By In Stuff

Analytics Guys and Heart Attacks

I’ve written this before: Bob Ryan is one of my favorite people on earth. Love his energy, his unflagging enthusiasm for sport, his storytelling … few things make me happier than seeing Bob on the road and just talking sports and life with him. Bob’s just fantastic.

Of course, he did tweet this:

Here are seven words I wish “the Analytics guys” could get put up on a billboard somewhere: “NEVER PUT FAITH IN SMALL SAMPLE SIZES.” If there is any one idea that is at the center of the enormous and conflicting and untidy world of sabermetrics, I suspect it would be those seven words NEVER … PUT … FAITH … IN SMALL … SAMPLE … SIZE. Those words are the gospel, they are the dogma, they are the core principle.

Heart attacks? Come on. No self-respecting Analytics guy would put much faith at all in the reliability of one at-bat or one game or one series. What Cardinals manager Mike Matheny did with Michael Wacha — sitting him for 20 days and then putting him in in the ninth inning of a tied lose-and-go-home game with lefties at the plate — defied every bit of baseball logic I know of, sabermetric, traditional or otherwise. But it absolutely COULD have worked. Billy Butler has stolen a base this postseason — anything CAN work.

Small sample size. Bill James has railed on this. Tom Tango has railed on this. Rob Neyer has railed on this. Joe Sheehan has railed on this. Baseball Prospectus has railed on this. Fangraphs has railed on this. Those aren’t Analytics guys telling you that this guy has hit that guy well because he’s gone 2 for 6 with a homer. Those aren’t Analytics guys telling you that you can expect a scoreless inning because a pitcher has now allowed a run in his previous three appearances. Those aren’t Analytics guys telling you that the best teams or the best managers win the playoffs.

Yes it’s 100 percent true that many Analytics guys — including some of us who aren’t smart enough to have the title — find many of Ned Yost’s decisions to be curious. It’s true that Analytics guys have probably underestimated the things Yost does well. In the Dentyne challenge more than four out of five Analytics guys surveyed would take Bruce Bochy over Ned Yost to manage their teams, I suspect there isn’t one who would be even mildly surprised if the Yost-led Royals beat the Bochy-led Giants in a seven-game series. Chuck Tanner beat Earl Weaver in a seven-game series, for crying out loud.

Small sample size. Now, look, if Ned Yost and Bruce Bochy and Buck Showalter were playing 10 seasons worth of Strat-o-Matic with the same teams, and Yost won that league, well, I’m not saying there would be heart-attacks for the Analytics guys, but there might be some mild indigestion.

Print Friendly, PDF & Email

86 Responses to Analytics Guys and Heart Attacks

  1. DjangoZ says:

    Thank you for this, Joe.

    The regular season is long enough to be a meaningful sample size and we place almost no emphasis on who won the most games during it.

    Playoff series are so short that they are not statistically meaningful and they are all we ever talk about after the season.

    One more reason to follow the EPL and not MLB.

    • Tampa Mike says:

      So if something isn’t statistically meaningful then it isn’t worth talking about?
      That doesn’t make any sense. Why bother with sports at all then if all you want to do is evaluate statistics?

      • buddaley says:

        No, that is not the point at all. Anything may be worth talking about-1 pitch, 1 AB, anything. And it is always fun for many people to fool around with trivia. No harm in that.

        The issue is not whether it is worth talking about but whether it is useful in trying to analyze the reality. I may curse my team’s pitcher for giving up a game winning home run, and I may remember and talk about it a lot. But if the entirety of his career indicates that he is not homer prone, I would be pretty silly to call him a gopher baller and say the manager should not trust him in key situations.

    • flcounselor says:

      Pardon my ignorance, but what is an EPL? And why would I want to follow it?

      I guess I’m not up on jargon as much as I probably should be.

      • The English Premier League. Soccer in England (and Wales, sometimes). 36-game seasons to determine the title, no playoffs. And you should follow it because it’s top-level soccer, and top-level soccer is awesome.

        • David in NYC says:

          No, the EPL season is 38 games — 2 games (home and away) against each of the other 19 teams in the Premier League. There are “playoffs” in the sense of knockout tournaments (the FA Cup being the most obvious example) where you go home when you lose. Finishing first in the league is most definitely more important than any knockout tournament.

          And yes, watch it, this is soccer at its highest level (well, I suppose the UEFA Champions League is higher since it’s international).

          Not to mention the fact that the announcers all speak English correctly, know what they’re talking about, avoid overstating the obvious, and — most importantly — know when to shut up so we can just watch the game. Ever notice how there is never more than 2 seconds of dead air in any American sports broadcast (games announced by Vin Scully are not included), so much so that you can’t even hear yourself think? The most common thing I yell at my TV when watching American sports is “STFU!” (but I don’t use initials).

          • Loztralia says:

            “The announcers all speak English correctly, know what they’re talking about, avoid overstating the obvious, and — most importantly — know when to shut up so we can just watch the game”

            I suspect you’ve only seen a small sample size of Premiership commentators. Jamie Redknapp alone disproves every single one of these claims.

        • whatisinthere says:

          watching a soccer game is a sure sign that you’ve given up on wanting to be entertained. i bet you could cure insomnia by prescribing the watching of soccer games to insomnia patients. they’d fall asleep within seconds.

