By In Stuff

Stolen Bases And Rex Hudler

Last night on the Kansas City Royals game, announcers Rex Hudler and Ryan Lefebvre briefly discussed Sabermetrics and the stolen base. I want to make a point about what they were saying without (I hope) insulting either one of them. Ryan is a friend, and while I don’t know Rex I hear that he’s a very nice guy.

Ryan made the point that there are studies that seem to show that the stolen base is not as effective or important as people used to assume. Ryan will tell you he’s not especially a fan of some of these advanced statistics, but he’s also an open-minded guy and we have had some fun discussions and disagreements through the years. Anyway, I think he was simply making the fair point that while statistics may show that attempting steals is not necessarily a prudent play — and he concedes that this absolutely might be true — he wonders if maybe the statistics do not pick up on some of the psychological force of the stolen base and its threat, such as how it can distract a pitcher and stress out a defense.

This could make for a very interesting discussion. Unfortunately, Rex Hudler took it in kind of a different direction.

First, Rex asked if these studies about the stolen base not being as effective or important were “Sabermetrics.” I don’t know, I found that kind of funny. He said it as if “Sabermetrics” is actually a person. Ryan just said that there were such studies out there that questioned the value of the stolen base. At the time, I should add, the Royals had runners on first and third.

Then Rex Hudler basically said this: If the guy on first (Alcides Escobar) stole second base, he would be able to score on a single. If he stayed at first base, he would not be able to score on a single. But if he made it to second base, he would be able to score on a single. Which he would not be able to do if he stayed on first base. So it would be better if he was on second base. That way he would be able to score on a single. He couldn’t do that on first base. But he could on second base.

“What’s wrong with that?” he asked.

Thus endeth the dissection of stolen bases.

I do not bring this up to knock Rex Hudler but to say that this often — way too often — is the level of argument among the old school. And, I really wish, it wasn’t like that. Old school doesn’t have to be wrong. Old school doesn’t have to be clueless. There are many, many sabermetric thoughts about the stolen base — most of them weighing the value of an extra base against the possibility of an out* — but best I can tell not one of them is, “A runner on first base is better than a runner on second base.”

*Most studies I’ve seen suggest the break even point is about 70%. Rex Hudler was successful 71% of the time.

This is why I wish former ballplayers and managers and people around the game would learn a little bit about Sabermetrics — not so they would AGREE with various arguments but so they would know what the point is and could make interesting and strong arguments why they DISAGREE. That’s what makes for interesting conversation.

I love compelling arguments for things I totally disagree with … I think those sorts of arguments open the mind. If you want to quote a pitcher or two saying that they really were distracted by a fast runner on first and often made bad pitches because of it, please do. That would be interesting. If you want to say that being a running team keeps players heads in the game, it keeps them aggressive, and you would sacrifice a few extra outs for that aggression, hey, go for it. If you want to say that the stolen base is exciting, and fans like it exciting, and in the end you are playing for fans, make that argument. LIke I say, I don’t care if it’s an argument I disagree with. I do care if it’s an argument against nothing.

And the “stolen bases put runners in scoring position” is an argument against nothing. It is like saying the sacrifice bunt is good because it moves the runner to second and then the next guy singles him home. It is like saying the intentional walk is good because the next guy hits it into a double play. It’s like saying asking a woman out in a bar is good because you have a great date afterward and then get married and stay married for 50 years. The best case scenario isn’t an argument. It’s just the best case scenario.

So that’s my plea to Rex and some other baseball folks out there. Just learn a little something about Sabermetrics. Maybe it’s stupid, yucky math stuff figured by the pajama-wearing nerds and it sucks the heart and soul out to the game. Or maybe, just maybe, there’s something in there to talk about.

Print Friendly, PDF & Email

115 Responses to Stolen Bases And Rex Hudler

  1. Brain1081 says:

    I don’t know if it’s worth noting but I did find it slightly ironic that after all the talk about how the pitcher wasn’t able to keep focus on Gordon (at the plate) because he was worried about the baserunners that he went on to strike Gordon out.

    • Tampa Mike says:

      Yea, but after Esky stole 2nd he couldn’t concentrate on Billy Butler. Maybe he was exhausted thinking about Dyson, Esky, and pitching to Gordon, so he couldn’t handle Billy.

  2. John Sharp says:

    There are SOME Sabremetrics that do make sense, like WHIP, or OPS.
    The there are things like BABIP.
    The latter is absolutely silly.
    I don’t know Miguel Cabrera’s BABIP, but I do know his average when he doesn’t put the ball in play… it’s 0.00000000.

    Good article Joe, and I’ll try being better at being SABR informed.

    • civil writes says:

      Actually, since BABIP excludes home runs, the only way for his…er…BABNIP to be 0 would be if he hadn’t hit any home runs. And we’re talking about Miguel Cabrera, here, not Willie Bloomquist.

    • Tux says:

      This comment has been removed by the author.

    • Tux says:

      civil writes beat me to it. And Miggy’s BABNIP (batting average on balls not in play) is essentially HR/K…which comes out to .291 on his career. It’s a flawed stat, and often horrendously misused, but it does have some merits.

    • VanHecke says:

      “This is why I wish former ballplayers and managers and people around the game would learn a little bit about Sabermetrics — not so they would AGREE with various arguments but so they would know what the point is and could make interesting and strong arguments why they DISAGREE.”

    • John: I have no idea why you dislike BABIP, or why you think it’s silly. I’d argue that if you have BABIP, HR rate, and K rate, you don’t even need batting average. (Now THAT is a silly stat!)

      It’s actually tremendously useful to know whether a hitter is getting lucky in finding holes (or, conversely, finding gloves), and if what you’re seeing is a fair indicator of their talent or if it’s likely to change. Carlos Gomez has a .432 BABIP, but his career BABIP is .120 lower, which means that his batting average is likely to be sub .300 the rest of the way. Mike Moustakas has a BABIP that’s barely over .200, and once that normalizes his apparently bad year will actually look just fine.

