Thursday, June 3, 2010

EYE OF THE TIGER: A. Galarraga: 9.0 IP 0 R 1 H* 0 BB 3 K; 28 UP*, 27 DOWN.

* But see, Joyce, J. "I just cost that kid a perfect game." June 2, 2010.

ETA: Joe Posnanski adds the perfect grace note to the call and its aftermath.


  1. Professor Jeff9:32 AM

    1) Upon further review, it's "Galarraga," not "Gallaraga."

    2) Does this incident make anyone more supportive of increasing the use of replay reviews?  I've been pretty set against it, but I'm growing more sympathetic to the idea, provided it's very limited in scope.  Say, managers get only one challenge per game, and they can only use it on safe/out calls at a base or on fly-ball catch/trap disputes -- i.e., calls where video evidence might provide conclusive evidence one way or the other.

  2. 1) There's already replay for home-run calls. I'm against it in other situations: bad judgment calls by umpires rarely make a difference (it didn't even make a difference-difference in this game), and they pale in comparison to the mistakes made a large percentage of the time on ball-strike calls.  (Evidence is quite solid that different umpires have different strike zones, and a number of umpires suffer from what I call the Enrico Palazzo effect with their strike zones.  All of which could be fixed with automated strike zones.)  Baseball games are already too long, and adding another five minutes on this sort of thing would be pretty ugly.

    2) NB MLB Rule 4.19, prohibiting review of "judgment calls."  Everyone harkening back to the Pine Tar game is forgetting that that was a protest over an umpire's erroneous interpretation of the rule book rather than a bad judgment call.  Habeas afficianados will be amused to note that MLB has a "harmless error" doctrine.

  3. The Pathetic Earthling9:42 AM


    I just can't see that, Jeff.  It's so rare that there are history-altering calls, that I think the cost here is tiny.  Sure Galarraga becomes a great sports trivia question in thirty years, but to avoid another blown perfect game or no-hitter or whatever, we add in a five minute delay in the middle of a ballgame?  More than that, those replays won't be available for game-ending events, as here, but will have been largely used up in some random fourth inning sac fly with a questionable call at second trying to keep a rally alive.  

  4. Meghan9:43 AM

    I was so bummed out by this.  I get that human error is a part of all games but it seems like there could be a way for replay to be used where it's not even in the manager's hands, just the replay ump.  When something is egregious enough, as last night's call was, the upstairs guy could phone it in pretty quickly. 

    Man, that was sad.

  5. The Pathetic Earthling9:44 AM

    Also, I think the way people will break on this decision may track nicely their view of substantive v. procedural due process. 

  6. J. Bowman9:50 AM

    Galarraga's reaction to the call, and all interviews with him in the 12 hours since, should be part of a seminar in how to deal with disappointment.
    I agree with Ted's thoughts on replay in general. Applying it to this situation, if you're going to use a single missed call to argue in favor of replay, wouldn't you want to use a call that actually had an effect on the game? The historical value of a perfect game aside, I'd be more sympathetic to the outcry over this if the Indians had then come back to win, or even tie.

  7. Benner9:51 AM

    Isn't this more of an equal protection issue?  

    I vote no replay -- a home run call directly implicates the score.  Sure, the individual lost a perfect game, but he got the next out.  The standings remain the same.  I may revisit on calls at the plate.  Alternatively, each manager gets one challenge per game? 

    The Flyers benefitted last night from replay, so I'm not against it in principal.  (I have a skewed view of "principle.")

  8. Separately, in the long run, the missed call does more for Galarraga's immortality.  We remember Ernie Shore and Harvey Haddix, who are a huge part of baseball lore despite not having an "official" perfect game, and now we'll remember Galarraga, too.  In contrast, I can remember about nine of the twenty or so actual perfect games: Young, Larsen, Koufax, Barker, Cone, Johnson, Buehrle, Braden, and Halladay, so I imagine Galarraga is better off than the median perfect game pitcher--especially since it seems unlikely that Galarraga is going to have a better career than Len Barker.

    Interesting statistic: of 32 pitchers who got the first 26 outs of a perfect game, ten of them (including Galarraga) blew it on the 27th batter (Haddix and Pedro Martinez are the others).  You'd think 8.2 IP of perfect baseball would give you more than a 70% chance of a perfect game.

  9. J. Bowman10:32 AM

    A .3125 OBP from the 27th hitter actually sounds reasonable.