          • Ron B says:

            Watching anything that you don’t understand will put you to sleep. I’m a baseball fan, but most people outside the US and some within find nothing more boring than baseball, and to be fair, baseball doesn’t have a heck of a lot of action. But it’s entertaining if you understand the intricacies.

        • Chris says:

          Love the idea of regular seasons actually determining the best teams. Used to be that way in baseball. Unfortunately America doesn’t agree. We live in a 1 game/pitch/pass/drive/moment/ determines all type of society. See: Bayless, Skip.

    • Loztralia says:

      The sample size in the Premiership is certainly large enough to prove that one of the very small group of teams with the most money always wins.

    • whatisinthere says:

      there’s never a reason to follow soccer over baseball. in fact, there’s never a reason to follow soccer over watching paint dry. i can’t think of something more mind-deadening than watching a soccer game. it’s making me sleepy just thinking about it.

      • You’re over 40 then and probably over 50. The younger set greatly prefers soccer and considers baseball very boring…. Which, unfortunately as games have gotten longer and offense has disappeared, is quite true…. Excellent playoff baseball aside.

  2. Brent says:

    Using large sample sizes, the Royals are (so far) a 97 win team in 170 games and the Orioles are a 99 win team in 169 games. Not sure I see a lot of room there between them.

  3. tombando says:

    Larger sample Size: whining about the Myers for Shields trade. Golly gee there Joe.

  4. MCD says:

    I have no problem with Ryan’s tweet as a short, off-the-cuff joke. However, if he appeared on a panel show or published a lengthy column claiming that a Yost victory “shocked” the analytics guys, then I would say he was missing the point.

    • DJ MC says:

      The problem is, this is essentially his default belief on the subject. Just because it is appearing in 140 characters instead of a ten-second sound bite or a 400-word column doesn’t make it less meaningful.

  5. Dave says:

    I’d also add another “beware of” statement, but I’m not quite as terse as the seven words. Beware of applying a large sample average to a specific situation. Take the analytic insight that on average the bunting game costs runs. While that is true from the numbers averaged over all teams and all players and all situations, that likely is quite wrong for a particular team and particular players in particular situations. The Royals for example as a team don’t have the slugging percentage of other teams. Thus, it may well be for them that getting a guy in from first with no outs is more difficult than getting a guy in from second with one out. And that could well be for a particular part of any team’s lineup coming up than for another part of that same lineup.

    Average analytic insights are great, but teams are not made up entirely of average players, nor are all teams average.

    • I agree that averages can be very misleading. Even a batting average, or OBP a can be misleading if it was built up in red hot April and May or against weak teams. It would actually probably be better to rate overall performance by tracking the percentage of games with at least one hit, or two hits, or with one extra base hit etc. That would factor out hot streaks or stats built up by slaughtering a few bad pitchers.

      • A reasonable-enough sounding comment, but still I might suggest not overthinking things, and say that the flipside of not paying undue attention to small sample sizes is that you can safely make inferences from large ones. Specific to your comment here, Jose Altuve led the American League in hitting this year, and also led the AL in two-hit games. And three hit games. And in hitting during the month of June.

        And not to flog a dead horse, but I can’t help myself: the league leader in RBIs will more times than not be a leader in RBI percentage.

  6. Tampa Mike says:

    I think the analyst community has gone off the deep end. Before Moneyball, teams didn’t look at statistics enough and were disregarding valuable information. At this point they have gone overboard and don’t look at anything but statistics. I appreciate the hidden insight it can give you, but it isn’t the end all be all just like “the eye test” isn’t either. Enough already. It’s a game.

    I know, I know… I’m an idiot. I have a flawed understanding of the game. I’m too stupid to understand how bad the Shields/Myers trade was. What I do know is I LOVE this Royals team and they are a helluva lot of fun to watch.

    • flcounselor says:

      Tampa Mike, what makes you think the criticism of that trade had anything to do with analytics? It was about giving up a top prospect under team control for six years along with several other young players for a starter who would remain only two years and another pitcher.

      Fortunately it has worked out, but two years ago there was a heck of a risk that it might not have.

      Either way, that trade is completely irrelevant to whether or not teams should use statisticsin their analysis. That you would even make the argument causes one to suspect that you have no grasp on what statistical analysis is, let alone its proper usage.

      While you think you are slaying sabermetrics with your example, you are simply revealing your ignorance of it.

      • Tampa Mike says:

        The trade was something of a side note. It was absolutely risky, but it was a move that had to be made. This team needed a leader for the pitching staff much more than another outfielder. Those kind of pitchers are very expensive. We can speculate all day about what free agents they could have pitched up instead, but we don’t know if any of them would have signed with the Royals. Supposedly Edwin Jackson refused.

        I never said I was “slaying sabermetrics”, nor do I intend to. I absolutely think that teams should use statistical analysis. I think any team that doesn’t is foolish. But any team that uses ONLY statistical analysis is also foolish. It’s still a game played by people with emotions that effect their performance.

        • invitro says:

          “But any team that uses ONLY statistical analysis is also foolish.”