      Also, you’re wrong about Miggy’s average when he doesn’t put the ball in play – it’s .250 (HR/HR+K).

      And if you like WHIP and OPS, you might want to take a look at FIP (which eliminates hits – Pedro had BABIPs of .200 one year and .300 the next, so accounting for BABIP is even *more* important with pitchers) and wOBA (which is essentially a more precise version of OPS – OPS undervalues walks, overvalues extra-base hits).

    • Tux says:


      Ah poo. Not sure how I botched it but you’re very much correct on that.

    • irishguy says:

      Mike Moustakas’ batting average will not go up as long as he keeps hitting infield pop-ups, which he does with great frequency. This is what gets me about BABIP and trying to say there is a “norm” to it. It also is an indication of how hard a guy has been hitting the ball. In other words, the higher the BABIP, the more likely he’s getting good wood on it.

    • Ian R. says:

      @irishguy – Sure, some hitters can hit for high averages on balls in play. Hitting the ball hard does it, as does hitting a bunch of line drives. Getting down the baseline quickly (think Ichiro) does it, too. Hitting a bunch of pop-ups like Moustakas brings it down. No sabermetrician in the world would argue that hitters have no control over their BABIP. What they will say is that if someone is hitting .450 on balls in play, there’s probably a lot of luck invovled.

    • IceCreamMan says:

      There are norms to BABIP’s and they are career averages or shorter term moving averages. You can look to those norms to see if and how great of a mean reversion you can expect to see. If someone is significantly above or below their babip observed over a significant period of time, you would expect it to revert, unless there has been a tremendous adjustment made (which would likely be observed in other stats as well).

      The issue with advanced stats and sabermetrics is not what is being measured, or how, or if it is significant or not, but in the interpretation and context- this is coming from someone who believes strongly in the value of quant analysis and does it for a career. The world – and baseball – is far too complicated to boil down to one number and many statistics are taken out of context and are not presented as part of a larger picture. Also, too often the proponents of the quant side are not entirely friendly in their attempts to explain them, and the qual side are equally unfriendly in their attempts to understand.

  3. djangoz says:

    But if they were able to think like this then they wouldn’t be “old school.”

  4. I don’t blame Rex on this one, if “Ryan made the point that there are studies that seem to show that the stolen base is not as effective or important as people used to assume.”

    Of course a stolen base is effective. The real point is that ATTEMPTING a stolen base isn’t necessarily the best strategy.

    • KC Nick says:

      Thank you. The point isn’t that a runner on second isn’t better than a runner on first, it’s that the risk of sacrificing an out often outweighs the chance of successfully stealing the base. To discuss the merits of a SB after a runner has successfully stolen is like discussing the wisdom of a bet in a casino post-payout.

  5. Martin F. says:

    Rex and the Old School don’t want to give Billy Beane any more credit for that book he wrote than they have to….you know “Moneyball and the Bomb: How I learned to Love Sabremetrics.”

  6. Patrick Mann says:

    I also think you’re missing the value of the exchange between Hud and Ryan last night. The value doesn’t come from the information presented about statistics. The value is entertainment. Hud is the color guy… and he’s damn good at it. When you say “I don’t know, I found that kind of funny. He said it as if “Sabermetrics” is actually a person,” you prove that he’s done his job.

    • But that can’t be right, can it? If that were the case, you’d let stand-up comedians do the color-commentary.

      There’s at least some basic expectation that commentators should provide informative, accurate information about the topic at hand, right?

    • SM23 says:

      well, Dennis Miller was Mon Night Football color wasn’t he?

    • rogerd says:

      I must disagree about Rex being a “damn good” color guy

    • deaconjones says:

      Me too. In fact, I would consider him a “very annoying” color guy who uses his baseball knowledge and background to “add nothing insightful to the broadcast.

    • Unknown says:

      I totally agree. I especially (don’t) love his “baseball talk.” For example: “jam sandwich” for getting jammed on a pitch; “steak” for RBI, because it sounds like rib eye and his new one from last night, “pea-rod.”

  7. Andrew Daull says:

    Some Sabermetrics, or “advanced statistics” have their place (remember, now-standard stats such as OBP, Slugging %, and Saves were considered along those lines and frivolous at one point), and they can be useful. However, I think Rex’s overall point here is valid.
    Not everything needs a number attached to it. If you want to discuss whether to steal second base while down one run in the bottom of the ninth, fine. If you want to argue that the runner should go because it’s an 0-2 count and you suspect the next pitch will be offspeed low-and-away, that’s cool too. But I think it’s a little too much to throw an actual percentage out there either for or against it.
    In this particular instance (or really, in all cases of a stolen base), you’re trying to get in scoring position. That is an important thing to do in baseball, and the game situation itself (inning, score, count, defensive position, catcher’s arm strength/accuracy) should dictate whether or not to run, not a mathematical formula.
    Overall, I think the biggest reason “old schoolers” dismiss sabermetrics is because not everything needs a number attached to it.
    Someone somewhere obviously thought studying the effectiveness of a stolen base was important. But if Charlie Manuel whipped out a calculator when deciding to send Jimmy Rollins, I’d be pissed.

    • “Not everything needs a number attached to it.”

      In real life, sure.

      But in a game where the only thing that matters is how many runs you score? You’re wrong – everything that can have a number attached to it needs a number attached to it, because numbers are the only things that matter.

      “the game situation itself (inning, score, count, defensive position, catcher’s arm strength/accuracy) should dictate whether or not to run, not a mathematical formula.”

      I think you’re actually contradicting yourself, here. If I asked you to give me an example, you’d probably talk about when it’s a better than average count for running, assessing the strength of the catcher relative to the average arm, when defensive positioning makes it a more likely success… and those are numbers. And whether you’re using an explicit formula or not, you’re still using a formula.

    • Andrew Daull says:

      “And whether you’re using an explicit formula or not, you’re still using a formula.”

