Mets 2015 preview: Can the Mets be better than the Marlins?

Deciphering which team is better prepared for the 2015 season.

The landscape of the National League East is vastly different going into 2015 than it was in 2014. The Phillies are finally rebuilding and will likely be even worse than they were last year. The Braves look to be rebuilding, as well. Meanwhile, the Nationals are still strong, and the Marlins have made a flurry of moves to improve themselves for the coming season. The Mets' offseason additions have been limited to Michael Cuddyer, John Mayberry Jr., and Sean Gilmartin, hardly big-name players and likely much less than what most Mets fans were hoping for. Still, it seems like the Mets and Marlins will be fighting it out for second place, which raises an important question: Which team is better?

To answer this, we'll need to project out both teams, understanding than any projections—particularly for teams consisting of so many young players—are imperfect. Then we must consider our own biases. At the start of this exercise, I thought the Mets were a better team, with a stronger rotation that would more than make up for the deficiencies in the lineup. Our projections will tell us if this theory is correct.

First, let's look at position players. Below is a table of the Steamer projections for the regular position players on both teams.

The only bench player included is Mayberry, as he's the only player who projects to see significant time in a platoon role on either team. As of right now, the Mets' bench looks to be deeper than the Marlins', particularly in the outfield department, but neither bench is finalized yet, making any calculations premature.

These projections have several issues on both ends that I'd like to correct. This is where the process gets subjective, but it is necessary to add information that Steamer does not consider. Here's a run-down, position by position on each side, starting with the Mets:

• The d'Arnaud projection looks accurate to me. Even after his putrid offensive output in the first half last season, he put up 1.6 fWAR. His total offensive output should be a good deal higher, and his defense should also improve following offseason surgery to remove bone spurs in his elbow.
• Steamer does a poor job of projecting Lucas Duda. If used correctly, Duda will hardly ever see lefties, which will limit him to about 500 at-bats. This will also push his wRC+ closer to his career 158 mark against righties. Given this, I feel comfortable projecting him for a 2.5 WAR season, similar to the mark he posted last year.
• Murphy is projected to regress offensively and post worse defensive metrics. I think Steamer is a tad bearish, but to be conservative we'll say the 1.5 WAR mark is accurate, given Murphy's poor profile (which I outlined here).
• Wright gets a rebound. The Captain has a couple of All-Star level seasons left in him.
• Flores is getting too much love from Steamer on the defensive end. He posted strong numbers in a short sample last season, but until he does so over a long period of time, we can't reasonably project him to be a plus defensively. Drop him down to 1.0 WAR.
• Granderson's, Lagares's, and Cuddyer's projections all seem conservative, but reasonably accurate.
• Lastly, Mayberry's Steamer outlook is far too gloomy. With about 200 at-bats against lefties—against whom he has posted a career 130 wRC+—I see no way in which he posts a mere 92 wRC+. As such, I would bump his total value up to 0.5 WAR.

Once we have adjusted Steamer's projections to better reflect real-world player usage, we arrive at an adjusted total WAR of 17.3.

Now, for the Marlins:

• Saltalamacchia is what he is at this point. Outside of his walk year with the Red Sox, he has never posted more than 2.0 WAR in a season. His projected numbers are essentially identical to what they were in 2014.
• Steamer has not considered the fact that Morse will now play first base, where his career defensive marks are middling rather than atrocious. We can comfortably add 0.5 WAR to his projection.
• Dee Gordon's regression started in the second half of 2014, and I don't see anything to indicate that it won't continue. I think Steamer has hit the nail on the head with this regression candidate.
• Prado keeps on providing slightly above-average offense and solid defense. Nothing to adjust here.
• Hechavarria is a defensive enigma (regarded highly by execs and commentators, disliked by the metrics) who posts poor offensive numbers. Steamer expects his offense to regress a tad, but also for his defensive marks to be higher than at any other point in his career. Ultimately, I think the final number is pretty close to what Hechavarria will deliver, though one could make a case that he's a young player with room for improvement.
• Yelich is projected to regress offensively and defensively, and there's some merit to Steamer's numbers here, given his high BABIP. Still, Yelich has always produced high BABIPs, and I think reaching the 4.0-WAR plateau is likely again.
• Ozuna's and Stanton's numbers look accurate.

