What It Is to Sit a Player

I’m sitting here, reading complaints on Twitter. Not long ago, a starting lineup was revealed for a Tuesday baseball game, and almost immediately some people took issue. They don’t like that a certain bench player is getting a start, and they don’t like that a certain regular starter is sitting out. Why, they wonder, would you ever make your lineup worse, especially when you’re playing against a team in the same division?

To be perfectly honest, yeah, this is inspired by the Seattle Mariners. But this isn’t just a post about the Mariners, because this can apply to every team and to every team’s fan base. There’s not a single fan base immune to lineup complaints, and as much as more people are beginning to understand that lineup order doesn’t make much of a difference, a bench player filling in for a starter always generates a negative response. After all, the bench player, presumably, is worse than the starter, which is why the starter is the starter. So what does it mean to take out a starter?

I’m talking about sitting a guy on a single-game basis. The negative commentary I’ve read today is about Willie Bloomquist slotting in for Dustin Ackley. Bloomquist’s worse than Ackley, most likely. But every manager starts a bench player sometimes. Let’s play around with the significance, and offer examples in the form of win expectancy. What I mean will become clearer in a minute.

The best player in baseball is Mike Trout. I don’t think anyone doubts that. Even the people who voted for Miguel Cabrera for MVP acknowledge that Trout is the best player overall. Two years ago, Trout was worth 10 WAR. One year ago, Trout was worth more than 10 WAR. This year, Trout is on pace to be worth more than 12 WAR. Combined, ZiPS and Steamer project Trout as a 10-win player over a full season. So the worst thing a manager could do would be to sit Mike Trout and start some replacement-level substitute. If that were to be done, then the Angels’ odds of winning the baseball game would drop by about 6.1 percentage points.

What’s the equivalent of 6.1 percentage points, in win expectancy? Let’s say you’re the road team. Let’s say your chances start about 50/50. To have your win expectancy drop about 6.1 percentage points, you’d need to have a scoreless top of the first, and you’d need to start the first batter of the bottom of the first with a ball. That’s what substituting Trout would be worth — three automatic initial outs, plus a ball.

And that’s Mike Trout. Let’s try some different players. Instead of Trout, let’s say you start a bench guy for Miguel Cabrera. The Tigers’ odds of winning would drop by about 4.0 percentage points. The equivalent of that in win expectancy? You’re the road team, and the top of the first begins with two outs. You don’t concede the whole half-inning, like you do in the Trout example, but you get most of the way there.

Bryce Harper? Starting a bench guy for Bryce Harper? The Nationals’ odds of winning would drop by about 3.1 percentage points. That’s like conceding the first out, and then having the second guy start out 0-and-1.

Starting a bench guy for David Ortiz? The Red Sox’s odds of winning would drop by about 2.4 percentage points. That’s like conceding the first out.

Starting a bench guy for Carlos Beltran? The Yankees’ odds of winning would drop by about 1.8 percentage points. That’s like conceding the first two strikes of the game’s first plate appearance.

Every single substitution — they all matter. A team is weakened by a bad player playing in place of a better player. If you sit Mike Trout for a game, it’s like giving away the first half-inning of offense. But, a few things. (1) No other player is Mike Trout. (2) Leverage is low in the top of the first. (3) Boy are there ever other considerations.

Figure that a bench player doesn’t start for a starter for no reason. There’s always some reason, as managers don’t fill out lineup cards randomly. And because there’s a reason, you have to figure the gap between the bench player and the starter might be smaller. Or, if the gap isn’t smaller on that particular day, the manager might have an eye toward future benefits. Maybe he needs to keep the bench player content. Maybe he needs to keep the bench player fresh. Maybe he needs to give the starter a rest, because if he played every day, the starter could wear down.