      A fair point. While “your” formula may involve some actual numbers and math, “my” formula would be a summation of the game situation.
      I guess my point would be, why run it through a mathematical formula when there doesn’t seem to be an apparent need for it?
      For example (an admittedly extreme one, at that):
      Tie game, bottom of the 9th, Mo Vaughn on 1st, 2 outs. 1990 Benito Santiago is behind the plate.
      Both of our separate formulas would tell us not to run. But why plug numbers when all you need to know is that Mo Vaughn would be running?

    • SM23 says:

      well, that’s obviously not when you’d need that info, cause Mo Vaughn ain’t running anyway. change Vaughn to 2005 Podsednik, Santiago to 2005 Mauer and now it matters like hell what the numbers say

    • Andrew Daull says:

      Also, I’ll allow for your reply, but I think we’re getting a little off track.
      The original argument, as Joe said, isn’t “are stolen bases ever effective,” it’s moreso “when is it effective.”
      Neither of us agree that stolen bases aren’t effective overall. Obviously it’s better to have a runner on 2nd than a runner on 1st. It’s that we disagree on how to go about determining when an attempt is effective. In my example above, it obviously wouldn’t be effective to send Mo Vaughn, both from a statistical standpoint and a situational standpoint. In that situation, nothing advanced is needed.

    • Andrew Daull says:


      Another fair point. And granted, I have no idea what this study about the effectiveness of stolen bases entails, or what math is involved in figuring it out. But in your scenario, again, I think game situation would be the best factor. You have (at the time) one of, if not the fastest to run the bases against a very strong an accurate arm. I think there, any metrics that can be applied even out, rendering the other moot. You know Podsednik will get a good jump. You’re confident Mauer will make a strong accurate throw. So then, it depends on what pitch you think will be thrown (80mph curveball? 93mph fastball? 87mph slider?), and you have to rely on the count and knowledge of the hitter’s tendencies for that.

      Again, maybe we’re using a different way to get to the same answer in the end. It’s just down to which way we prefer.

    • Tracey says:


      If you’re an experienced carpenter, you might be really good at estimating how long a foot (i.e., 12 inches) is. If so, you might just estimate where to cut that board, because messing with numbers by doing some arithmetic is time consuming and not as much fun as going with your gut. Sometimes, you might get it just right. Other times, it might be close enough. Do you think your success rate is going to be as high as the guy who carefully measures and calculates?

      They’re just different ways to get to the same answer in the end…except for when the answer doesn’t come out to be the same. Applying a little intellectual rigor never hurt a soul.

    • irishguy says:

      So baseball is like carpentry? I don’t think so.

      Here is why. A carpenter will always measure a board instead of guessing. And his outcome, if he is any good at all, is guaranteed.

      No matter how many numbers you run, the outcome in baseball is never guaranteed.

    • Patrick says:

      The 70% is actually very easy to arrive at and involves almost no theory to devise. The key is to start with the run expectancy chart. I will be citing the one from Baseball Prospectus from 2011 (found here

      It is important to note that this chart is not theoretical or hypothetical. It is derived by simply examining every game played in the 2011 season and watching how many runs on average resulted from each situation.

      From this chart we can see that a team with a player on 1st base with no outs will score an average of .85 runs. If that runner successfully steals second the will score an average of 1.06 runs. If the runner attempts to steal but fails the will score an average of .26 runs.

      So, in 2011 a successful steal attempt was worth (1.06-.85)= a gain of .21 runs. But a failed steal is worth (.26-.85)= a loss of .59 runs. With that we can simply solve .21x=.59(100-x)

      So, in order to break even on a steal of second with 0 outs in 2011 you would have to succeed 73.75% of the time.

      Sorry for any formatting issues. I posted this from my phone.

    • Andrew Daull says:

      Irishguy kind of beat me to it, as I was going to post just about the same thing.
      If I’m a carpenter, of course I’m caring about the accuracy of my cuts. If a board is 11 inches instead of 12, it ruins the whole project. That is a certainty.
      I don’t believe that certainty exists in baseball. If the league average of stolen base success is 70%, and I’m 7/7, does that mean I’m not going to run because the statistical average tells me I have a 100% chance of being thrown out my next three attempts? Conversely, if I’m 0/3, does that mean I’m going to be successful in my next 7?
      Of course it doesn’t. The possibility exists I could go 0-for my next-3 (or 7-for my next-7), but that possibility doesn’t exist because math says it will.
      I fear that sometimes these metrics are viewed as certainties for the future, when they are no more of a guarantee than any other “conventional” statistic.

    • Ian R. says:

      @Andrew – if you’re a 70 percent base-stealer, statistics absolutely do not say that you’re guaranteed to go 0 for your next 3 if you’ve gone 7 for your last 7. A true talent 70 percent base-stealer would be expected to steal at a 70 percent success rate in any sample. In this case, you’re most likely going to go 2-for-3 (which is the closest you can come to 70% in three attempts).

      Anyway, a metric like that isn’t designed to predict future performance. What it’s saying is that, in general, you should only run if you’re confident you have a better than 70 percent chance of stealing the base. Obviously that changes somewhat depending on the game situation – with runners at the corners and two out, for instance, getting caught would be devastating because you’d strand the runner at third, so the break-even point is higher. On the other hand, in a tie game with nobody else on, the break-even point could be lower.

    • Andrew Daull says:

      Understood. My question is, if you’re on the basepaths, are you calculating in your head if there’s a 70% chance of success? Most likely not. You’re studying the game situation, pitch-by-pitch, to decide whether or not it’s a good idea to try to steal. Am I being held on? How is this guy’s pickoff move? How much of a lead will that allow me to take? Is he pitching out of the stretch or slide-stepping? Will he throw fastball or offspeed? Who will be covering second? Where is he aligned defensively? Can my speed beat the catcher’s arm?
      None of that can be quantified in the 10 seconds between pitches. It’s only a number that has been calculated based off of what has already happened, just like any other stat.
      The “break-even points” don’t need a percentage on them. That’s just sound strategy to not strand a runner in scoring position, or to get yourself in scoring position if you think you can.
      Somebody thought/thinks these numbers are important and can be applied. The A’s were on to something when they put a higher emphasis on OBP, but as they found out, OBP isn’t the apex of statistics (however, now we see that OBP has become a much-ballyhooed statistics, one that even anti-saber people use). The game needs to be played and let the statistics happen as they will. The statistics shouldn’t dictate the game.