Ultimately, we wind up adding 1.2 WAR to the Marlins' total to bring it to 19.6, which is 2.3 more wins than the Mets'. This is likely not a surprise to most Mets fans, as the Marlins have one of the best young outfields in baseball and have made several additions in the infield. The gap is not huge, and I think the Mets' deeper bench will help make up the gap a tad; nonetheless, the Marlins have an edge here.

Unfortunately, this approach will fail horribly when looking at pitchers. Below are the Steamer projections for Mets and Marlins pitchers in 2015:

Clearly, these numbers are wildly inaccurate. The Mets' rotation, full of young former top prospects and a returning ace, is projected to accrue less WAR in 2015 than it did in 2014. Similar problems exist on the Marlins' end of things. The problem here is twofold: firstly, Steamer often fails miserably in projecting young talent due to its use of MLEs (major league equivalents) to translate minor league stats to the majors. Secondly, fWAR's fundamental flaws hurt the Mets and benefit the Marlins (more on this here). We might be able to get a little closer by using RA9-WAR projections, but those are not provided and would still rely on MLEs.

Therefore, our best option is to make our own projections, working off of RA9 numbers from Baseball Reference. This process is highly dependent much more on stats than on scouting, and will likely introduce a good deal of bias into the process. To counteract this, I will include bear, base, and bull projections for all pitchers—both Mets and Marlins—in an attempt to get a better grasp on how the rotations stack up.

Here are the Mets' projections, followed by justification:

*Harvey's stats come from 2013, his last season in the majors.

• Matt Harvey was a legitimate monster in 2013, posting 6.0 WAR in 178.1 innings. His value in 2014 is very much tied to the number of innings he pitches. As a conservative base case, I chose 150 innings (assuming he's limited in the hopes of a playoff berth) with variations off of that for bear and bull cases. I've assumed that he can be 70% of himself in his first year back from Tommy John surgery in the base case, since it often takes pitchers a while to get back to full strength.
• Zack Wheeler did essentially what we expected last season: a decent amount of strikeouts, but too many walks. However, he was able to improve his wOBA from .312 in the first half to .299 in the second half, striking out a higher percentage of batters while walking fewer and giving up fewer hits. Furthermore, recent work at FanGraphs indicated that he has one of the best pitch arsenals in baseball. There's certainly a lot of room for optimism. As such, I've projected a modest improvement to two wins in 2015 as our base case, and then scaled that WAR/IP rate in each direction for the bear and bull cases.
• Jacob deGrom's fantastic performance was backed up by his peripherals, and he also appeared on the arsenal score list mentioned above. Still, the safe assumption is that he's not quite that good. Assuming he takes a small step back but is able to increase his innings total, deGrom still offers a great deal of value for the Mets in 2015. The bear case assumes that he regresses further, while the bull case assumes that he's able to maintain his performance from his Rookie of the Year campaign.
• Jonathon Niese essentially is what he is at this point. The ace-caliber breakout Mets fans hoped to see for many years never came, and Niese has settled in as a solid mid-rotation starter. He had a solid 2014, but the injury risk is real and he showed a drop in fastball velocity for the third consecutive year. Hopefully, Niese can stave off the shoulder explosion that seems inevitable.
• Things are a bit murkier with Bartolo Colon, given both his spot on the aging curve and the chance he gets traded midseason. My base projection gives him 125 innings before he is shipped out to make room for Syndergaard. I don't expect him to be much more than a replacement-level innings eater in any case, but there is a chance that age finally catches up with Colon and makes the innings he provides below replacement level.
• For the third consecutive season, Mets fans will likely be excited by the arrival of a new young pitcher when Noah Syndergaard inevitably takes his place in the major league rotation. I expect him to deliver 100 solid big league innings, 25 coming from spot starts for Harvey and the remainder coming in the latter half of the season as a regular rotation member. To play on the conservative side, we'll say his debut is more Zack Wheeler (though I like Syndergaard more) than Jacob deGrom for our base case, project rookie struggles (with fewer innings) in the bear case, and allow for a glimpse of the potential upside in the bull case.
• Rafael Montero provides some solid depth for this group. I don't see him getting more than 50 innings barring a major injury (a distinct possibility with pitchers), and to be conservative I'll say he provides no excess value in the bull case.