          Which teams do you think are using only statistical analysis? You must have some teams in mind, or else there would be no point in what you said.

    • DJ MC says:

      Generally, if you think everyone around you has gone crazy, their mental states probably aren’t the ones with a problem.

  7. Alan Cunnigham says:


    Isn’t it Trident (and not Dentyne) that publicized that “4 out of 5 dentists surveyed chose sugarless gum for their patients that chew gum?


    (By the way, as a math teacher and baseball coach, I agree whole-heartedly with the sample size dilemma).

  8. Michael Green says:

    A couple of points.

    One, The Great Boswell, as in Tom of the washington post, which used to be a newspaper, has said, “Judge slowly. Even more slowly than that.” Bob Ryan is a great sportswriter, but there’s an affliction in the Boston sportswriting corps that leads them to judge foolishly. That doesn’t mean Ned Yost is a good or bad manager, but it does mean that looking for a simple answer means judging too fast.

    Related to that, The Vin has said that Al Campanis used to quote a Branch Rickey dictum: do not judge your team until it has played every other team in the league. That was possible when he said it in the 1940s. Not today. But I am reminded of the large numbers of Mets fans who called in to WFAN to demand Davey Johnson’s firing when his team lost on opening day one year. Deciding who the best manager is simply on the basis of one event or series is stupid.

    Now, I wasn’t convinced of Don Mattingly’s talents either way, but the NLDS convinced me that he should not be allowed to make a single strategic decision. I watched Tommy Lasorda for 20 years and figured out that the same thing was true, except in post-season, when the rules go out the window. That doesn’t make me an expert. But it appears to make me a better judge of things than Bob Ryan, which is scary.

    • Bill Caffrey says:

      I call shenanigans. First of all, the only year you could be thinking of is 1990, because, other than the very first game of his tenure with the Mets in 1984 (when WFAN didn’t exist) that’s the only time the Mets lost on Opening Day under Johnson.

      Second of all, any talk of firing Johnson after opening day in 1990 was likely to be based on the soul-crushing way in which they lost the 1988 NLCS (if only Johnson had brought in Randy Myers to pitch the 9th inning of Game 4 or even mid-inning to face Mike Scioscia), followed by a disappointing 87-win season in 1989 (down from 100 in ’88 and their lowest win total under Johnson) and the sense of a window closing (Carter and Hernandez were gone; Gooden was clearly not the same pitcher; Strawberry was in the last year of his contract). It was hardly the result of a single opening day loss.

      And in fact, talk of firing Johnson was swirling throughout the offseason between 1989 and 1990 and, 42 games into the 1990 season with the Mets sitting at 20-22, they did in fact fire him.

      Lastly, no matter how many people called into WFAN on the day you’re thinking of, it was not “large numbers.” Playing in NYC, the Mets literally have millions of fans. The vast majority of those millions have never called a sports-talk radio show. If a few dozen called for Johnson’s firing after their opening day loss in 1990, that would not be large numbers, nor would it be a representative sample of fans.

      • Michael Green says:

        If shenanigans answers, tell him I said hello. I was listening that day, and almost every one of the fans who called into the post-game talk show said Johnson should be fired. Why, Howie Rose asked (I am almost certain it was Howie who then hosted that show). Because, each fan replied, he lost today. Some offered other reasons. But all of them came back to that.

        • Recollections from that long ago are very unreliable. I remember clearly that Sandy Alomar Sr., a punch less infielder for the Angels, hit 2 HRs in a high scoring doubleheader against the Senators. I told that story often. Recently I checked the game logs on BBR and immediately knew my story was in trouble because Alomar only had two seasons with the Angels where he had as many as two HRs. I found the game….not a doubleheader….. Which WAS an 11-10 slugfest against the Senators in which Alomar did hit a HR. Not a doubleheader and only one HR. So, there were elements of truth in my recollection. I didn’t make it ALL up. But the years had muddied up my recollections and turned them into something that never happened. BBR game logs don’t lie.

          • Michael Green says:

            And I’m not saying I can’t be wrong. But I remember that loss and I remember the phone calls. And I remember how much more pleasure I took from recalling the Dodgers beating the Mets in the 1988 NLCS.

          • Bill Caffrey says:

            Howie Rose did host Mets Extra in those days, which was the pre and post-game show for Mets broadcasts on WFAN. The show was not more than an hour long and included giving the scores of games from around the league. And lots of commercials. It also usually included a post-game interview.

            It did include call-ins, but it was unlikely that there were more than 12-15 callers for a given show. But let’s say there were 20. By your own acknowledgment, “some offered other reasons.” I don’t doubt the callers were saying JOhnson should be fired, but I do doubt any of them giving that day’s loss as a reason. Again, there was talk of firing Davey all offseason because of their poor 1989.

            I think it’s much more likely that one of the callers gave that day’s loss as a primary reason why Johnson should be fired, and it struck you strongly because that was such a stupid opinion for a fan to have. And so it stood out among all the callers to you and over 25 years it has morphed into “large numbers” of fans and “all of them” coming back to that day’s loss as the reason why he should be fired.

          • Michael Green says:

            For some reason, I cannot reply to the last comment below, so I guess it goes here. I don’t want the commenter to go away unhappy, so, here goes: I hereby acknowledge that anything I have ever believed or remembered in my life is wrong. Thank you.