Maybe the starter’s sick, or in some discomfort. Maybe there’s something about the particular pitching matchup. There’s a lot that can go into a lineup decision, and while it’s an essential pastime of the pastime to criticize team managers, managers need to be given some benefit of the doubt. They are particularly informed, and while that doesn’t prevent them from making bad decisions, it does mean that perceived bad decisions could have very reasonable explanations. You might not like when a starter sits for a bench player, but it could have a long-term benefit. In the moment, unless the player is really, really, outstandingly good, the impact of the substitution is pretty small. It might be like one out. It might be like one strike. It might be like one gust of wind blowing one speck of dust into one hitter’s eye before the first pitch of the game. This doesn’t even say anything about the possibility of the sitting starter pinch-hitting later on.

There’s just precious little point in complaining about single-game lineups. Even perceived suboptimal decisions make little practical difference, and we’re not always privy to reasons for the decisions that might reduce the impact. Which isn’t to say you can never justifiably complain. A pattern of weird substitutions would be worse. Consistently sitting a decent player would be worse. Consistently running out a lousy lineup arrangement would be a problem. Starting the wrong guys in the playoffs would be terrible, when the stakes are their highest and there’s an abundance of off-days. But as I write this, it’s April 15, and a bunch of teams are preparing for a regular-season baseball game. Some of those teams are going to have bench players starting in place of superior regulars. Some of the fans of those teams are going to really not like it. And it’s probably going to really not matter. Complaining’s a good tool. You don’t want to wear out your good tools.

Print This Post

Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.

44 Responses to “What It Is to Sit a Player”

You can follow any responses to this entry through the RSS 2.0 feed.
  1. KDL says:

    Love the analogy. Fun piece.

    What I hate though is what I’ve seen a bunch already this season. (Maybe it’s gone on in the past and I just didn’t notice, but…) Managers giving multiple regulars the same day-off. I understand the day-off idea. I’ve never been one to attack the idea of a day off. And this piece does a nice job of explaining how little damage is done. I also get that injuries become a factor. But the idea of sitting 2 or more regulars on SCHEDULED rest is odd to me. Even if it only amounts to 2 out and 1-2 count.

    Vote -1 Vote +1

    • Mr. Jones says:

      Ron Washington does this regularly, and it bugs the hell out of me as a Rangers fan. You’d think they could plan well enough to stagger the planned off-days.

      Vote -1 Vote +1

      • Richie says:

        Heck no you don’t stagger. A day game after a night game, good time to rest whomever needs it. Left handed starter going for the other team? Get any of your righty bench bats in there that day. No, having 3 50% win chances is not better than having 2 55% and a 40%.

        Vote -1 Vote +1

        • Michael says:

          Actually, .55^2*.4 = .121 < .125 = .5^3
          But, I agree with you in sentiment.

          +8 Vote -1 Vote +1

        • John says:

          Michael, you calculated the chance of winning all 3 games, which is better in the 3×50% scenario, but the chance of losing all 3 games is also higher. The expected number of wins is the same in both scenarios (.5×3 = 1.5 vs .4+.55+.55 = 1.5).

          +15 Vote -1 Vote +1

        • stevenam says:

          Of course you stagger. Day game after night game? Give one guy the night game off, another the day. 6 of one VI of another. Not difficult. If you’re talking about resting more than two players, do it over more than two days. How much are these guys (managers, that is) paid?

          Vote -1 Vote +1

      • Ruki Motomiya says:

        It seems to me like it would be better to do it all in one game and not staggering it. Staggering it means you decrease your odds of winning multiple games (Which, in turn, means you’re increasing your odds of losing 2+ games, as your odds of losing either one individually is higher), while stacking them all on one day means you have a higher chance to lose one game, but a then higher chance to win all the other days you would have rested more people.

        Finally, you can rest multiple guys against worse pitchers or same handed guys while bringing in platoon bats, either marginalizing the disadvantage via platoon or taking the odds from a game you have a heightened chance of winning.

        Vote -1 Vote +1

        • Ruki Motomiya says:

          Considering how this seems to be a common “I wonder if it’s true” thing in the comments so far, maybe we’ll get to see a fun research article on it. :)

          Vote -1 Vote +1

        • tz says:


          It would also be good to see who the opponents are when managers rest their players. I’d have to think that this is less likely to happen against divisional foes, or later in the season against pennant contenders.