    • BobDD says:

      Andrew seems to me to be saying that you do not need a stat to know if it is night or day. Well that is true though I do not see any reason to brag about it. However, there is a stat that you can look up as to the exact time of sunset, sunrise, and so on. You do not need that to decide to water your lawn or not – congratulations you get to do that numbers free. But evidently there are other people (they write large annual almanacs about this) like to know things exactly, because they find it interesting and/or because they find it valuable and the argument that you don’t always need the number sounds to me like head-buried-in-the-sand.

      Likewise you do not need a computer and an internet connection to live, but many of us like them anyway.

    • Andrew Daull says:

      I’m not seeking to discredit sabermetrics, I’m genuinely trying to see what advantage they have over “old-school” on-the-spot situational analysis. If sabermetrics are just the quantifications of that situational analyses, then there ultimately is no difference, and if I were managing, I’d make my decisions off of in-game scenarios, not a formula. My decision may end up being the right one, or it may backfire, just the same as any decision made using a sabermetric analysis.
      It’s just another way to go about making a manegerial decision, and it’s subject to the same bad hops, shifting winds, untimely shadows and mental lapses that have hovered over every game for the past 150 years. Therefore, I fail to see how it’s any better than the “old” way, like so many people claim.

  8. bigsteveno says:

    That comment from Rex Hudler was probably one of the LEAST stupid things he said during that broadcast. He may be a “nice guy,” but he’s a disaster in the booth. He’s a plague on the Royals broadcasts, and I hope to god the organization finds someone competent next year. Or sooner.

    • Patrick Mann says:

      You’re missing the boat. He makes the game much more fun to watch. Fun. Remember that? A lot of Royals fans don’t.

    • tomemos says:

      He absolutely doesn’t. He’s annoying and artificial, in addition to not providing useful information or insight about the game.

    • mgh329 says:

      I preface this by saying that I could not do Rex’s job for a multitude of reasons, and from all accounts I have heard he’s a good guy doing his best…But, he’s a horrible broadcaster. I personally get pissed off listening to the ignorant stuff he spouts out and eventually change over to the opposing team’s coverage. In retrospect, maybe I could make it more fun by creating a blog, “Stupid S**t Rex Says”

    • Unknown says:

      There’s already a twitter account for that…

    • I had to listen to the Rex the Wonder Dog for years with the Angels, and my own dog knew more about baseball than he did. The man is an idiot. Please don’t use him as an example of “Old School”—he’s more like a drop out.

    • I had to listen to the Rex the Wonder Dog for years with the Angels, and my own dog knew more about baseball than he did. The man is an idiot. Please don’t use him as an example of “Old School”—he’s more like a drop out.

  9. I understand (and even appreciate) the numbers. And indeed, many managers attempt the SB (not to mention the sac bunt) far too often. But as someone whose favorite team ever was the 1985 Cardinals, I have a hard time believing they would have been better if they had just never stolen bases. I guess it comes back to Ryan’s point about the intangible impact on the pitcher. Interesting discussion though.

    • Flax says:

      The 1985 Cardinals stole 314 bases as a team and were caught 96 times. That’s a 76.5% success rate, above the line that sabermetric types think is the cutoff for where stolen base success is too low to be worth the tradeoff of the out. So really in your example neither side is wrong.

    • thrillho says:

      No one is saying a team is better if they don’t attempt a stolen base at all. For instance, those ’85 Cardinals were a great team and were great on the basepaths. The point is to look at success rates. Joe alluded to 70% being the break even point. So on that Cardinals team, Terry Pendleton probably should’ve stopped trying to steal (17 steals, 12 caught stealing) while Coleman, McGee, Smith, Herr – were successful enough that the steal attempt was worth it.

    • adam119 says:

      Who suggested they should never attempt to steal bases? All he’s saying is to consider all aspects of the decision, including whether it’s worth the risk or not. If a SB success rate is around 70%, then there are some situations where taking that risk is well calculated, and others where it’s foolish. Is it so wrong to have a discussion around that?

    • The ’85 Cards had about a 77% success rate (the league-average was 10% lower) and were 22.6 runs above average on baserunning alone – the first number is good, but the second is huge. And the only 30+ base-stealer on the team with a worse than 80% rate was Ozzie, and he was barely below that number. So, no, they wouldn’t have been better if they never stole bases!

    • BobDD says:

      second to what adam said; if there are eleven different influences that could be considered, why are some cutting out dreaded sabermetrics/numbers and insisting that since only three influences are needed to consider sometimes (the Mo Vaughn/ceramic turtle argument), that additional factors should therefore not ever be looked at. Luddites.

  10. Mark says:

    Lord, thank god you weren’t listening to the Twins broadcast when they tried to talk about them. That could have been a 5000 word post on how out of touch Blyleven is….

  11. Music Guy says:

    Put succinctly, Rex committed the classic fallacy of begging the question.

  12. Joe says:

    The ’85 Cardinals are a good example of the value of basepath pressure. They were a totally different team before Coleman came up and a shell of themselves when he was lost for the year. I don’t know what his success rate was but the game changed when he made it to first. It was a blast to watch.

  13. jim louis says:

    I like Hudler’s spirit and positivity and he seems a real good guy. But I miss Frank White on the broadcast. White was interesting, informative, critical, and liked his dry, laid back humor and laugh.

    Another new guy, Steve Physioc (sp?) is insufferable.

  14. irishguy says:

    So Joe, explain to me how the sabermaticians arrived at the “break point” of 70 percent success. How did they actually determine that?