    • Bob Lince says:

      “…Boswell … has said, ‘Judge slowly. Even more slowly than that.'”

      Would help if the Heisman Trophy voters would learn that lesson.

  9. I think the bigger point is one of statistical significance. This is a function not just of the size of the sample, but of the actual impact that the independent variable (the manager in this case) has on the dependent variable (the outcome of the game, or series, in this case). It’s the fact that managers have such little influence on the outcome of the game in the first place (which is a different topic altogether I realize, but which is also generally accepted) that such a large sample is needed.

    If you wanted to show the impact of starting 12 people on a team’s defense vs. 9 people, a smaller sample size would be fine. If you want to prove that Ned Yost’s appetite for sacrifice bunts makes him a worse manager than Buck and Bochy….well, then, to Joe’s point you’re going to need more than a seven game series.

    • Thomas says:

      You might wonder about the premise of the question, give that Ned’s team had 52 sacrifice bunts in the full 162 games, while Buck’s team had….51. If you think managerial skill declines that sharply with one bunt (and without considering whether the bunts were otherwise a good strategy), you are probably someone who routinely castigates Ned.

      • DJ MC says:

        Actually, according to Baseball Reference the Orioles had 35 non-pitcher sacrifice bunts in 2014, compared to 33 by the Royals. However, that’s not the only thing to look at.

        Of those 35 sacrifice bunts, the Orioles bunted 15 times in the first six innings, and 20 in innings 7-9 plus extras. So the majority of the sacrifices came late in the game, when you are likely to only be playing for a run or two.

        In the case of the Royals, 12 of the 33 came in the first three innings, with the remaining 19 split between 4-6 (7), 7-9 (9) and extras (5).

        So while officially Showalter bunted more overall, he was bunting in situations that called for it. Yost was bunting in the third inning (a full 7 times).

        • Thomas says:

          So the difference between the best and worst manager is 4 bunts, right? No need to consider WPA, no need to compare the last-in-the-majors power of the Royals to the power in the O’s lineup.

          Somewhere along the way analytics became as dumb as what preceded it. Yost is the best evidence for that.

      • The bunting comment was meant to be representative of the difference between the two, not the single distinguishing characteristic; I think you missed my point.

  10. Dave B says:

    The mere mention of Chuck Tanner put my brain into an intense early-80s memory of the smell, taste, and powdery texture of the stick of gum inside those 1984-ish Topps baseball card packs.

    And then I got very sad. I’m old. Boo.

  11. NevadaMark says:

    Chuck Tanner? The Pirates got 81(!) hits in the 79 Series. That will cover a lot of managerial inadequacies. Holding the O’s to two runs over the last 27 innings also helps.

  12. MisterMJ says:

    Baseball playoffs is ridiculous, considering a 162-game season that stretches half a year and then there’s a SINGLE play-in game (for wild-cards) followed by a best of five and best of seven. The Royals were 4 runs down and 4 outs away from losing to the A’s and then go on to win EIGHT consecutive games. They also happened to lose 13 of 19 games to the Tigers during the regular season. Is more teams worth it? From a ratings, revenue generation, and fandom perspective, I guess it does. More the merrier. But it doesn’t mean the BEST overall team wins every year. In a game like baseball, I have no doubts that if a 75-win team was given access to the playoffs, they’d make the World Series every now and then.

    The EPL is a good comp. The team with the most points after 38 games wins the title. It’s very clear-cut. Sure, sometimes a team stretches a lead so there’s really no suspense going into the final month, but all in all, not a bad way to go. But as only a handful of teams spend egregious amounts of money every single year, there’s the FA Cup competition (knock-out). But lower teams like Wigan, Portsmouth, Stoke City, Cardiff City, and West Ham have made the finals in recent years. Great for those fan bases but it goes to show how unpredictable a “playoff” can be in sports.

    So baseball rewards the teams that play well in the end – and whether it be opportunistic, luck, or randomness, all those variables weigh heavily.

  13. Dr. G says:

    I find it funny that writers like Bob Ryan assume the position of people who enjoy analytics/sabermetrics/etc. is basically, “the numbers are gospel.”
    I can’t even remember when I’ve read or heard a sabermetrics/analytics/etc. writer/blogger/etc. make that kind of umbrella statement. I seem to find the opposite, such as Dave Cameron on Fangraphs, particularly when he’s on Carson Cistulli’s delightful podcast each week (a weekly Monday appearance), when he often goes out of his way to note that sabermetrics and scouting go hand-in-hand; neither one is the “only answer.”
    So I hear the old guard like Ryan pretend their “rivals” in the sabermetric community take only one view point, but clearly those are the people that have the more diversified and open-minded approach.
    It’s just sad that Ryan (and others) has to be so polarizing… sadder still that some of his readership feed off of that.
    And one final note… my personal experience is that sabermetrics have enhanced my enjoyment of the game and “small sample size” moments. In this way, I can understand Travis Ishikawa, for instance, his skills, talents and weaknesses on a larger level and then find even more drama when he delivers a walk-off home run to win a pennant. It’s marvelous and only made better, for me, with sabermetrics.

