Assessing the loss of Stephen Strasburg on the NL playoff picture
Over the weekend, the Nationals revealed that ace Stephen Strasburg will make just two more starts, the final one on September 12, before being shut down for the remainder of the season. Whether protecting his valuable right arm as a precaution against future injury while compromising the upstart contender’s chance at a championship is indeed the right decision is a hotly contested topic that will be debated for years to com. It’s easy to sit back and second-guess the team’s handling of him up to this point if they knew October was a possibility, to suggest that they should have delayed the start of his season until May or found a phantom injury to sideline him for a few weeks during the summer, but by now it’s a moot point. Nationals general manager Mike Rizzo hasn’t wavered from the plan he set out earlier this year, and he won’t, no matter how much public pressure he receives.
While we can’t estimate what losing Strasburg means to the Nats on the psychological side, we can grapple with the impact on the team’s rotation in runs, both in the regular season and the postseason, knowing that they have anointed John Lannan to take over his slot for the remainder of the regular season. A 27-year-old lefty, Lannan was a staple of Washington’s rotation from 2008-2011, but with the additions of Gio Gonzalez and Edwin Jackson, the returns of Strasburg and Jordan Zimmermann from Tommy John surgery, and the emergence of 2007 first round pick Ross Detwiler, he was squeezed out. Despite drawing a salary of $5 million, Lannan has spent nearly the entire season toiling for Triple-A Syracuse, making just two spot starts for the Nats. In a very basic way, he’s the antithesis of Strasburg, in that he pitches to contact and misses few bats. Relying on two- and four-seam fastballs that average around 89 mph, Lannan generates a ton of groundballs while striking out just 4.7 per nine, a career rate less than half of Strasburg’s 11.3 per nine. His walk and homer rates (3.4 and 0.9 per nine) are both inferior to Strasburg’s as well (2.3 and 0.7 per nine), but then the same can be said for most pitchers.
Based on the remaining schedule, Lannan figures to make two or three starts. If the Nats stick to a strict five-man rotation the rest of the way, he would start on September 19 against the Dodgers, September 24 against the Brewers and September 29 against the Cardinals. However, Washington has off days on September 13 and 17, making it possible for his first turn to be skipped. That could allow manager Davey Johnson to realign his rotation with an eye toward the Division Series, a destination that now seems probable given the team’s 6 1/2 game lead over the Braves in the NL East; the Baseball Prospectus Playoff Odds report shows the Nats with a 95.8 percent chance of winning the division.
Using the methodology that I deployed to illustrate the potential impact of trading for Ryan Dempster, we can quantify the step down from Strasburg to Lannan. That method assumes that each pitcher will perform as anticipated based upon his Fielding Independent Pitching metric — an ERA estimator that uses strikeout, walk, home run and hit-by-pitch rates — that he will maintain the same level of stamina in terms of innings per start the rest of the way, and that the bullpen will collectively perform at a level equivalent to the entire unit’s FIP to date. For Strasburg, I’m using his 2012 numbers; for Lannan, I’m using 2011, when he set career lows in both ERA (3.70) and FIP (4.25), thereby putting his best foot forward:
|Pitcher||IP/GS||IP remaining||FIP||Runs||Bullpen IP||FIP||Bullpen Runs||Total Runs|
Under those assumptions, the impact comes out to exactly three runs, less than one-third of a win using the basic rule of thumb that every 10 marginal runs scored or saved produces an extra win. If we instead use career FIP for Strasburg (2.39) and Lannan (4.58), the gap increases to 4.1 runs; if we use the actual 2011/2012 ERAs in question, the gap shrinks to 1.5 runs. Given the size of their lead, the chances of that difference having an impact on the team’s fortunes is quite small, though it is worth noting that at 82-52, the team has just a one-game lead over the 82-54 Reds for the league’s best record and thus home field advantage through the playoffs.
Lannan isn’t likely to take a start in the playoffs; instead, Detwiler (3.15 ERA, 3.71 FIP) is likely to fill the slot, and as the fourth starter behind Gonzalez, Zimmermann and Jackson (in some order or another). We can estimate that impact using a method that Baseball Prospecuts’ Nate Silver devised based upon the historical distribution of postseason starts; Silver found that teams’ No. 1 pitchers took 31 percent of those starts, with number twos at 28 percent, threes at 23 percent, and fours at 18 percent. Using FIP and assuming Johnson would align his starters accordingly — but turn to the postseason-tested Jackson instead of Detwiler for his four in the initial iteration — the rotation goes from this:
Assuming the aforementioned distribution of starts, a constant level of bullpen performance and the maximum of 19 playoff games over one best-of-five and two best-of-seven series, the Strasburg rotation would allow 63.9 earned runs, the non-Strasburg one 67.5 earned runs, a difference of 3.6 runs. Again, that’s less than half a win, albeit in a situation where such a difference is magnified. Assuming the team’s offense maintains its level of scoring (4.42 runs per game) and that the pitching and defense will continue allowing unearned runs at its same rate, the team’s Pythagorean winning percentage would drop from .593 to .570, which still looks quite good.
