upped that to four years and $56 million but admitted, "In a year, it's either going to look like genius or foolishness, with the in between options looking less likely."
Here's a bet that it will look foolish.
He didn't change fundamentally in Pittsburgh
There's a halo around Pittsburgh pitching coach Ray Searage. He has taken in many a struggling pitcher and turned him into gold, or so the story goes. The problem is that some of the stories ascribed to him are more myth than fact.
J.A. Happ, for example, was already turning to the fastball more often before he got to Pittsburgh, and he told me that the only thing he worked on with Searage was being direct to the plate. Joe Blanton, who has been a revelation in the bullpen after pitching 30-plus innings with Searage, benefitted as much from moving to the pen as from the tutelage. Neftali Feliz was lights-out with Texas before, so maybe he was just healthy with the Pirates.
Then there are the players who were about as good in Pittsburgh as they were previously. Juan Nicasio, Ryan Vogelsong and Jon Niese are the most recent examples of guys who remained the same. Add Ivan Nova, despite the cosmetic changes to his outcomes.
Take, for example, Nova's pitching mix. This doesn't look like radical change, does it?
He traded a few fastballs for non-fastballs, but if we're talking about a 100-pitch start, we're replacing three or four fastballs with three curves and a changeup. That doesn't seem like radical change.
If you're noticing the velocity increase, that's natural. Velocity peaks in August, probably because of fitness and weather. Other than a slight change to the release point on the curveball, there was no significant difference in Nova's movement on his pitches after he moved to the Pirates.
The ways he did change are not likely to stick
You don't go from an ERA over five to one right at three without changing something. Nova did change something; with the Yankees, he walked just over two batters per nine, right in line with his career rates, and with the Pirates, that dropped to less than half a walk per nine innings. He also cut his home run rate from nearly two homers a game to basically half a homer a game.
If you can't strike out more batters -- and Nova didn't -- you can try to control your walk and home run totals. The problem is that the swing Nova underwent was so drastic that it seems unsustainable.
It's more obvious with the home-run rate. Nova has given up 1.04 homers per nine innings for his career, on a 13 percent home-run-per-fly-ball rate. In New York this year, those numbers surged to 1.76 and 21 percent respectively, which we would point to as bad luck. In Pittsburgh: 0.56 and 8 percent.
The home run per fly ball rate in the league this year was 12 percent, and that's important. Pitchers suffer major swings in that rate from year to year -- the year-to-year stickiness of the number is about the same as the stickiness in wins totals, which isn't good -- and so you normally want to use the league number when projecting a pitcher.
There's little evidence that the process on the fly balls he allowed was any different. With New York, Nova allowed an exit velocity of 92.0 mph on balls with a 10-degree or more launch angle. With Pittsburgh: 91.7 mph. Batters hit balls in the air with about the same authority.
When it comes to walks, you can see what Nova changed to improve the walk rate. He showed the best first-pitch strike rate of his career (66.9 percent vs. 58.6 percent) and was in the zone more often (48.1 percent vs. 42.8 percent). First-pitch strike rate is strongly tied to walk rate, so that's great news. But first-pitch strike rate is sticky year to year -- basically about the same as fly ball and walk rates in that regard -- and pitchers tend to stay within a narrow band over the course of their careers.
Also, take a look at the called strike rates for Nova with the Yankees and the Pirates. Every three batters, Nova got an extra strike call with the Pirates that he didn't get with the Yankees. That's a strike call that might have turned a 2-1 count into a 1-2 count and one he can thank his catcher for. Francisco Cervelli was the seventh-best catcher in framing runs on Baseball Prospectus. What if Nova's new catcher isn't as good at stealing strikes?
The resulting walk rate? Nova walked 1.1 percent of the batters he saw in Pittsburgh. Nobody has done that over the course of a season this century. Carlos Silva in 2005 walked 1.2 percent, and he regressed to his roughly 4 percent career rate the next season. Phil Hughes had a 1.9 percent rate in 2014, then gradually returned to his roughly 5 percent walk rate over the course of the past two seasons. That's the full list of sub-2 percent walk rates.
Nova should return to his customary walk and home run rates next year.
Nova doesn't have the tools to be a long-term successful starter
Here's the biggest problem with signing Nova to a big-money deal: He has to be a starter to be worth that money because he hasn't proven himself as a late-inning option with the big fastball.
We know adding multiple pitches helps a pitcher with the third time through the order in particular. We know the job of a starter is to eat innings, which takes good health (and multiple pitches), and nothing predicts future injury like past injury.
Nova fails on both accounts. He has had major triceps and rotator cuff problems, and he underwent Tommy John surgery in 2015, which means he's a little more risky than average when it comes to health.
But Nova also threw either a fastball or a curveball 93 percent of the time last season. That's rare! Only 45 of 822 qualified pitcher seasons in the pitch-tracking era featured two or fewer pitches thrown more than 5 percent of the time. Of those 45, nine were from knuckleballers. Let's take those out.
If you're betting on Nova to keep doing this, you're betting on him to remain in the 4 percent of pitchers who have managed to qualify for the ERA title while throwing two pitches. You're betting on his being A.J. Burnett or Justin Masterson without their track records.
That bet probably shouldn't cost $50 million.