The Cardinals should trade for Jorge de la Rosa for bullpen relief

The debate about what to do at the trade deadline rages on. Generally, I have felt that the most realistic thing the Cardinals could do was add a reliever, and I wrote about the possibility of adding Jake McGee on Saturday. Others have felt similarly, with Joe covering the possibility of adding Andrew Miller, and John making a case for Fernando Salas.

The problem is, there just aren't that many quality candidates. There's quite a few bad teams, but those bad teams are really bad, and their bullpens are particularly bad. The relievers that have been good will have quite an interest, and a tendency for relievers to be valued higher in MLB trade than public valuation could get out of hand. For my last article, I claimed that the Cardinals would probably look to buy-low on someone they think will bounce-back. For that, I created a list of every pitcher on every team that should definitely sell: The Braves, Phillies, Reds, Brewers, Rockies, Padres, Diamondbacks, Athletics, Angels, Twins, Rays. More teams will likely join that list, but for now, those are the selling teams.

Anyway, I looked at every reliever on those 11 teams, taking their current FIP and xFIP and their Rest of Season (R.O.S.) projected FIP. From there, I could identify what players are likely to bounce-back according to the projections. There weren't that many good candidates. Chris Capuano, for instance, had the biggest difference between his FIP and projected FIP, but his projected FIP was still 4.13. So he'll probably bounce-back, but what he'll bounce-back to still isn't all that valuable. To John's credit, Salas did very well by this method. It was also the inspiration for me to look into Jake McGee, although that ended with disappointment. Brandon Maurer was also a bounce-back candidate, but was already traded to the Marlins.

But wait, Brandon Maurer! Another starting pitcher who, after having a stretch of underwhelming results, reinvented himself by moving to the bullpen. It's a lot easier to pitch out of the pen than out of the rotation: despite relievers overall being the weaker group of pitchers, the league average reliever has an FIP 25 points lower than the average starter, 4.32 to 4.07. In general, that's explained by two things: (1) relievers are able to throw harder since they have shorter appearances than a starter and (2) relievers don't make it more than once around the order, so hitters don't get to see them multiple times.

So, I had a new project. I set out to measure the advantage a reliever has by essence of being a reliever. I looked at seasons within the Pitch f/x era (which began in 2007) for seasons in which pitchers threw both 20 inning out of the bullpen and 40 innings as a starter.

That got me 144 seasons. That wasn't quite enough though, so I did back-to-back seasons as well. Any time a pitcher threw 20 relief innings one year, and 40 starter innings in the year before or after, I added into my data set. But I didn't want to double count any performances, so if I already had a single season from that player that covered one of those years, I didn't count the year to year match.

That brought my sample size to 276 instances of a player pitching both out of the rotation and bullpen in a short time frame. With that list put together, I took each player's differential at a bunch of different pitching stats, to see the average difference pitchers had between the rotation and bullpen. Here is the average advantage when pitching out of the bullpen:

When pitching out of relief, a pitcher becomes better at almost everything. I was surprised that BB% increased, but with a 3.4% increase in K%, it seems relievers become more effectively wild. They do better on batted ball stats as well, getting more grounders, less homers per fly ball, and more infield flys as a percentage of fly balls.

The velocity increase holds up as well, with pitchers on average adding about a mph to their fastballs. Hitters become more aggressive as a whole when a pitcher moves to relief, and they make less contact. Zone% remains nearly unchanged.That 65 points of FIP though, is the number I'm looking for. That's the amount that, on average, pitchers gain in run prevention ability when coming out of relief rather than starting. And that finally brings us to the subject of today's post: Jorge de la Rosa.

de la Rosa combines both Saturday's premise and today's: a bounce-back candidate, and a struggling starter that could reinvent himself as a reliever. He has had an ugly season so far that already got him booted from a bad rotation temporarily, and even after righting the ship a bit he has a 2016 FIP of 5.85.

Like McGee, de la Rosa pitches in the harshest environment to pitch in, Coors field. Thus that 5.85 comes out to a FIP- of 120, meaning his FIP is 20% higher than average. He's been better by xFIP, with a 4.96 xFIP (109 xFIP-) and the projections like him even better than that, at 4.43, which is lower than average in his environment. That's also up from his preseason projection of 4.21, so it's not like the projections are just being stubborn. They see a skill change in the wrong direction, and still think he'll be much better going forward.

