Nationals Arm Race

"… the reason you win or lose is darn near always the same – pitching.” — Earl Weaver

Statistics, Deception and Boswell


In case you weren't aware, Clayton Kershaw is really good.  Photo via wiki.

In case you weren’t aware, Clayton Kershaw is really good. Photo via wiki.

Was casually reading Tom Boswell‘s article in the post today (which can be summarized as “1,000 words on why you don’t want to face Clayton Kershaw in the playoffs, duh) and he posted the following intriguing statistic as part of his argument:

  • In games started by Kershaw, the Dodgers are 19-4
  • In games NOT started by Kershaw, the Dodgers are 59-58

Ergo, the Dodgers are really beatable when they don’t throw Kershaw.

This sounded like a great stat.  Until you dive a little deeper and you do the same analysis for our own first place/better record than the Dodgers team.

Los Angeles is, as of the moment of this writing, 78-62.  By starter, their team record breaks down like this:

  • Kershaw: 19-4
  • Ryu: 15-9
  • Greinke: 14-13
  • Haren: 13-14
  • Beckett: 9-11
  • Other randoms: 8-11

All others besides Kershaw and Ryu: 44-49
All others but Kershaw: 59-58

Pretty compelling, eh? They’re a losing team when not throwing one of their two aces in Kershaw or Hyun-jin Ryu.   Surprisingly, the team under their supposed #2 Zack Grienke sports just a .500 record on the year through 27 starts despite his known quality.

Lets do the same analysis for the Nationals, currently sporting a better record of 79-59 and having just beaten the Dodgers two out of three at their house.  Here’s the Nats team record under each starter this year:

  • Zimmerman: 19-9
  • Fister: 14-7
  • Gio: 13-10
  • Stras: 16-13
  • Roark: 15-12
  • Jordan: 1-4
  • Treinen: 1-4

All others besides Zimmermann and Fister: 46-43.

Hmm.  So, much like the Dodgers, if you play the Nats and you’re not facing one of OUR two best pitchers, we’re basically a .500 team.

I guess the point is this: in a sport where a team that is winning 57% of its games has the best record in the majors, the margins for winning and losing are pretty slim.  Or maybe the point is this: you can use stats to support pretty much whatever hypothesis you wish to postulate.

I think everyone knows that beating Kershaw is nearly impossible, and facing him twice in a short series may be the difference between advancing and going home.  But then again, this is the same Kershaw who has (believe it or not) a career 4.23 ERA in the post-season and got hammered by the Cardinals for 7 runs in 4 innings in last  year’s NLCS.   I guess that’s why they play the games.

But, I think it is also safe to say that the series in Los Angeles showed why it’d be one heck of an NLCS if the seedings held and the two top NL teams  held form in the divisional series.  Lets hope it comes to pass.

11 Responses to 'Statistics, Deception and Boswell'

Subscribe to comments with RSS or TrackBack to 'Statistics, Deception and Boswell'.

  1. I was surprised to see Kershaw regularly hit 96 the other day. I had formed an impression of him as a good FB, great curve pitcher, but that FB was dynamite. The only thing to hope for in the playoffs is that randomness spins out one of his mediocre starts.

    Not sure that I agree that the comparison to the Nats is valid though. According to your numbers, they are at least +3 with every normal starter (or +9 with the bottom three), and it is the fill in guys who have been highly beatable. Being that it’s Sep, everyone is healthy and they will drop down to 4 starters for the playoffs, it looks unlikely that one of the (-) fill in guys will make a start in the playoffs. Whereas LAD only has one guy with a positive spread out of their back three guys (and that is just +1), so they will be starting two guys in each round of the playoffs with a .500 or losing team record.


    4 Sep 14 at 5:37 pm

  2. well, my bigger point was that you can do nearly the same analysis with our pitchers (taking away our team’s record with our best guy) and get nearly to a .500 record (give or take a couple games over 130, a small percentage). Is there that much of a difference if a team is 14-13 in a starter’s games versus 13-10? Two games over the course of 25? that’s a pretty thin line.