  10. I was going to say pretty much what Ted said -- Galarraga is going to end up way more famous for this than if he had thrown a "real" perfect game, both because of the unusual circumstances and because of his reaction.

    I am firmly in the "no replay" camp.  Baseball games are agonizingly long as it is.  I'm a huge Red Sox fan, and sometimes I don't even feel like watching Sox-Yanks games because I know they're going to go 5 hours.  Even a small delay is going to push it even further over the top. 

    There's also the slippery slope issue.  Keith Olbermann wrote a blog entry last night saying MLB should just reverse the call because there's no cost in doing so.  But I think we can all see where that would lead, and it wouldn't be anywhere good.  If you're going to review a call where a perfect game was at stake, how about a no-hitter?  How about a case where a batting or ERA title is on the line?  How about if some guy has a record errorless streak going, bobbles the ball, recovers to make the play, and the runner is called safe erroneously?  How about a situation where a guy has a contractual incentive to hit a certain offensive/pitching number and a blown call costs him a million dollars?  How about close calls that really do affect the outcomes of games?  They'd all have to be reviewed and the game would become unbearable. 

  11. Oh, forgot to make my other point:  not only are umpiring controversies part of the game, but they are actually good for baseball.  I think way more people are aware of and talking about last night's game than would otherwise be the case, even with a perfect game.  Yes, many of the comments are critical of MLB, but I think in a situation where baseball is competing with so many other forms of entertainment, any press is good press in a case like this.  This reminds people that for all of its flaws, baseball can be an exciting game where anything can happen.  (Now please don't quote this back to me the next time the Red Sox lose on a blown call.  Thanks.)

  12. isaac_spaceman11:06 AM

    Upon what do you base the notion that bad judgment calls by umpires rarely make a difference?  For example, the difference in run expectancy between no outs, one on and one out, none on is 0.656 runs.  The difference between no outs, two on and one out, one on is 1.000 runs.  Make those two mistakes in one team's favor in a single game, and you've spotted them a run and a half.  How many times in 2500 games do you think that happens?  Or here ( is a game summary from yesterday, when a blown call essentially resulted in a 50% swing in win expectancy. 

    We now have technology that can call every ball and strike correctly.  Yet we still rely on humans who frequently call them incorrectly, sometimes inconsistently so.  Why do we do this (apart from the umpires' union)?  That umpire errors usually don't make a difference only makes it feel so much more wrong when they do.   

  13. isaac_spaceman11:08 AM

    Umpiring controversies are good for baseball the same way that growing certainty about biased referees is good for basketball, which is to say it is not. 

  14. No replays. Blown calls are part of the game. Not so much "good for the game" as "the cost of doing business."

    No matter what the record books will say, or whether or not the commissioner intervenes, we all know that <span>Galarraga pitched a perfect game and then schooled us all about class.

  15. isaac_spaceman11:43 AM

    Okay, the argument against replay is that (a) it will add too much time to the game; and (b) it won't usually matter.  I accept the premise of (a), that it will add time to the game.  I disagree with the premise of (b) -- it matters often enough that it matters.  In all likelihood, your estimate of how often calls are blown and how much they affect games probably dictates how you feel about replay. If we were all working from the same assumptions on that side of the ledger, there probably would be a much greater consensus.  Apparently I think that there is more error than most people think. 

    Two follow-up questions: 

    1.  If you're so worried about the five minutes that a replay might cost every time there's a close call, why aren't you up in arms about the lack of enforcement of time limits between pitches?  Every batter steps out of the batter's box between pitches; most pitchers fail to observe the 10-second rule.  You'd get back more than your five minutes of replay delay if you enforced pre-existing rules designed to keep the game moving quickly. 

    2.  What is the argument against electronic ball/strike calling?  It's as faster than human umpiring and it moves error rates to ~ zero.  Actually, it would speed up games, because it would eliminate every instance of arguing over balls and strikes.  Quicker games, more accuracy.  Why not do this again?

  16. isaac_spaceman11:44 AM

    "It's faster," not "It's as faster"

  17. Limit replay to safe/out, hr/not hr, fair/foul.  Limit # of uses per game, and strictly enforce other timing rules.

  18. I don't think an honest mistake on a judgement call is the same thing as an essentially rigged system where superstars get more calls, certain players are discriminated against because the refs don't like them, etc.

  19. Another aspect here that's worth mentioning is Joyce's words and grace after he'd seen the replay. No beating around the bush, no hiding behind anything, no avoidance. He admitted his mistake and apologized for it. Posnanski could mention that to his daughters as well.