How does that stack up against the other NL contenders? Divining which starters each of a half-dozen managers will choose and how they will deploy them is no easy task. Here’s my best educated guesswork, applying the same assumptions as above, keeping in mind each team’s injury situation and the amount of postseason experience under each candidate’s belt. Note that for pitchers who have worked for more than one team, I’m using their full-season stats.
Cardinals: Adam Wainwright (3.02), Jaime Garcia (3.14), Lance Lynn (3.63), Kyle Lohse (3.68), bullpen (4.14). Pythagorean estimate: .611.
On paper, the defending world champions look quite formidable, but several caveats apply. Garcia just returned from two and a half months lost to a shoulder injury, Lynn has been flagging lately and may face workload concerns and despite the league’s most potent offense (4.82 runs per game), the Cards have underperformed their Pythagorean projected winning percentage by 32 points (.573 versus .541). They’re 8 1/2 games back in the division, and lead the Dodgers by just half a game for the second wild card spot. Note that swapping out Lynn for Jake Westbrook (3.84 FIP) only bumps the Pythag percentage down to .608.
Braves: Tim Hudson (3.87), Kris Medlen (2.28), Ben Sheets (4.28), Paul Maholm (4.25), bullpen (3.35). Pythagorean estimate: .583.
The Braves, who lead the wild-card race by three games, have more potential starting candidates than any other team, and actually figuring out who they might use is a challenge. I’ve assumed that struggling Tommy Hanson (4.59) and young Randall Delgado (4.21) will take a back seat to in-season acquisitions Maholm and Sheets, but the latter is on the disabled list now due to shoulder inflammation. Swap in the more experienced Hanson and the team’s estimate drops to .573, but that doesn’t assume any regression on the part of Medlen, who has averaged 7.0 innings per start while delivering a 0.54 ERA in seven turns, both unsustainable. If we instead use his career averages of 6.1 innings per turn and a 3.10 FIP, that cuts their Pythagorean estimate another 14 points, to .559.
Giants: Matt Cain (3.47), Madison Bumgarner (3.47), Ryan Vogelsong (3.83), Tim Lincecum (3.92), bullpen (3.79). Pythagorean estimate: .542.
Given Lincecum’s 3.76 ERA since the All-Star break, that FIP doesn’t seem unreasonable, but his inconsistency and short leash — one start of more than six innings out of his last four — suggest that manager Bruce Bochy isn’t going to trust him at the front of the rotation.
Reds: Johnny Cueto (3.18), Mat Latos (4.13), Mike Leake (4.23), Bronson Arroyo (4.24), bullpen (3.19). Pythagorean estimate: .540.
Because of the dropoff beyond Cueto — nearly a full run per nine innings — the Reds don’t appear particularly formidable, even with the league’s best bullpen. Then again, they have the NL Central virtually sewn up, which means they’ll get to forego the one-game playoff. Note that swapping in Homer Bailey (4.32) for Leake or Arroyo would only push that percentage down a few points.
Pirates: A.J. Burnett (3.82), Wandy Rodriguez (4.06), James McDonald (4.06), Jeff Karstens (3.41), bullpen (3.70). Pythagorean estimate: .502.
With an 11-20 record since the beginning of August, the Pirates’ ship is taking on too much water; they’re now 2 1/2 games back for the second wild card spot, and an unsurmountable 11 games back in the division race. McDonald has been particularly pummeled lately, but when the alternative is Kevin Correia (4.66 FIP), manager Clint Hurdle doesn’t have a viable alternative at his disposal, particularly given the release of Erik Bedard, who for his troubles had still pitched to a 4.14 FIP.
Dodgers: Clayton Kershaw (2.97), Chris Capuano (3.74), Joe Blanton (4.12), Josh Beckett (4.18), bullpen (3.60). Pythagorean estimate: .499.
The Dodgers’ chances take a huge hit with the loss of Chad Billingsley (3.41), who if he returns from his latest bout of elbow inflammation will likely be limited to a bullpen role. That leaves the number two spot in the hands of the fading Capuano unless Beckett can rediscover his lost form. Note that the team’s chances fall if Aaron Harang (4.24) is swapped in for any of the starters. On the other hand, note that the 4.01 runs per game produced by the offense is based on a lineup that lacked Matt Kemp for two months and that more recently underwent a significant facelift via the additions of Hanley Ramirez and Adrian Gonzalez, massive upgrades over the unproductive Dee Gordon, Juan Uribe and James Loney. Swap in the team’s second half scoring rate of 4.37 runs per game, and the Dodgers’ estimate rises to .541, even without Billingsley.
All told, the Strasburg-less Nats’ estimate of .570 is the third-highest among the league’s seven postseason contenders, but the two teams above them are likely headed for a one-game wild-card playoff, a consideration that Silver’s distribution model can’t account for, and one I’ve thus not accounted for either. But suffice it to say that while the Nationals take a hit with the loss of Strasburg, they should still be a force to be reckoned with in October.