    Todd Boss

    5 Sep 14 at 9:07 am

  3. You just highlighted why I hate the 5-game DLS format, namely because it helps teams with unbalanced rotations and hurts a team like the Nats who can throw four good arms at you. Kershaw pitching 2 out of 5 is a much tougher chllenge than him pitching 2 out of 7, and if the Dodgers used him on short rest in the final game he’d be less likely to be dominant.

    If I had my way, they’d ditch the stupid wild card play in game (seriously? 162 games to get to the playoffs and the ONE to get eliminated?) and increase the divisional series to a full seven games.

    But of course, baseball won’t do that because like with the NFL the owners prefer a watered down, mediocre product in which 80% of the teams still have a chance to make the playoffs as the season winds down. Keeps to idiot casual fans who show up knowing little or nothing about the game coming through the turnstiles.


    5 Sep 14 at 10:35 am

  4. Todd – I was just trying to point out that the biggest negatives for team losses for the Nats came from guys not expected to pitch (or at least start) in any postseason games, while that wasn’t true for the LAD. Is the spread or the sample big enough to make a real conclusion? Probably not.

    But I think it gets to bdrube’s point: the Nats real rotation advantage in the postseason is getting to the 4th starter. Just compare it differently with LAD:
    Kershaw v. JZimm (big advantage LAD just because, you know, Kershaw)
    Ryu v. Stras (slight adv Nats or push)
    Fister v, Greinke (slight adv Nats or push)
    Roark/Gio v. Haren (big adv Nats)

    And this is probably the worst we look against any contender (maybe similar to OAK). You could quibble with that statement, or my conclusions for 2 & 3, but I think 4 is where we’ll always hold a big advantage.


    5 Sep 14 at 12:54 pm

  5. Wally-
    If that’s the case, and we are assuming a game 1 loss, then why not put Roark/Gio in there just to eat it, or Fister to at least keep it a little competitive. Zimm v. Greinke seems like a better use if we’re assuming whoever faces Kershaw doesn’t stand a chance.


    6 Sep 14 at 2:09 pm

  6. Anon – I could see the logic in that, but it rarely happens. I don’t think a team can overtly concede that they just can’t beat another player, plus the randomness of any 1 baseball game is so high, you never know what will happen. Nothing prevents a team from manipulating its rotation for ‘other reasons’ to get to the same result, but in a way that prevents the media from calling them out as chicken s**t.

    On a different topic, I have been in the camp that the Nats should let ALR go in the offseason (spot for Zim, save the $$), but holy cow! This guy has been huge for them this year. I mostly follow the advanced stats, so Rendon clearly looks like their most valuable player this year, but my eyes tell me that ALR has had so many important hits this year.


    8 Sep 14 at 8:11 am

  7. Rotation Matchups: never happen; not in todays’ game of internet “blowups,” hyper aware media and bruised egos of players. It happened all the time two generations ago (sacrificing your lesser starters on the opposing team’s better starters) but not now. But wouldn’t you love to see someone try to do it? Assuming that Kershaw is unbeatable, you line up your #4 against him, then go your 1-2-3 from there.

    LaRoche: *classic* contract year behavior out of him. See 2009 and 2012. If you keep him, you have a massive positional issue. It is clear that Zimmerman cannot play 3rd anymore, so he has go somewhere. And the only two places it makes sense for Zimmerman to go is 1B or LF. If its LF, then you cascade moving Harper somewhere and displacing someone else. Werth can’t play center, and despite my advocating to play Harper in center that doesn’t seem to be in the mindset of this executive/managerial group either. But if you’re putting Harper in center, then you’re also saying good-bye to Span. Which I don’t think is happening; I think they’re going to exercise Span’s option and use Souza as a bench guy, look to move Moore, stick Taylor in the minors another season or use him as a backup OF (though that means lots of backup OFs on this bench thanks to McLouth’s deadweight of a contract).