    You see, you oversimplied “old school” argumentation to make your point. Allow me to make mine.

    Too often I read from the sabermetrics crowd “well, somebody has studied this,” without really understanding the nature of statistical analysis enough to judge whether the study in question is valid.

    Secondly, sabermetricians didn’t come up with the idea that you attempt to steal only when there is a high probability of success. In other words, even an old school guy wouldn’t give Billy Butler the green light. But in the situation at hand? Yeah, why not send Alcides Escobar? At the very least, the threat of it would make the other team change their defense.

    But as somebody noted, the Vince Coleman Cardinals and the Lou Brock Cardinals used the stolen base quite effectively to score runs that otherwise would not have been scored.

    • tomemos says:

      This comment confuses me. You start by wanting to know exactly how the 70% figure was arrived at, which is fair enough. It’s good to be skeptical and ask for evidence. But then you go on to say that of course everyone knows that stealing is only worth it in high-percentage situations anyway. In other words, you’re accusing the research of being both suspect and unremarkable, which is kind of a contradiction.

      If we agree that there is a break-even point for stealing, isn’t worthwhile to determine as precisely as we can what that point is? Wouldn’t that be better than the “yeah, why not” approach endorsed by you and Hudler, where we just suppose we can’t know what’s best, so just try anything?

    • irishguy says:

      It is suspect in that I want to know why the figure isn’t 60 percent. Or 75 percent. Or whereever, and how it was arrived at. But statheads always quote numbers like that as if they were carved in stone and came down from Mt. Sinai because someone else has “studied” it. In other words, it cuts to the heart of the sabermetric argument.

      It is unremarkable because that still won’t that wasn’t known long before sabermetrics. You don’t steal unless you have a high chance of success? Well, duh! Statheads are always discovering answers to questions that aren’t worth asking, then act like they were the first people who thought of it, and they have discovered the secret to life.

      And what on earth does putting a precise number on that tell us? That a guy with a 60 percent success rate shouldn’t run against a pitcher with a long, slow delivery and a catcher who is just as likely to throw it into centerfield as second base?

    • tomemos says:

      But how can you say “you want to know why the figure isn’t 60 percent” and simultaneously say that these are “questions not worth asking”? You don’t want to know why; you just want to criticize an approach you’ve made up your mind to hate. You accuse people who use statistics of arrogance but you’re the one who just dismisses your opponents by namecalling (“statheads”).

      It’s not that a runner must have a 70% or better success rate to run; it’s that the win expectancy of attempting a steal only makes sense if you have roughly a 70% chance or better chance of making it. Now, of course we’ll never know at any given moment what the chances are, but at least having a figure to estimate from is better than the total guesswork of “why not?” Surely you’ve come across managers who just reflexively run into outs without any cost-benefit analysis; I agree that we don’t need statistics to explain why that’s wrong, but isn’t studying the issue bound to enhance our understanding of it?

    • Wilbur says:

      It sounds as if Irishguy is asking for a citation to the study or studies which determined this about stolen bases, win expectancy, etc.

      It’s not an unreasonable request, it’s just lazy.

      The best place to start is the Bill James annual abstracts from the 1980s. Your public library probably has some or all of them. They are very entertaining reading and not just dry statistical junk.

      I believe there are websites which define and explain these sorts of things as well. If you’re interested in learning, the material is there.

    • tomemos says:

      I don’t even think it’s lazy to ask for that information. It’s just that it’s obvious Irishguy has already made up his mind that the studies could not be valid or useful, so his request is not in good faith.

    • irishguy says:

      Wilbur, no. Laziness isn’t asking someone to back up their statement. Laziness is saying “there are studies” then not providing the citation.

      And thank you for your suggestion, but I already own the Bill James Abstracts from the ’80s. You see, I was quite the budding stathead in my day, until it began to dawn on me that very little of what James was saying would actually came true. And I learned that the hard way, taking all this wisdom and finishing in the middle of the pack of my Rotisserie League every year.

      It dawned on me that when James was right, he was merely expressing conventional wisdom in a new way, wrapping that around numbers and formula that were designed more to confuse with B.S. than to enlighten with newly discovered keen insight.

      And when James was wrong, and he was quite frequently, he was spectacularly wrong. Two examples. His 1985 Abstract, written in the wake of the Tigers historic season, predicted that they were on the brink of a dynasty because no one on that team had a season that was far outside their career norms. In other words, you could expect the Tigers to continue to perform like that as far as the eye could see.

      The next year, when they fell to 84-77, third place and 15 games out with all their regulars putting up virtually the same numbers, James wrote that he failed to account for how much the bench and bullpen contributed to the 1984 success.

      I won’t even go into his utter failures to predict future performance with the PECOTA and BROCK2 gibberish. Suffice it to say once again that I once took that seriously and got embarrassed by it.

    • irishguy says:

      tomemos, since the point eluded you the first two times I tried maybe the third time is the charm:

      Don’t be so gullible as to accept “studies” without knowing their basis.

      I find it ironic that Posnanski uses “there are studies” to cut short a discussion at the same time he accuses Hudler of trying to cut short a discussion with the perfect reasonable statement that you steal bases to advance runners into scoring position.

    • tomemos says:

      With respect, I don’t know if a roto league is the most convincing means-testing of statistical analysis in baseball. I might look instead at teams like the A’s, and Rays, which achieved success far beyond their payroll by relying on stat analysis. Or we might consider the championship teams which made statistical analysis a big part of their strategies, like the Red Sox, Cardinals, and Giants. If analysis is so worthless, why are baseball teams paying to have it done? Are they all being led by the pied piper?

      Your account of Bill James as a huckster who merely repackaged “conventional wisdom” is a flat misstatement of what the conventional wisdom was in baseball in the 1980s, or for that matter is today. No, Bill James did not discover the value of the walk or the waste of the bunt (and he gave credit to managers like Earl Weaver who already had this understanding). But neither were those things “conventional wisdom.” He and those who came after him provided the data that made those propositions undeniable, and the baseball world has by and large fallen suit.