          Vote -1 Vote +1

  2. Ron says:

    Can I still complain about the fact that Morrison starts over Saunders most of the time this year?

    One thing you overlooked is the watchability factor. Willie Bloomquist drops the watchability of any game a much larger percentage than his WPA.

    Vote -1 Vote +1

  3. Stan "The Boy" Taylor says:

    That was fast. Saw all that on twitter and figured you or Dave would have something up tomorrow or later in the week.

    It is a perfect storm of the M’s fans still thinking Ackley can turn into an impossibly good player, like .350/.500/.450 or something, and the replacement being none other than WFB.

    Vote -1 Vote +1

  4. Westside guy says:

    Wait – we’re only allowed to complain if it’s justified? Where’s the fun in that?

    +9 Vote -1 Vote +1

    • Richie says:

      Many ‘+++’. And actually I think pyschologists have determined that complaing whets your appetite for further complaining, it doesn’t sate it against further.

      Vote -1 Vote +1

  5. Reptar says:

    So if an umpire screws the Yankees out of 2 strikes, then he might has well have removed Carlos Beltran from the game? Wow, I am going to be even more pissed when they get a ball/strike call wrong.

    +5 Vote -1 Vote +1

    • Ruki Motomiya says:

      But you have to counter it with all the balls/strike calls you get wrong for your side. Considering how well the Yankees have been framing I don’t think you have too much to be mad about. :)

      Vote -1 Vote +1

    • Iron says:

      “In a completely unexpected move, Joe West reset the count to 0-0 and ejected a surprised Carlos Beltran.”

      +14 Vote -1 Vote +1

  6. Lex Logan says:

    It isn’t obvious to me that benching two regulars on the same day, and playing both on the next day, is worse than benching one on each day. Let’s say your odds go from 50% to 44% in one game, or two games at 47%. How is that clearly worse?

    Vote -1 Vote +1

    • bilbovibrator says:

      i think in 162 examples of this “coin flip” u wanna stack the percentage high as possible as frequently as possible to rack up the most wins. but im just a dumb hs drop out

      Vote -1 Vote +1

  7. Kinanik says:

    I wonder if it makes sense to sit multiple players at once. If two teams are evenly matched, and game-run-differential is a normal distribution, it is the departure from the average that matters most. If you’re 50/50, cutting the first 3 WAR player out will move you to 46/54, while sitting the second might only move you to 44/56. If you sit players on different days, you’re taking a 4% cut twice; if you sit both on the same day, you get one 6% fall.

    It’s the same principle as in H2H fantasy baseball; the first standard deviation above average earns you more wins than the second.

    The math changes if the matchups don’t start out even. An above average team might be better off sitting one at a time.

    Vote -1 Vote +1

    • Ruki Motomiya says:

      Something else to consider is that by splitting it over two days, you increase your odds of losing 2 games compared to losing 1 and winning 1, since your odds of losing both games has gone down.

      Vote -1 Vote +1

      • Kinanik says:

        Using my numbers (obv. wrong, but to demonstrate…)

        No sitting (baseline): (50/50)
        Win 2: 25%
        Win 1: 50%
        Win 0: 25%

        Sit one day: (50/50 and 44/56)
        Win 2: 22%
        Win 1: 50%
        Win 0: 28%

        Sit two days: (46/54 and 46/54)
        Win 2: 21.16%
        Win 1: 49.68%
        Win 0: 29.16%

        *Assuming* my original numbers were in the ballpark of correct (that deviations from the average are more important nearer to the average), then the chance of losing two games is higher if you sit one player per game than if you sit both at once (albeit a tiny amount).

        Vote -1 Vote +1

  8. Lex Logan says:

    With last year’s Reds it seemed that every regular given a day off played much better subsequently. I would’ve given everyone a day off every week or two starting at the all-star break. Instead, Baker kept them playing and they withered at the end. Purely anecdotal but I would certainly like to see some data on performance effects of days off.

    Vote -1 Vote +1

  9. Richie says:

    Let’s be blunt. Much of this is fantasy-driven. Unless you’re in Vegas, which case then it’s sportsbook-driven. Bench the guy during a week I don’t have him in, dagnabit!