  14. rucksack says:

    Is this in reference to the Odorizzi-Davis trade? I mean, let’s focus on the guys in this trade producing the most value.

    If you want to focus on Myers and Shields, though, that seems to me a small sample issue in that Myers season being a bust is likely a blip amidst a fruitful career and Shields’ value in 2014 relative to options that didn’t require getting rid of an asset like Myers was probably attainable through other means.

    That trade has worked out well THIS YEAR (small sample). And it may produce extreme (but still small sample) value in the form of a World Series win. But large sample value is still likely to go to the Rays on this.

    I”m not saying that anyone should be (or is still) whining at this point but people whining about expected value at the time of the trade were being smart. It is small sample size making this trade look good. I love that it has worked out, but sometimes you flip a coin ten times and get ten heads. That doesn’t mean it was smart to bet on it happening.

  15. rucksack says:

    This is @tombando. Not sure why the reply link didn’t nest the response correctly.

  16. MikeN says:

    Another backlash against Ned Yost. Now it is just luck i his team wins, because the numbers say his team is terrible and the manager is worse. You say it is just a small sample size. How about the sample size of postseasons going back 40 years. How many won 8 games in a row, or even six or seven? Now consider that the Royals are a weaker team, appearing in the play-in game. If Yost is a terrible manager, takes their 40% chance and turns it into even lower odds for each game. Now given that they won 8 games in a row this suggests that perhaps Yost is a good manager who has boosted this team’s ability to win substantially.

    • nightfly says:

      Luckily it’s not hard to look that up with Baseball Reference.

      A cursory search indicates that in 1976, the Reds swept the playoffs, 3-0 vs. Pittsburgh and 4-0 vs. New York. (It’s 8 in a row for the Big Red Machine if you include game seven from 1975’s Series, and they had substantially the same club.)

      In 1998, the Yankees went 11-2, and in ’99 they went 11-1. They built a 12-game win streak in the playoffs across those two seasons.

      The 2004 Red Sox famously won the final four games of the pennant series when trailing 0-3, then swept the Series – 8 straight wins.

      The 2005 White Sox also won eight straight games on their way to an 11-1 overall record.

      The 2007 Colorado Rockies, like these Royals, swept their way into the World Series, only they didn’t face a play-in game so they only won 7 straight.

      It’s an achievement for sure, but it’s not unheard of, and I don’t think that Yost really has suddenly proven he’s a miracle worker. For one thing, it’s pretty darned insulting to his players to think that they couldn’t possibly prevail unless Ned Yost was magicking them past opponents. For another, the solution to Yost-bashing is not Yost-worshipping. He can simply be what he has proved to be over the course of his career: a guy who makes curious gut-based decisions, follows a Script for Winning Ballgames that sometimes will work in spite of itself, and is riding a hot team in perhaps the only way they can maximize their chances of being successful. They’ve hit homers over these 8 games at a better clip than they did in the regular season, and that may well be enough to have fueled the streak.

      God bless baseball, at any rate.

  17. Frog says:

    Small sample size is a killer and so often not appreciated. When I trying to explain the importance of sample size I use the first at bat of a rookie – if you take that single at bat and extrapolate it then he will either be a 1.000 hitter or a 0.000 hitter. Most people quickly get the point that it is absurd and they need to see more to get a proper idea.

    Also, people so often look at “averages”, when they really need to be looking at medians and modes. I think playing cards would be good training for managers.

  18. ralf says:

    Although your blog is entertaining, I often dont see why the questions/ problems you pound on could have any relevance to enjoy the game with a greater understanding.
    Evaluating individual skill sets based on statistics should not be mingled with actually staying as a team in a competition with another team. Unimportant wether in a single game, a series or a pennant. This quote is not worth a notice.
    Statistics are a tool however advanced you may title it. And a tool itself has no value without the proper use. For determining the propable production of a single player – very useful, for understanding how long a certain battery runs usually – even more excellent, but for understanding or even predicting complex structures like a baseball game, or even a series of them – not controllable. For creating a happy family live or a love affair – deadly poisonous. Thanks god, otherwise nobody would like to see a baseball game anymore, nobody would stick to his family or risk a date.

    And another thing I do not like, enjoy very much. But which was present for a long time in your blogs, especially in regards how you treat the Royals:
    How a game is decided, How a world series is won, how you qualify for a postseason is a given situation. Any team going through this successfully has rightfully accomplished it. It deserves respect. It does not mean, no of course not, to discuss single aspects. But anything even slighly hinting on luck or undeserved should be attacked fiercely.
    Wether you understand and like the offensive ways teams like the Giants or the Royals of the last two years create (although they had lot a lot longballs in the postseason), can not matter as long they are winning their games. And in the present run of the team you actually are attached very much, you hopefully will find time to reflect, why this was a surprise for you. May be you have missed something.

    I’m looking forward wether the Royals will be able to take on the incredible tense battles the Giants had in the NL series.
    I am looking forward to more moments like the slightly surprised grin of Bochy after Duffy had scored on the wild pitch from second base. A game finally they lost later.
    I look forward to this series because I like to see good, intense, fair baseball games. Nothing more.