    So it kind of has to be 1B. At first he can still make highlight reel diving stops, he can use his infielder instincts to dig out ground balls. Heck, he could turn into a GG at the position.

    Todd Boss

    8 Sep 14 at 8:44 am

  8. On matchups and Kershaw, maybe you would be tempted to do it in a 7 game series, if you don’t think that the Dodgers use Kershaw on short rest (1-4-7). But it would be madness to do it in a five game series because you’re probably going to have to win a game that Kershaw starts to win the series. Note that I didn’t say “beat Kershaw.” You need a starter good enough to fight Kerhaw to a draw and have a battle of the bullpens.

    Now with the Nats, because there really isn’t that much of a gap between their starters (in a good way), you don’t need to “sacrifice” one of them against Kershaw anyway because they are all pretty capable of “having a day” and essentially battling Kershaw to a draw. Interesting note: of the 14 shutouts the Nats have tossed this year, five were in games started by Jordan Zimmermann (no surprise there). Tied with that is … Gio Gonzalez with another five. Roark has started two, and Fister and Strasburg one each.

    I agree with letting ALR go after this year and moving Zimmermann to first. The Nats are deep in OF, not IF, and the most likely result is Span’s option being picked up and having Souza and McLouth as a R/L combination on the bench. Not too shabby. I’m fairly optimistic about McLouth being able to contribute something next year because I don’t assume 2014 results dictate 2015 results.

    John C.

    8 Sep 14 at 10:12 am

  9. Remember: Kershaw is not unbeatable in the post season. He got shelled last year in St.L and (believe it or not) has a 4+ ERA and a losing record in the playoffs. I can’t remember where I read it, but it said something like Kerahw has pitched like once in his life on short rest and got killed, so the odds of him going 1-4-7 in a 7-game series seems dim. Worst case is meeting LA in the divisional series and having Kershaw pitch 1st and 5th/decider at home.

    I’m hopeful of this scenario: Wash and LA finish with 2 best NL records, each wins their NLDS series. But LA has to fight it out with St Louis and burn Kershaw in the 1st and 5th games to get to the NLDS. Suddenly he can’t go til game 3. Meanwhile Washington’s depth at starter means we don’t have much of a drop off and can throw whoever in games 1 and 2 in DC, give up on game 3/Kershaw in LA, grab one against LA’s #4 starter in LA, and clinch NL pennant in game 6 at home.

    Todd Boss

    8 Sep 14 at 10:28 am

  10. I agree that freeing up 1B for Zim or even Werth is important, and ALR should be allowed to leave. I was just giving him a shout out for how well he has hit for them this year. And they will miss the LH power bat next year, but it doesn’t change the conclusion.

    Btw, here is my two cents on Zim: if 2014 represents Zim’s throwing power going forward, he has to give some thought to moving to the AL and DHing. Neither 1B nor LF will get him out of making throws that will be important from time to time. And if this is it for him and they won’t/can’t move him, I think LF is the better option because (i) he’ll make less throws overall, (ii) also less that matter, and (iii) can usually be less precise. And to be clear, that comment isn’t based on what is best for his shoulder, it is based on what would hurt the team less. He’ll still have the occasional throw that is cringe -worthy and teams will start taking the extra base on him quite a bit, I just think he’ll be less involved than at 1B. At 1B, Zim will still have the occasional DP throw to make, maybe a throw home, each of which need to be hard and accurate. I don’t think most agree with this, but just my two cents.


    8 Sep 14 at 3:50 pm

  11. Kershaw is not unbeatable in the playoffs. Greg Maddux had a losing playoff record, as did many other good pitchers. I do agree that a best of five format favors teams w/stronger 1 2 pitching punch. All playoffs should be best of 7, but hard to do with wildcards and northern teams making post season…brrrr. League could always shorten the season, but this will never happen.


    10 Sep 14 at 2:36 pm

Leave a Reply