    Vote -1 Vote +1

  10. ghug says:

    Ackley sat for WFB two days ago and Michael Saunders is also on the bench. I think a bit of annoyance is justified.

    Vote -1 Vote +1

  11. Ruki Motomiya says:

    Plus, there’s the simple fact that players need to rest sometimes. Who knows how much you get back that you lost by having the player fresh next game? (Yes, I know this was pointed out in the article, I’m just gonna guess the majority of times that a non-platoon player sits it is for rest.)

    Vote -1 Vote +1

  12. jose luis says:

    Damn, Sullivan. I love you, but you are leaving so little for my fanhood.

    Vote -1 Vote +1

  13. bilbovibrator says:

    “…and while it’s an essential pastime of the pastime to criticize team managers”

    hahaha so true

    Vote -1 Vote +1

  14. Ian R. says:

    It’s worth noting that an MLB bench player is often somewhat better than replacement level. Heck, some teams have bench players who could start for some other clubs. A team with decent bench guys will see even less of a drop-off, relatively speaking.

    Also, platoon splits. When the Red Sox bench Ortiz, for example, it’s often against a LH starter. He still hits lefties well enough that taking him out hurts the offense, but it’s less of an impact than benching him against a right-hander would have.

    Vote -1 Vote +1

    • tz says:

      Great point Ian. Interestingly enough, this somewhat explains two apparently different items:

      – Some of the “extra” success that the Rays and A’s have generated from having plenty of positive WAR players on their bench.
      – The continuing thought process that a combination of three 2 WAR players are better than a combination of one 6 WAR player and two 0 WAR players. While I don’t believe this holds for starters, improving the “floor” level of your bench clearly has merit.

      Vote -1 Vote +1

    • Jack Carter says:

      “It’s worth noting that an MLB bench player is often somewhat better than replacement level.”

      –Not on the Mets.

      “Heck, some teams have bench players who could start for some other clubs.”

      –Not on the Mets.

      “A team with decent bench guys will see even less of a drop-off, relatively speaking.”

      –Not, however, on the Mets.

      –Sorry. The spirits took hold of me. In his first season as GM Sandy Alderson showed the expected knack for replacing Minaya’s disastrous bench guys (who in the case of Little Sarge Minaya even intended to have start) with players worth half a win or even a win. Guys with a little upside who occasionally helped the club win. It was cause for real optimism, especially since Minaya had punted a couple of division titles by needlessly surrounding the team’s stars with replacement level and worse players.

      Since then, though, Alderson’s been an almost complete failure, giving us the less than zero Omar Quintanilla, and farting around with Lucas Duda and his Glove From Hell for four seasons, sapping completely what modest value Duda’s bat provides. Alderson has also failed to do what we so often claim an upper crust GM should be able to do–assemble a competent pen on the cheap.

      I suppose, after all, it’s only been the difference between 75 and 77 wins but with the Mets farm and young pitching it might matter in 2016. I’ll be sorry to see a Mets team miss a close title race due to Alderson’s inability to work the margins. In 2010 that was about the last thing I would have expected to be writing about the team today.

      Vote -1 Vote +1

  15. bookbook says:

    Point taken. No need to complain. And YET…

    The move seemed odd because…
    * Ackley had just had the day off on Sunday, April 13th. So, that was two games not starting out of three after playing every game before that…
    * He’s also one of the few M’s who is hitting right now.
    * AND sensitive enough that maybe you don’t want to mess with him when he’s swinging the bat well.
    * Michael Saunders was on the bench. It’s hard to think of a situation where you want to start Willie in the OF and leave Saunders on the bench.
    * If the more salient point was to get Willie into the starting line up for a (second) game (in three days), both Brad Miller and Kyle Seager actually need a game off (I think they’ve started every game, and near-rookie Miller in particular is losing his plate discipline a tad).

    So the particular substitution was odd, to an outsider without whatever wisdom and insider acumen Mr. McLendon brought to the table.

    Vote -1 Vote +1

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>