  20. isaac_spaceman1:04 PM

    You assume that errors are honest mistakes on judgment calls.  Conventional wisdom, and I have no idea whether it's correct or not, is that superstar pitchers get larger strike zones than everybody else.  And there are umps who have quick trigger fingers with players who have reputations.   

  21. Jenn.1:41 PM

    Linda Holmes has an interesting take on the situation, with a thoughtful windup and then this: 

    I don't think instant replay, on a limited basis, is a bad idea. But before you get too OMG INJUSTICE MUST BE REPAIRED REVERSE REVERSE FIX IT FIX IT, keep in mind that baseball statistics are full of things that are just as unrelated to the pitcher's performance as an umpire's mistake undoing a perfect game. One of them is a third-base error undoing a perfect game. "Justice" does not demand a different result. "Justice" really has very, very little to do with baseball statistics to begin with. Fairness — yes, there's fairness. But fairness doesn't demand that all opportunities for other people's mistakes to affect the outcome be removed, because they are everywhere.

  22. J. Bowman3:15 PM

    That would seem to cover everything besides ball/strike. And, of course, pine tar/no pine tar.

  23. J. Bowman3:19 PM

    Yeah, but she thinks Russell should have lost Survivor, based solely on the fact that almost no one voted for him.

  24. I'm all for enforcing the timing rules.  I haven't watched a baseball game on tv since Game 5 of the 2005 NLCS because it's unbearably slow, and 25-year-old Ted would be shocked at that 4.3 year lapse.  I'd even be okay with combining enforcement of the timing rules with a one-shot-per-team instant replay safe/out fair/foul hr/not-hr rule. 

    I've been for electronic ball/strike calling for at least a decade, probably longer, but the archives are down so I can't check.

    But one of the great things about baseball is that decisions are instantaneous.  Very rare that you see the on-field referee-by-committee consultations that you see one play in five in football.  I'm sure there are blown calls, but over the course of a 162-game season with 80 or so unique batter-pitcher confrontations a game, the blown calls average out, and any team that misses out on a wildcard because of a blown call didn't really belong in the playoffs in the first place.  In football, by contrast, there are only 16 games, and individual plays can be much more significant to the playoff implications because of the small-sample issues.

  25. Separately:

  26. J. Bowman5:01 PM

    1. If the players are unduly delaying the game, then fix it for the sake of fixing it; don't fix it for the sake of adding in a different delay.
    2. As you mentioned earlier, the Umpires Union is the major hurdle for Questec. I do wonder how the players feel about it - one assumes it will be bad for approximately 50% of them, at least in the short term.

    Anyway, the problem with instituting replay isn't overturning the egregious calls. It's the close plays, and there are a lot more of those. If you start looking at every bang-bang play at first, that's adding another 30 minutes to the game. If you decide to limit the reviews, then you're just moving the line, and now there will be blown calls that don't get overturned because the team was out of challenges or it happened in the sixth inning instead of the ninth or it just wasn't quite bad enough.

  27. isaac_spaceman5:44 PM

    It's not Questec, it's Pitch-f/x, which is infinitely better.  Unlike Ted, I would not have been in favor of electronic ball-strike calling ten years ago.  I think the change came about three years ago.  The ump union problem is not insurmountable.  There would be no reduction in the labor force, since you still need someone to call plays at the plate, balks, check-swing strikes, foul tips, etc.  But I suspect the umps' union would still oppose on principle, because the umps would lose some of their control over the game, which is exactly the problem we're trying to solve. 

    The notion that blown calls would average out is a guess.  It is neither supported nor refuted by data.  For the most part, they might average out, but you would expect a couple of outliers every year.  It is not true that a team that misses the wild card (or a division title) by only a blown call didn't deserve to make it -- that's your subjective judgment, with which I imagine a lot of people would disagree.  And if you think of the utility people get from watching sports, my own pet theory, perhaps no better but certainly no worse than yours, is that the fan disutility from losing a game because of a blown call is greater than the utility from winning a game because of a blown call.  In other words, by definition the blown calls can't all balance out, because the hard losses are heavier than the embarrassing wins. 