      I do know the basis of the stolen base studies, in fact; they’re based on the expected win percentage of three situations: runner on first with (say) no one out, runner on second with no one out, and nobody on with one out. For the attempt to be worth it, the gain in win expectancy that comes with the additional base has to exceed the loss that comes with the risk of losing a baserunner and an out. I’ll admit that I haven’t looked into these studies carefully, but I’m not really here to defend the studies; I’m here to defend the principle of analyzing the question statistically. If you accepted that this was even worth doing, we wouldn’t have a disagreement.

      And finally, relating to the post that we’re all talking about: do you actually think that just assuming the success of the stolen base is a “perfectly reasonable” rebuttal to statistical analysis regarding base-stealing, or is it simply convenient to ally with the enemy of your enemy? I hope you’re playing the knave and not the fool here.

    • Patrick says:

      The 70% is actually very easy to arrive at and involves almost no theory to devise. The key is to start with the run expectancy chart. I will be citing the one from Baseball Prospectus from 2011 (found here

      It is important to note that this chart is not theoretical or hypothetical. It is derived by simply examining every game played in the 2011 season and watching how many runs on average resulted from each situation.

      From this chart we can see that a team with a player on 1st base with no outs will score an average of .85 runs. If that runner successfully steals second the will score an average of 1.06 runs. If the runner attempts to steal but fails the will score an average of .26 runs.

      So, in 2011 a successful steal attempt was worth (1.06-.85)= a gain of .21 runs. But a failed steal is worth (.26-.85)= a loss of .59 runs. With that we can simply solve .21x=.59(100-x)

      So, in order to break even on a steal of second with 0 outs in 2011 you would have to succeed 73.75% of the time.

      Sorry for any formatting issues. I posted this from my phone.

    • Ian R. says:

      It’s also worth noting that the break-even point for a stolen base isn’t a static value – it changes over time. In the low-scoring ’80s, for instance, teams could be successful with stolen base rates that would be harmful in the high-flying late ’90s.

      Really, this makes intuitive sense. If most hitters in the league are singles hitters, you gain a lot of value by moving the runner from first to second – now he can score on a single, and in a low-scoring league that one run could be the difference. On the other hand, when most hitters are hitting for power, you’re not gaining as much – a runner on first can score just as easily on an extra-base hit as a runner on second. Furthermore, if you’re in a high-scoring league, you probably need a multi-run inning to win, and a stolen base doesn’t do much to help you put together a multi-run inning. Keeping the runner on base, though, does – which is why the damage inflicted by a CS is higher, relatively speaking.

    • BobDD says:

      IrishGuy, sabermetrics are not a magic talisman/bones/dice that decide when the stars are aligned and whether to do something. They are simply more information. Why would anyone turn down additional good information? Sabermetrics can be used properly or improperly, but if you do not understand them then by all means leave them alone. As for me, they light up my life.

    • DavidJ says:

      irishguy, The Book by Tango, Lichtman, and Dolphin has a chapter on base-stealing; pages 333-339 in particular cover the break-even points for the stolen base attempt in various situations (because, of course, there’s no one-size-fits-all break-even point; it depends on all sorts of contextual factors, and they do a very good job of sorting these out). That should give you what you’re looking for. Also, you might want to wander over to Tango’s blog; he’s very friendly, transparent, and responsive, and I’m sure he’d be willing to help you out with any questions you have about how these kinds of numbers are arrived at.

  15. Wilbur says:

    I’ve wondered what the best announcers I’ve ever heard – Harry Caray (with the Cardinals) and Jack Buck – would’ve had to say about the sabremetric trend. I saythis as a life-long Cub fan.

    If you heard Caray in the 60’s you would know him as the sharpest, keenest-eyed broadcaster ever. He was anything but the beery buffoon as he is caricatured now.

    Jack Buck was a well-rounded man, cultured and intelligent, and may have readily accepted the broadening of baseball knowledge.

    Both were around when Bill James was putting out his annual abstracts, but very few within the game appear to have been aware of them.

  16. Mike Caruso, 1999:
    12 SB, 14 CS

    Craig Biggio, 1993:
    15 SB, 17 CS

  17. Principe says:

    A quick check with Baseball-Reference suggests that, overall, baserunners were successful on 74% of steal attempts in 2012. Of those who attempted at least 5 stolen bases throughout the season (a conservative benchmark for players who we might consider “base stealers”), they were successful on 77% of attempts.

    If the break-even point truly is 70%, as Joe remarks the studies show, then I’d say that this is an interesting indication that managers may actually know what they are doing when it comes to utilizing the stolen base.

    This makes me wonder if this is due to more of a sabermetrician’s perspective in front offices, potentially talking to managers about optimal strategies, etc., or if managers actually have more of a “feel” for the game than we give them credit for. In 1982 and 1983 (just two random years before sabermetrics blew up), league averages were 66% and 67%, respectively. Maybe the sabermetric emphasis on smarter SB strategies has actually had an influence.

    I’m not sure what my point is exactly, but maybe the stolen base, as it is employed strategically today, is actually a sound strategy.

    • irishguy says:

      But this is where assigning a specific number to it and saying, “above this number good, below this number, bad” is flawed.

      First of all, you don’t know the outcome in advance.

      Second, if you are playing in a bandbox with a Murderer’s Row, stealing bases with an 80 percent chance of success might not be a good idea. If you are playing in Yellowstone with a bunch of line drive hitters, stealing with a 65 percent chance might be a good idea. The odds are better than getting the two base hits you’d need to bring that run home from first.

    • clashfan says:

      Irish, where did you come up with those numbers? Can you cite the study used to ascertain them?

      No one’s saying to never take the game situation into account. Just that it’s a good idea to know what the baseline is.

    • BobDD says:

      Irishguy strawman. I’ve never heard a numbers guy say “above this number good, below this number, bad”. Situations are customizable, but if someone who steals 25 bases a year has only a 60% success, you will probably have him attempt less while having the guy who steals 25 at an 85% rate attempt more. Why wouldn’t you use those success rate stats?