  19. Joe Bonomo says:

    Is it accurate to say/complain that baseball reward teams that play well at he end? Historically, there are last-pleace teams who go on 12- or 15-game winning streaks in August/September but who have no or little chance of making the playoffs. A team has to play well throughout the long season to have a chance at the playoffs, WC or no.

  20. likedoohan says:

    Anyone commenting on sports in a media position should be required to take a basic statistics course. The lack of understanding of even basic principles is a real detriment to any discussion of, well, just about anything. Many people believe one example can totally disprove any theory. It was cold last night, so climate change is bunk.

    • MikeN says:

      It was Einstein that said one example can disprove any theory. There was a book written by 100 authors arguing against his theories. His response was, if my theory is wrong, one author is enough.

  21. G says:

    I find it odd that in these situations we go to the “monkeys on typewriters” analogy and start expecting someone to write Shakespeare.

    Fact is–these managers are doing the job they have, not the job created in the imagination of an analytics writer or a computer program. They manage 162, then 1-29 games in the playoffs. If Tommy Lasorda–of all people–can be considered a poor manager with 8 playoff appearances in 20 years, then I’m not sure how you get there from here.

    Anyone who has been a leader in “high leverage” situations knows that the data is the easy part. The hard part is surveying the people you’re working with–with all the challenges and opportunities actual human personalities present, and even harder is having the guts to make the call and stick with it. An MLB manager can’t call SABR and ask for a position paper on the RHP in the bullpen. If he’s prepared with the data and prepared on the human side, it is up to him and him alone to make a tough decision in a matter of minutes, not hours or days.

    Data analysis has done enormous things in terms of how we understand what we think we know, but Napoleon probably has more to say about seizing the moment and understanding the situation than most baseball writers.

  22. EnzoHernandez11 says:

    One of my favorite examples of the perils of generalizing from small samples is the case of Barry Bonds. Had Barry retired in 2000 (as a no-doubt first-ballot Hall of Famer), he would have spent the rest of his life listening to idiots diminish his career by carping about he always choked in October. As of 2000, Barry had appeared in the post-season five times and had an OPS of under .700 to show for it.

    Then came 2002.

    By the time he actually retired, Barry had a post-season OPS over .900, which, as it happens, exceeds that of Mr. October and Mr. November. But Barry wasn’t smart enough to hit three home runs in the deciding game of a New York-L.A. World Series back when there were only three TV channels.

    (And, no, I don’t want to talk about steroids. My point is that adding just a few more observations to a small sample can reveal a completely different “truth”.)

  23. John Leavy says:

    Bob Ryan would have been a lot closer to the truth had he predicted Ken Tremendous would have a heart attack if Brian Sabean won his 3rd World Series while Billy Beane and Paul DePodesta were still waiting for their first!

  24. JTF says:

    OK. A couple of thoughts here.

    First, Bob Ryan’s comment seems to be about the narrative that Yost and the Royals winning creates, not a definitive analysis of this year’s (possible) champion. Yost winning the Series would offer a strong anecdotal suggestion that would negatively impact analytics. This seems clearly true. The point is still to win the World Series. And the Royals put themselves in a position to do that over a fairly large sample of the regular season.

    Here’s the other thing. At what point does a sample size become sufficiently large when you keep getting the same result?

    The Royals are 8-0 in the playoffs this year. You can flip a coin and not be surprised to get 5-3 or 6-2 split even though you know over a larger sample you’ll end up at about 50/50. But if you get heads 8 straight times (less than 1% chance), don’t you start to wonder about the coin?

    Playing an inferior strategy can succeed on occasion or even get a slight majority. However, it’s not likely to make a long run of consistent success. I mean really, what does statistical analysis tell you about eight straight wins over three quality teams at the most crucial time of the year?

    Of course, this is all best explained by the five words that supersede all else:
    Baseball is a funny game.

    • David in NYC says:

      Statistically speaking, it takes 30 observations before analyses like standard deviarion and confidence interval can be used to make a meaningful analysis. In the coin toss example, if the number of heads in 30 flips was between, say, 13 and 17, that would within the range of statistically normal deviation. If you got a result of 20+ heads, that is *probably* not a random result.

      My “favorite” misuse of this is batter v. pitcher stats. If Buster Posey is 4-for-6 lifetime against Michael Wacha, that tells us literally nothing useful. OTOH, if Posey is 20-for-30, that is useful information; 40-for-60 would be even more meaningful. The more observations, the better; this is what causes regression to the mean.

      • JTF says:

        That’s interesting and useful. Thanks. However, it doesn’t quite answer the question as posed. At what point does a uniform result become statistically significant? That is, when you are getting no variation in the results.

        In the example you cite, Posey is 4-6 not 6-6. I don’t think you would need to get 30 consecutive flips resulting in heads before you can make reasonable statistical inferences.

    • MikeN says:

      Exactly. Then throw in the fact that very good games have not done it over 40+ years of opportunities. Only about a dozen times have teams won six in a row, with the 98 Yankees winning 5 and 6 around Pedro.

      This unlikely to happen with an inferior manager of a great team, and no chance for an inferior manager on a mediocre team.