There's also a lot to like about his contract, in that it's bad, so it won't demand a lot in return. He's in the final year of a $25M/2 year extension with the Rockies, and is making $12.5M this year. On August 1st, with a third of the season left, he'll be owed about $4.2M. With de aa Rosa's performance by FIP putting him at an incredibly marginal level over replacement level (0.1 fWAR), and his ERA even higher than his FIP, it certainly doesn't appear like there will be much demand for him this month.

According to MLBTR, the Royals have kicked the tires on de la Rosa, but they've also been connected to quite a lot of pitchers according to the report. Between his cost and bad performance on the year, the Cardinals could acquire him for little in addition to picking up the $4M+ remaining on his 2016 salary.

I headed over to Brooks Baseball to see how his stuff looked. His velocity is down but only by a bit, and his movement on his pitches is normal. His contact% has decreased a little more than the league as a whole has, so he's a little harder to make contact against than last year. His split-finger and curve is getting more whiffs/swing than last year, his cutter less. Here's a little tidbit that may be evidence his possible success as a reliever. Look at this breakdown of his whiffs/swing and pitch usage for each time through the order in 2016:

His stats the first time through the order may be fairly indicative of how he pitches in relief: about a 45/35 fastball-splitter combo, while also mixing in cutters and curves, more so against lefties. That is a lot of whiffs the first time through the order, and that doesn't even take into account the extra mph he'll likely gain on his fastball that should make each pitch a bit more effective. The sinker seems likely to be dropped entirely from a move to relief.

How can we project de la Rosa's move from Colorado's rotation to St. Louis' Bullpen? We'll start with the Coors affect. According to FanGraphs park factors, Coors has a 108 FIP park factor, meaning pitching for the Rockies will, on average, raise your FIP 8% (these numbers are already halved for playing half of your games on the road). So I took his FIP, xFIP, and projected FIP (which I'll call pFIP), and divided each by 1.08, to bring it down to park neutral.

Then, to account for moving to the bullpen, I deducted 0.65 points off of each. Notice that I'm not adjusting his numbers to playing for the Cardinals, which has a 99 FIP park factor. We'll leave that off for a little margin for error in our direction. Here's how those numbers compare to the Cardinals' bullpen, and it's likely makeup if the Cardinals added him:

Even basing it on FIP, where he comes out worse, he still comes out better than fellow lefties Siegrist and Lyons. The upside is that he could end up in a position as one of the Cardinals' primary set-up men. And he could do so very cheaply compared to other relievers. As a declining starter, de la Rosa isn't thought of as a relief option, he's just thought of as a bottom of the barrel starting pitching option. I mentioned earlier a list of relief pitchers on selling teams, here's the top 10 of that list, in terms of projected FIP, with de la Rosa added in:

He would place ninth with the projected FIP as a reliever I've calculated here, which probably doesn't sound all that good. Really he should be tenth, because teammate Jake McGee's FIP would be lowered by more than 4 points if he was traded away from Coors.

However, there's no reason to expect that all the players shown here will be traded. Liam Hendriks, Ian Krol, Xavier Cedeno, Cam Bedrosian, and Jeremy Jeffress are all still pre-arbitration players and thus are making very little and controlled for a while, so their respective teams don't necessarily have to move them. The A's liked Sean Doolittle enough to give him a long-term extension, so it's not a given he'll be shopped hard either.

The most obvious trade chips on the list above are Boone Logan, Arodys Vizcaino, and Jake McGee, with Jeffress being the only likely pre-arb guy getting moved. That's a pretty pitiful top of the market as far as the buyers should be concerned, but they still might command something a little too similar to top of market prices. de la Rosa offers an option out of that mess.

The Yankees could add Andrew Miller and Aroldis Chapman right to the top of this list of course. Even though the Cardinals would be long-shots to add Miller or Chapman, their availability would help drive down prices on the relievers the Cardinals would realistically add, and would lower the need to experiment by turning a starter into a reliever. However, it's also still not the most likely scenario. The Yankees are probably only going to sell if it's really bleak for them at the deadline.

It appears that below average starting pitchers are just waiting to be above-average relievers.. Being that above-average relievers typically get better returns in trade than below average starters, it certainly looks like a market inefficiency waiting to be exploited. Jorge's handedness, current performance and remaining contract all make him a good fit for the Cardinals to try out this type of conversion. Trading for a struggling starter and making him a reliever isn't a very likely move, but I think the analysis here shows it's a compelling alternative to dealing with the weak reliever market.