To visualize the probability distribution of the Mets' rotation WAR, I did some graphical simulations (if anyone is interested in how, I'll do my best to explain in the comments). Below is a good visualization of where we should expect the Mets' rotation WAR to fall.

Here are the Marlins' projections, followed by justification.

*career statistics used due to sample size

• Jose Fernandez is expected to return from Tommy John surgery after the All-Star break, and thus will not be able to log much more than 100 innings in 2015. To be conservative, I've assigned him 80 innings, and then scaled his career WAR rate by a factor slightly lower than the 0.7 I used for Harvey (because Fernandez will have less time to recover over the span of his innings). The bear case is scaled even lower with fewer innings, while the bull case assumes he's able to just about maintain his career performance rates.
• Henderson Alvarez's profile screams regression. His LOB% was nearly 80% in 2014, and he didn't even strike out six per nine innings. His FIP and xFIP marks sat in the mid-threes, far above the shiny 2.65 ERA he posted (though still slightly above average). As outlined here, very few pitchers with Alvarez's skill set maintain this sort of ERA. One such pitcher is Greg Maddux, and no one is Greg Maddux. Since we are using RA9-WAR, Alvarez appears to be an ace, but I feel confident saying that he'll struggle to post even three WAR next season. Since we want to be conservative, we'll make that Alvarez's base case. His bear case includes more regression, while his bull case assumes he can maintain last year's performance.
• Mat Latos saw his fastball velocity continue to decline, dropping 1.8 mph to a career-low 90.7. Unsurprisingly, his strikeout rate fell as well, coming in at 37.7%. While the spacious Marlins Park should help mitigate the latter problem, neither of these trends is positive, and we haven't even mentioned the litany of injuries he suffered in 2014. As Jeff Sullivan pointed out here, Latos is an extremely volatile asset, and there are a ton of warning signs. Again, to be conservative, I'll say that he can maintain his 2014 rate of production, allowing for a rebound to career norms in the bull case and projecting an injury-plagued, continued decline in the bear case.
• Jarred Cosart is another starter who put up a solid performance in 2014 but has plenty of warning signs. He too does not strike many batters out, and unlike Alvarez, his control is mediocre at best, resulting in a high WHIP. There's still some upside here: Cosart displayed stronger strikeout rates in the minors, and is a former top prospect. However, his major league performance to date—specifically the continued control problems—makes me rather bearish on adding much improvement in his base case.
• A former first-round pick, Tom Koehler finally found some degree of major league success in 2014. However, it's tough to see much more upside in his profile given his unspectacular stuff and mediocre minor league numbers. A repeat two-WAR season seems by far the most likely outcome, with some limited upside available. His projection comes out identical to Cosart's.
• Brad Hand should probably not be a starting pitcher given his extreme platoon split. He allows a .347 wOBA to righties versus a .266 wOBA to lefties. His performance as a starter was not actually much better than that as a reliever, so it might be that Hand is just not good period, or should be limited entirely to lefty batters coming out of the bullpen. We're trying to be conservative, so I'll give him a zero WAR projection for his time in the rotation. I imagine he'll shift to the bullpen when Fernandez returns, though that will be contingent on Haren not retiring.
• Dan Haren's innings are tough to project due to his previous statement in regards to retiring if not playing on a West Coast team. Arguably, Mets fans should hope he chooses to play out his contract. Haren hasn't been an above-average starter for a couple of years now, and the outlook doesn't appear very rosy for 2015 either.

Similar to above, here is a visualization of Marlins' rotation WAR:

And here's an overlay of the two distributions:

The rotations finish in a dead heat. Using the base case of our projections, the Marlins will accrue 0.05 more WAR than will the Mets. If we extract this number by sampling from the probability distribution, we get something closer to 0.1 WAR. Regardless, the difference is fairly miniscule.

Something this analysis neglects is the superior depth of the Mets' rotation. I've entirely neglected Dillon Gee, who is being forced out of the team's rotation by excessive depth. On the farm, Steven Matz is approaching major league readiness, and other more fringey arms such as Matt Bowman (who was recently compared to Brandon Webb) are also available in worse-case scenarios. Meanwhile, the Marlins' minor league depth is much more limited: Justin Nicolino is their best remaining pitching prospect, and I remain skeptical of his ability to contribute at the major league level while striking out a paltry 4.28 batters per nine innings in Double-A this past season. Ultimately, I think this depth will almost entirely eliminate the small gap that we've seen from the rotation projections.