  28. The Seattle-Minnesota example you give is an extreme case.  Blown safe/out or fair/foul calls happen, what, once every ten games?  And they make a difference in scoring half the time, maybe?  And a third of games are one-run or extra-inning games.  So, with all these optimistic guestimates, we're talking something that affects the outcome of one in 120 games ex ante -- and the postseason maybe one year in hundred, though I have yet to hear anyone identify any blown calls that made a difference in baseball history beyond the 1955 and 1985 World Series--and it's far from clear that the call was blown (rather than controversially close) in 1955.  And frankly, I don't believe in the wildcard at all, so I have little sympathy for any team who loses out on the wildcard because of bad luck in blown calls not averaging out for them.

    Now, the ball-strike problem is a different issue.

  29. "<span>The notion that blown calls would average out is a guess."  I would say it's more of a null hypothesis.  There's no reason ex ante to think that blown calls benefit one type of team over another.  A team in a hitters' ballpark or gets a lot of men on base or that does a lot of aggressive baserunning might experience more blown calls over the course of a season, but, again, half those blown calls will go for them and half against. </span>

  30. A .3125 OBP from the 27th hitter sounds reasonable against an average pitcher.  But you have to account for the Bayesian fact that this is against a pitcher who has retired the previous 26 batters, suggesting one or more of an umpire with a big strike zone, a pitchers' park, and, most importantly and most likely, an above-average-to-great pitcher who is likely pitching at the top of his game.

    That said, small sample size.

  31. isaac_spaceman9:00 PM

    I'm not comfortable with any of your numbers, since they appear to be just made up, and I know that in your day job you would never accept that from someone (friend or foe).  I think blown calls happen more often than that, but neither of us has any evidence.  Anyway, let's live with your assumptions for the sake of this discussion.  Over almost 2500 regular-season games, you're talking about affecting the outcome of 20 games.  That means you're affecting 40 outcomes in the standings -- adding 20 wins here and there and subtracting 20 wins there and here.  With a normal distribution, you'd reasonably expect that for some teams, those games would even out, but every year or two you would have a team that added two wins and a team that added two losses (occasionally, even more).  With four-to-six team divisions, the changes are pretty good that every five years or so one or two of those teams would be first or second in their division, and you could see how a two-game difference will, within a reasonably short period of time, change the order of finish in a division.  And with only six divisions, within a longer (but still not that long) period of time, you'd get the team who benefited the most from the blown calls in the same division as the team that benefited the least, for as much as a four (or more) game swing.  So your assumptions lead to the conclusion that blown calls do affect divisional championships (not just wild cards).  If you change your assumptions, we would have to conclude that they affect divisional champtionships less often, but statistically, we'd still be confident that it would still happen. 

    In my opinion, that's BS.  I'll take the extra two-to-five minutes (the rules would say two, but we know from the NFL that it runs closer to five) to review close calls every five games (assuming that half the reviews would be of close calls that turned out not to be blown). 

  32. So once in 30 years there will be a four game swing within a single division involving the first-place team, and in a smaller fraction of the time, that four-game-swing will mean a would-be division-winner doesn't even make the wildcard. (What percentage of non-wildcard teams finish within four games of first place?)

    Part of the reason this concerns me a lot less than you is because you seem to be unaware how much randomness is already built into the system: the issue of blown calls is a tiny fraction of existing statistical noise.  If you take a platonic 90-win team and a platonic 81-win team, and have them play 100 seasons, the 81-win team is going to finish ahead of the 90-win team 30 or so times on average.  Turns out that even 162 games is a pretty small sample, and there's no reason to feel confident that the team at the top of the division is the best team in the division.  Add the fact that the playoffs are essentially coin-flips, but baseball insists on three levels of playoffs, and, yeah, the benefits vs. costs of an extra five minutes a game are a lot more ambiguous.  Especially since Tony LaRussa is going to figure out how to game instant-replay to give him extra time to warm up a relief pitcher and then every other manager will similarly game the system.  Especially since any reasonable IR system will limit challenges and thus won't cover every single blown call so we have little way of knowing how much improvement we're actually going to have.

  33. J. Bowman10:25 AM

    You're also looking at a pitcher who has likely thrown close to or over 100 pitches, against either a hitter (admittedly, the #9 hitter) who is facing him for the third time that day, or a pinch-hitter who would, presumably be chosen for a platoon advantage (perhaps not, but there's a depth to that decision I don't want to get into). I still think a ~30% success rate for the hitter (which is about 10% below the MLB average for last year) is reasonable.
    However, small sample size, to be sure; with only 32 plate appearances, we've seen that OBP go from .310 to .300 to .290 to .3125 in the last two months. My suggestion is that we cryogenically freeze ourselves, and take this up again in 2210.