      • invitro says:

        “This unlikely to happen with an inferior manager of a great team, and no chance for an inferior manager on a mediocre team.”

        Why do you think this?

        • MikeN says:

          The fact that even the good teams have not had such winning streaks, though it may have been for lack of opportunity(Torre’s Yankees had only Pedro between them and 11-0). So if you have a 60% chance of victory, over 8 games you are down to 1.7%. A mediocre team has maybe a 50% chance, for a 1 in 256 chance, and of course with the good competition, there odds are lower, and an inferior manager would be even lower still. If you say a 30% chance for each game, you are at .006561 % chance of an 8 game win streak. I think it is likely that Yost is a good manager who has elevated his mediocre team to well above a 60% chance of victory for each game.

          • invitro says:

            “The fact that even the good teams have not had such winning streaks,”

            Many teams have had winning streaks of eight or more playoff games.

    • invitro says:

      “The Royals are 8-0 in the playoffs this year. You can flip a coin and not be surprised to get 5-3 or 6-2 split even though you know over a larger sample you’ll end up at about 50/50. But if you get heads 8 straight times (less than 1% chance), don’t you start to wonder about the coin? ”

      You didn’t flip just one coin. Lots of teams have been in the playoffs, over a hundred of them. What you are doing is flipping a hundred coins, then finding one that flipped 8-0 and saying it must be a strange coin. Is this logical?

      • MikeN says:

        Only here we know that some teams are better than others. When we see coins that give out 5-3, 6-2 we suspect they are better but of course sometimes luck makes an unfair coin go the other way. Now seeing an 8-0 coin, you would naturally suspect it is weighted in favor of heads substantially, especially when none of the other hundred coins have done so.

        • invitro says:

          “Now seeing an 8-0 coin, you would naturally suspect it is weighted in favor of heads substantially,”

          Perhaps it is natural for someone ignorant of probability.

          “especially when none of the other hundred coins have done so.”

          Do you think that no other playoff team has won eight consecutive playoff games?

      • JTF says:

        I think the analogy here is being taken too literally.

        You seem to be suggesting that if you do something enough times, eventually you’re going to get some strange (statistically unlikely) results. This is of course true.

        However, we’re not really talking about coins. And I don’t think it’s sensible to reduce the games to a simple end result of a statistical eventuality. It’s more likely that the Royals are doing something that has actually enhanced their chances of winning to something greater than 50/50. And I think that this is something not being captured by the normal statistical analysis.

        Moreover, I did not say that after an 8-0 run you believe that you have a ‘strange’ coin. However, I think it’s pretty reasonable to start to wonder about it at that point. That is, I would begin to question whether or not the coin really represented a 50/50 chance of either result. Or if instead there was some factor I hadn’t considered that makes heads a greater likelihood.

        And if 8-0 is not sufficient, at what point do you wonder?

  25. KHAZAD says:

    I have been as critical of Ned Yost as anyone over the years. I hate his tendency to bunt early in the game against offensive teams. I once said that the easiest way for an opposing starter to get the first out of the game against the Royals is to allow a leadoff double and Ned will give you an out. There have been alot of inexplicable moves during the season, (some of which have been written about by Joe) which he makes worse with his explanations. (Yostifying!)

    Going into the playoffs, I was worried that Yost would try too hard and make the type of mistakes that would doom the Royals, and he very nearly did that in the wild card game, with his ill fated fake double steal trick play, and using Ventura on one day’s rest.

    He got pretty lucky in that game, as the poor result of the first play led to a hand injury to Oakland’s defensive catcher, which led to the insertion of the catcher that couldn’t throw, which led to the Royals stealing at will, and stealing back the game.

    Since then, though, I think his decisions have been pretty right on through the next 7 games. I thought he outmanaged Scioscia and Showalter, and basically pushed all the right buttons. I might still be worried about him making the big mistake in the World Series, but I will not be surprised if he doesn’t make that mistake.

    Also, there is one thing that is always overlooked by analysts, but is the most important job of a baseball manager. That is getting your team to buy in, keeping their spirits up, keeping them together etc. A baseball manager manages people and different personalities and gets everyone going in the same direction. I don’t think you can watch the Royals without thinking he has done a stellar job of this, and in the end, that might be the most important thing.

  26. MikeN says:

    Basically the moneyballers have elevated stats above all. So when the Royals win it is luck, and if a moneyball team loses it is bad luck. The models say they are right, so everything else is just luck and not skill. Only when a team playing in a moneyball fashion is it skill.
    This is what yields things like deification of Theo Epstein, for saying that JD Drew is a good player because he has a high OPS. I pointed out here that his overall moves were quite shaky, and the Moneyball worshipers responded that it is all part of being a good manager because he knows his team has the resources to cover up for mistakes. The irony is that JD Drew was actually a clutch player.

    I saw the same thing happen two years ago, where people at a blog were criticizing moves made by the Mavericks, and Mark Cuban came in and defended his moves. The responses were that his moves were terrible and even if they won it would be just luck. And this was in response to a guy who is stat heavy.

    • EnzoHernandez11 says:


      I don’t think anyone is saying that the Royals’ run is luck. The Royals are a good team; they were 16 games over .500 and had the fourth best record in the AL. And good pitching and timely hitting–not luck–have produced an eight-game winning streak.