  18. Frank says:

    Let’s imagine baseball with a rule that stealing bases was not allowed – like Little League. How different would the game be? No need to hold runners and no wide hole on the right side. No throws over to first and no errors on throws over to first. No hit and run / no movement of the infielders taking them out of position. No balks. I could go on.

    These are some major changes to the game. So what happens when a team self-imposes a rule to the effect that it will not steal? We get somewhat of the same effect.

    The threat of the steal – even if not actually used – is of great value. But if a team (or player) never steals (or seldom does), then the threat is removed (or reduced).

    The problem is that this is something that stats are going to have a hard time measuring.

    • tomemos says:

      Sure; and if a team tried to steal every time it had a runner on first (as an insanely literal, obviously unfair interpretation of Hudler’s words would suggest), then it would lose most of its baserunners on caught stealing or pickoffs and would never get a rally going. But no one is advocating always stealing, just as no one is advocating never doing it. So what we need is more consideration of what precisely the “fear of the steal” is worth. I agree that this will be difficult to measure, but I think it should be possible to at least observe. That’s a starting point that could lead to more analysis.

      Anyway, do you notice that intelligent defenses of the hidden virtues of the stolen base, like yours, are exactly what Joe is saying he wants, and exactly what Hudler is not providing?

    • You know you just described baseball as it was played in the 40s and 50s. Some people claim that was baseball at its best. And hardly anyone ever stole a base.

    • Frank says:

      You make a great point about the way the game has been played differently in different eras. Brock and Wills started changing that in the 1960’s. But the hay-day for stealing was the 1980’s, with the Cardinals being the prototype. Astroturf had a lot to do with that. If you could steal your way around the bases, you didn’t need extra-base hits from much of your line-up, and you built the team accordingly.

    • Luis says:

      I believe Frank is actually making Joe’s point. It isn’t whether or not stealing is a good idea, it is that Rex’s argument against “the studies” was dumb. Frank makes good arguments in favor of how the intangibles from the steal may help a team, you may be in favor or against it, but it is a fair argument and that’s is all we can hope for.

    • Dan Shea says:

      There may be a good reason there were fewer stolen bases in the 1950s – baserunners were on average far less likely to succeed then. 8 of the 10 worst success rates from 1939 to present are from the ’50s. They had success rates in the 50-60% range.

      And yes there were far fewer steals then – 8 of the 10 lowest average number of steals/game are from the 50s as well.

      Which perhaps goes to prove the sabermetrics point, that the stolen base is a good idea when your success rate is above 70%. When you’re making an out nearly half the time it’s not so attractive.

  19. RedsManRick says:

    The core of the problem you identify is this:

    Many old school people think Sabermetrics is a specific set of strategies/preferences, as if it were a science-based style of baseball that hates defense, SB and bunts and loves walks and homers. They think that brand of baseball is either wrong aesthetically or that it fails to appreciates the nuances of the game that they see.

    What they’ve failed to understand, at a fundamental level, is that sabermetrics is simply the formalization of the kind of analyses and trade-offs they’ve always made intuitively. It’s not about them being wrong. It’s about more rigorously understanding how their intuition works, which gives us the opportunity to learn and improve from data.

    Perhaps they do get the point is to have an enlightened conversation around learning , but that they are simply upset that opening the doors threatens their special place in the world as holders of this knowledge. Or perhaps they’re offended by the idea that people who haven’t walked in their shoes could possible contribute.

    In any event, comments like these and those from Hawk Harrelson, Joe Morgan, Tim McCarver and the list goes on suggests to me they have a very, very wrong perception about what sabermetrics is all about.

    • Andrew Daull says:

      @RedsManRick I lean toward being “anti-saber,” but I can appreciate your explanation that “sabermetrics is simply the formalization of the kind of analyses and trade-offs they’ve always made intuitively”, and would love it if sabermetrics were presented this way. But too often, I see and hear it used to explain that “so-and-so isn’t really good, he’s just lucky because the opposing team’s infield has the second-lowest UZR in the majors,” or “this guy will be the best one day because his BABIP+OBP would be first in the majors if he had the qualifying number of atbats (I’ve seriously seen that. I know BABIP and OBP are things, but adding them together seems completely arbitrary). To think in those terms, to me, completely discounts that a certain player just may be good or that maybe he’s just bad.
      I stated in another post above that one of my biggest turn-offs toward sabermetrics is that I feel some present it as future certainties, rather than use it as a tool to quantify what has happened.
      Maybe to better explain, if this were 2006, I would want Albert Pujols at the plate with the game on the line. Based on his past history, I feel he would give me the best chance to win. But to say “he is going to get a hit because the other team’s UZR is 30th in the league, and this divided by that equals whatever, so therefore we win” seems overkill to me.
      I guess I’d feel better about sabermetrics if more people used it as just another statistic to quantify what has happened in the past. But unfortunately, I see it used to predict the future, and if you disagree with them, they just throw more numbers in your face.

    • tomemos says:

      Andrew, that’s an interesting way of putting the issue. First of all, I’ll say that you’re totally right that some people are way too dogmatic about stats, and treat some questions (like defensive analysis) as totally settled when they’re not. At the same time, we shouldn’t judge all of stat analysis by its worst practitioners, any more than we should assume all stat skeptics are as bullheaded as Joe Morgan or Hawk Harrelson.

      But getting to your example about Pujols: I think you’re actually supporting RedsManRick’s point. Your saying that you would want Pujols up at the plate because his “past history” indicates that he’s got the best chance of helping your team. In other words, you’re projecting/predicting what he is likely to do compared to other players. Yet you also criticize sabermetrics being “used to predict the future”! What RedsManRick is pointing out is that we’re *all* looking at past results to predict future performance. It’s just that sabermetrics types are trying to do so systematically, using numbers that will allow us to be more precise than the old numbers some use (RBI and such).