Lastly, the bullpens, evaluated using the same method as above. Keep in mind that these projections are much less useful given bullpens' general instability:

• Familia had a very nice 2014, but still walks too many and struggles against lefties. Unless he corrects these problems (and I see no evidence he will), he's unlikely to improve on his 2014 performance. In the bull case, he sees some improvement, while in the bear case, these problems catch up to him and Familia regresses.
• Mejia improved significantly after moving to the bullpen, and his numbers as a starter deflate his value to a degree. I think Mejia will continue to be a solid reliever and ultimately post WAR numbers essentially identical to those of Familia.
• As much as we like Vic Black, he walked nearly five batters per nine innings last season. That alone is an extremely frightening statistic, and he'll have to do a lot of work to overcome the control problems that have plagued him throughout his career; until he does, I think Black is in for some regression.
• Carlos Torres somehow still has an arm attached to his right shoulder. Assuming he survives another year of abuse from Terry, he should offer the same steady performance. The upside and downside are fairly limited.
• Edgin was extremely effective in his LOOGY role down the stretch for the Mets, and his peripherals back up the strong performance. Assume some regression in the base case with upside remaining if he can indeed maintain his high WAR accumulation rate from 2014.
• Parnell had a very solid 2013, but he's suffered two consecutive serious injuries, and assuming he can come back just as strong is overly optimistic. Furthermore, Parnell will likely not be ready for the start of the season given the date of his Tommy John surgery, so we'll have to limit his innings as well.
• I am actually quite high on Sean Gilmartin. His minor league dominance against lefties could make him an extremely effective weapon in 2015, if used correctly. I may be a tad bullish, but given the small values, the induced error is minimal.

And now for the Marlins:

• Cishek's WAR probably understates his performance in 2014. He increased his strikeout rate to 11.57 K/9, a new career high, while giving up an inflated .331 BABIP (though he did allow line drives at a career-high rate). Despite this, SIERA still thinks he was unlucky, and I think it's fair to project a slight rebound in the base case. There's not a whole lot more upside unless he turns into Mariano Rivera, so the bull case is fairly limited.
• A.J. Ramos put up a shiny ERA and lots of strikeouts in 2014, but his walk problems are even worse than Black's. Giving up six free passes every nine innings is simply not a formula for success. This, combined with the absurdly low HR/FB mark, screams regression for Ramos.
• Mike Dunn's ERA is in line with his FIP and xFIP, and none of the other typical flags for good or bad luck stand out. His 2014 performance is probably a pretty good estimate of what he'll do going forward.
• Bryan Morris is another Marlins reliever who put up a nice ERA that is not at all backed by his peripherals. His 4.82 FIP is more than two points higher than his 1.82 ERA, and other metrics like xFIP and SIERA aren't very optimistic either. Given this, it seems unlikely that he'll be able to replicate his high WAR mark from last season, even in the rosiest of scenarios.
• Crow hasn't been a good reliever since 2012, and things got even worse in 2014 with a precipitous drop in his strikeout rate. The Marlins are supposedly entertaining ideas about making him a starter, which seems to be an even more disastrous proposition. Crow is little more than a mediocre middle reliever.
• Acquired for Logan Morrison, Carter Capps had some very poor luck in 2014, compiling a 3.98 ERA against a 2.35 FIP, largely due to an inflated .340 BABIP. Assuming his luck stabilizes in a larger sample size, Capps should provide some decent value in the Marlins' bullpen, with his elite fastball velocity offering some further upside.
• Chris Hatcher seems due for some improvement in 2015 given his strong peripherals in 2014. However, he is a 29 year old who saw his first extended taste of big league action last season, so it's tough to get any more than a little optimistic.

Again, here's the probability distribution for the Marlins, followed by an overlay of the two:

Here, the Mets project to have an advantage of 0.6 WAR by our base case and 0.74 WAR by the average difference in the distributions. This is the most volatile portion of our projections, so marking it down as the only clear win for the Mets is a tad concerning.

Summing up the three portions of our projections, we find the Marlins ahead by 1.75 WAR. With all the random variation that occurs throughout the regular season, that gap is almost meaningless. Still, it seems that the Mets need to upgrade something in order to achieve the goal of true contention that they hope for in 2015.