      All we’re saying (or I’m saying, anyway) is that the Royals–regardless of what happens against the Giants–are not the best team in baseball. The regular season proved that. If you replayed the ALDS and ALCS a thousand times, the Royals would probably fall short of the World Series in the majority of cases.

      Think of Villanova in college basketball in 1985. (Do folks remember that one, or am I just old?) They won six in a row, beating Patrick Ewing’s Georgetown, to win the National Championship. It was a great run, and there was nothing lucky about it. They deserved the crown; they absolutely outplayed the Hoyas in the championship game, couldn’t miss a shot. But nobody considered ‘Nova the best team in the country. Hell, they were an 8-seed that had already lost to Georgetown TWICE in the regular season.

      “Moneyball” was simply about taking advantage of strategies that other folks had not yet discovered. Over the course of a long season, those strategies–if they are valid–will give a team a small edge over its opponents. If the team is good enough, that might be the difference between an 84-78 record and making the playoffs. If the team sucks, it doesn’t matter if they play “Moneyball” or small ball or whiffle ball. And as the Angels (not exactly a Moneyball team) can tell you, once you get to the post-season, anything can happen.

      If the Royals sweep the Giants, it will not be a miracle, and it will not be luck. It will simply be one of those rare things that you see only once or twice in a lifetime and get to tell your grandchildren. It will not vindicate Ned Yost as a manager; if you want to vindicate Yost, start with 89 wins over six months.

      • MikeN says:

        Plenty of people have been saying it is luck. More particularly that the Royals are winning despite Ned Yost. Indeed, there has been comments that the Royals would have run away with the division without him.

        If the Royals sweep, it is indeed a vindication of Ned Yost. Even without it, the current run is a vindication. As you say lots of teams fall short and lots of teams do well, if you replay the Royals would come up short many times. But how many times would they win this many games in a row? That’s just it, they are a weaker team than their opponents or perhaps equal. Such winning streaks have not happened in the past, even with better teams. So Ned Yost has clearly elevated his team to the level of one capable of running off such a win streak. It is not possible for a bad or mediocre manager to win so many games in a row even with a very good team.

      • invitro says:

        “I don’t think anyone is saying that the Royals’ run is luck.”

        I don’t know if it is luck, but it’s certainly consistent with luck being the cause.

        “Think of Villanova in college basketball in 1985. (Do folks remember that one, or am I just old?)”

        I remember watching every minute of that game in my bedroom, on a 13-inch black and white TV, opening packs of Topps cards, and cheering on the Novas every play.

        “They won six in a row, beating Patrick Ewing’s Georgetown, to win the National Championship. It was a great run, and there was nothing lucky about it.”

        But it was definitely consistent with luck being the cause. And the improbability of the event is likely overstated. The Big East dominated basketball in 1984-85. And the refs gave Villanova 27 free throws to Georgetown’s 8.

        “If the Royals sweep the Giants, it will not be a miracle, and it will not be luck.”

        It would be consistent with luck being the primary factor.

      • dookie says:

        You mean nobody knew that strong starting pitching was important before Billy Beane came along? Because that’s how his early 2000s teams won.

  27. EnzoHernandez11 says:

    I don’t know. I think that, in general, we greatly overrate the role of the manager. The difference between the best and worst manager is probably no more than a handful of games, if that. Baseball is a sport of limited strategies.

    (Obviously, when I say “worst” manager, I mean the worst plausible manager. If some clown came along that sat all of his best players, put his worst pitchers in the rotation, and gave Prince Fielder the “steal” sign every time he reached first, then, sure, that manager would have a huge impact on his team.)

    It’s all about the players. Joe Torre was, at best, mediocre before he got the Yankees job. Did he learn something in the meantime? Sure, he learned that it’s great to have Jeter, Rivera, Pettitte, and Posada on your team. Was Sparky a great manager? I don’t know: Could anyone have screwed up the Big Red Machine or the Trammell-Whitaker-Gibson-Morris Tigers? Sparky had the best team of the 70s and arguably the best team of the 80s, and he had two rings to show for it. That’s not nothing, but still…

    There is, of course, the question of how well the manager handles the clubhouse, relates to the players, makes them believe they can succeed, etc. I imagine there’s something to that, though it’s impossible to know how much. Dick Williams and Tommy Lasorda related to their players in very different ways; both were successful.

    So, no, I don’t see that the current streak vindicates Ned Yost. But I also don’t think Yost deserves all the vitriol he has received over the years, either. He’s a subpar strategist, but he’s probably an above-average handler of players. To my mind, that makes it a wash, and doesn’t have too much to do with the Royals’ current run.

    (Along those same lines, shouldn’t we let Bruce Bochy get his managerial record above .500 before we nominate him for Cooperstown? And, no, I haven’t forgiven him for letting Mark Langston [ERA+ of 67] pitch to Tino Martinez in Game 1 of the 1998 WS.)

  28. Shagster says:

    Believe we may be groping our way to a “law of LARGE Sample sizes”.

  29. Bob Lince says:

    When the Great Scorer
    Comes to award the prize
    He asks not whether you won or lost
    But what was your sample size

Leave a Reply

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