Matt Harvey and multiverse theory

Matt Harvey has been on a lot of people's minds and radio talk shows in the last few weeks. The 2015 Mets find themselves in the surprising situation of having to plan for October baseball and with Harvey less than two years removed from Tommy John surgery, there are some interesting debates to be had: how do teams best keep pitchers safe, how do we evaluate present wins versus future value, how do we optimize the delicate balance players must strike between themselves and their team?

Unfortunately, the discussion that resulted had little to do with those questions. Harvey first seemed to float a Scott Boras-shaped test balloon, with the super-agent declaring his client was under a strict 180-inning limit. When that led to incredulity on the part of the Mets and the media, with some suggesting nefarious or manipulative intent on the part of Boras, Harvey himself endorsed the idea, albeit without much gusto. The next day, however, Harvey [or a ghostwriter] published an article at the Players' Tribune entitled "I Will Pitch in the Playoffs," sticking to the 180-inning figure but portraying it as the figure that would allow him to, well, pitch in the playoffs.

Whatever you think about Harvey, or Boras, or the Mets, everyone can agree this entire situation has been a huge mess. A frequent refrain has been that the timing of Harvey's announcement was the real problem. If he had told the team his limit earlier, the argument goes, they could have spread his starts out, skipped him when appropriate, and not encountered this problem at all. Maybe his inning count shouldn't matter -- Russell Carleton published an article Tuesday at Baseball Prospectus, looking at prior pitchers returning from TJ, and found no evidence of increased injury risk with lots of innings -- but unless you can convince Harvey and the Mets of that, planning all this out ahead of time seems like the next best thing.

Tom Tango wrote a brief blog post about this last week, proposing a "once per week" rule for pitchers returning from a missed season, which harkens back to this post from 2012, the year of the (in)famous Stephen Strasburg shutdown.

Tango points out how easy it would be for most teams to do this, requiring no additional starts from a "sixth man" than they normally get over the course of a season. I wanted to put this to the test, focusing on the Mets in particular. Let's gaze into the multiverse and hypothesize a Mets' team that pitched Matt Harvey only once per week, under the premise that this would not require substantial contortions of their rotation.

The Mets have five games remaining, including tonight; I'm assuming the starters for those games will be Steven Matz, Logan Verrett, Noah Syndergaard, Jacob deGrom, and Bartolo Colon base on the pitching probables listed on the official Mets site and two educated guesses (honestly, it doesn't matter that much). Here's how the real Mets' starts broke down in 2015:

Harvey, for all the talk of the inning limit, really hasn't stood out in terms of numbers of starts. True, he has the lowest total of the non-Syndergaard starting five, but it's not by a large margin, and he's at 183.1 innings (and likely done for the regular season). This next chart shows how many days of rest each starter had between games, with 4 conforming to a normal 5-man rotation.

First starts of the season were excluded.

Almost two-thirds of all starts from Mets pitchers were made with at least five days of rest, and more than one in five were made with at least six. I'm not sure if that's different from the rest of the league, and if so by how much, but it's certainly interesting. Also interesting: not a single start was made on three or fewer days of rest. Old Hoss Radbourn sheds a single tear.

In another universe, we can still keep these general constraints and design a different Harvey pitching schedule. No short-rest starts; no overuse of the backups (Gee, Matz, Verrett, and Montero). In the real world, Harvey's first start was on Thursday, April 9th, so Thursdays in this hypothetical become 'Matt Harvey Days' (and one of our editors cringes). When there isn't a game on Thursday, Harvey's start is moved to Friday. It's like a holiday! Holidays are fun! Already this plan has benefits [Ed's note: grrrr].

I also tried to roughly mirror the personnel moves of the Mets. Noah Syndergaard didn't start the year in the big leagues, so he doesn't in the universe either. Steven Matz spent more than a month on the disabled list, so he gets a break of similar length here as well. There's no guarantee that injuries and the like would be the same in the this alternate world, but I felt it was important to show this was plausible in a real scenario, with injuries and paternity leave and service time considerations. That said, this also proved the hardest constraint, and the one that did get bent the most.

Overall, though, this didn't disrupt things too much! Here's the chart of distribution of starts by pitcher from above, with our new distribution overlaid in orange.

Granted, the backups are used more frequently, with Montero and Matz picking up an extra start, but the biggest change is Syndergaard going from 24 to 26 starts. In this scenario, he debuted on April 28th, two weeks earlier than in our world. I think this is a pretty amenable adjustment; this season Syndergaard stayed in the minors long enough for an extra year of service time, but not long enough to avoid Super Two status, and the same is true in in our created universe.

With Syndergaard called up earlier, he, Colon, deGrom, and Niese started as frequently as possible, with Harvey pitching Thursdays, and Gee, Matz, Verrett, and Montero filling in as needed. The impact on Harvey was fairly minor -- two starts fewer in this scenario -- but this program has advantages over what transpired this year for the Mets.

Harvey last pitched on Saturday, and luckily, this last series with the Mets doesn't matter, allowing him to (most likely) not pitch again until the playoffs. In the alternate universe, Harvey pitches consistently all the way through to the end of the season, and makes his last start this Thursday. If the Mets hadn't surged quite so much and the Nationals hadn't collapsed quite so abruptly, having Harvey able to start and pitch as many innings as he wanted these last two weeks would be a big deal.

It also would likely have playoff advantages. The Mets are starting to think about home field advantage, and how they'll deploy their pitchers against their likely Divisional opponent, the Dodgers. Harvey having two fewer starts, and presumably 10 to 12 fewer innings than he does in reality, would allow for additional flexibility, or just more starts from the man with 2.80 ERA/3.16 FIP.

What impact might this have on the other pitchers? It would certainly disrupt their routine; there's no denying that. If they could all buy in, however, the actual change in their rest patterns would be relatively small. Here's the chart of rest days from above, with the new scenario in orange:

The number of starts made on a "normal" four days of rest falls almost by half, with the number made with six and seven-plus days both almost double. This is including Harvey, however, who moved from mostly four and five to entirely five, six, and seven-plus; a chart without him shows a more measured shift.

In our alternate universe, the Mets keep Harvey to one start a week, with minimal disruption. Ideally, they end up in a far better competitive position and avoid a nasty public relations fight with one of their stars.

If you're curious about the details, I put the entire distribution in this spreadsheet. I don't expect any team to buy into this strategy wholeheartedly, but six-man rotations for at least stretches of the season seem likely to be more common and in the case of elbow injuries, may make a lot of sense. Tommy John isn't going away, and teams are going to start adapting.

. . .

Henry Druschel is a Contributor at Beyond the Box Score. You can follow him on Twitter at @henrydruschel.