      I don’t think you’re going to find that a lot of saber-types would disagree with you about Pujols in 2006, any more than you’d find any who would say that it was wrong to send Dave Roberts in Game 4 of the 2004 ALCS. Those are the easy cases. It’s in marginal cases—do we intentionally walk this guy? Who should we pinch-hit here?—that sabermetrics, if it’s done right, may have an edge over the traditional way of doing things.

    • Tux says:


      The reason that happens is because being able to identify the “true value” of a player more accurately (or earlier in his career) is a very valuable thing. Stock brokers do it every single day and sometimes it works and sometimes the housing bubble collapses.

      If someone is using stats of any kind to say that any single outcome is given, they’re a fool. There are simply too many factors in the game that can’t be controlled. I once read that there’s something like 2 million things that can happen when a ball leaves the pitcher’s hand. Fly balls occasionally hit birds and drop in for a double when the outfielder was already camped under where the ball would’ve come down, and sometimes a runner trips over second and gets caught in a rundown when he should’ve scored on a line drive down the line. But in the greater picture, events like that balance out and we can be left with a large enough sample to really dig into the nitty-gritty of a player’s talents and abilities.

      I’d say that you can predict the future on a larger scale, over the course of an entire season. You can see trends over the course of history that point to things in the future (looking at how players improve towards their peak years, for instance). We can’t know everything and we never will. But if we can shrink down the stuff that we can objectively calculate, then that makes the luck factor that much more important.

      And whoever added BABIP and OBP should be smacked in the head.

    • Andrew Daull says:

      I gotta say, it’s refreshing to have a sabermetrics discussion without it devolving to “mom’s basement” and “fuddy-duddy” name-calling. That’s unfortunately rare.
      As I’ve said, my main hangup on these metrics is how they’re used as “certainties” for the future. One of the cooler “saber” stats I’ve seen is (and I forget what it’s called), but the one where the number of wins is projected by the number of runs scored (e.g. Team A scored 700 runs last year, which, over history, projects to 85 wins, or whatever. So if you want to win 90 games, build your offense to score 900 runs). It may project to that, but of course there is no certainty that if you score 900 runs, you will win that many games. Other factors come in to play. You could win less, or you could win more. I think what’s being lost is just being able to watch a game and think “hey, with this offense, this team’s never out of it.” (’95 Indians and ’09-’10 Phillies come to mind). Can’t we just say that a team is good and dangerous, rather than “Team A is avering 5.67 runs a game, and Team B is averaging 3.8?”
      That’s an example of what I mean when I say it’s used to predict the future. It’s presented as fact or certainty, when it’s not. As Tux said, there are 2 million things that can happen when the ball leaves the pitcher’s hand. We just differ on whether there should be a number attributed to those 2 million scenarios.
      To sum up my posts (cause I should do some work haha):
      “Traditional” stats (BA, W-L, ERA, etc.) quantify the past.
      Sabermetrics seek to quantify the future.
      I just prefer to look at the past (like in my Pujols example)

    • Robert says:

      This comment has been removed by the author.

    • Robert says:

      The “predicting the future” of sabermatrics is kinda where the value of it lies, though. If I lived in Vegas in the early 90s armed with today’s predictive statistics and future performance probabilities…I’d own my own island in the Carribean by now.

      In real baseball, this predictive value manifests itself in determining probability of future performance to determine contract value/length/etc. If sabermetrics holds itself to only analyzing the past, it’s value is almost completely destroyed, since we already know what happened then.

      Pretending that sabermetrics claims any sort of future certainty is disingenuous to me. Statistics lends themselves so naturally to probability density functions that it goes without saying (or it should) that that is what sabermetrics are talking about. Opening every sabermetric discussion with “I’m talking about future probabilities here” would be annoyingly redundant.

    • Andrew Daull says:

      You won’t get any argument from me on that. It’s when you’re discussing sabermetrics with someone and they present these metrics as absolute certainties. They’re not. And they’re no better a predictor of future performance than a traditional statistic such as BA.

      We are competing GMs evaluating the same player.

      Player A is a 25 year old with 3 years experience with yearly slashes of .290/18/86, .268/24/97 and .281/19/88.

      We see very solid rookie season, an uptick in power and maybe some good hitting with RISP in year two, and a swing more towards the mean in year three.
      An eyeball projection would tell you that for the next 4 or 5 years, until age or injuries start affecting him, he will hit within that range. You don’t need anything advanced for that. If I’m a GM, I’m not signing a player because of factors like the other team’s UZR, his WAR, or his BABIP. I’m looking at those numbers and seeing a young player who has put together some good years, and can be my everyday rightfielder (or whatever) for the foreseeable future. I’m sure that no matter what metric you evaluate him with, you’d end up with the same evaluation.

      Now, if Player A’s BAs were .180/.320/.250, I’d obviously need to do more thinking. And even then, I’m not researching the UZRs of defenses in his down years. WAR is useless here because you know it’s going to be below average. While your evaluations show he was a victim of bad breaks in years 1 or 3 and pay him according to what you think he might do, I might see someone who has shown some potential but needs to work on his batting eye, plate discipline or swing mechanics. I offer him the chance to win the starting job in Spring Training. You offer a starting job. He signs with you.

      Now, regardless of either statistical scenario up there, Player A could go on to hit .300/25/100 every year. Or he could hit .200/3/10 and be demoted after a year and a half.
      Either way, whichever way we decided to evaluate the player, he ended up the same. He didn’t succeed or fail because mathematical probabilities projected he would, he succeeded or failed because of the work he put into his craft and how that translated on the field.

      My point isn’t to discredit sabermetrics. They’re the same as any another traditional statistic; a number multiplied or divided by another to give you a percentage point of what’s happened. It’s that I don’t see their value as being any greater, or being a better indicator of future performance, than those traditional stats. A hit is a hit regardless of UZR, same as an out. A win is a win and a loss is a loss no matter what a team’s collective WAR is. Play smart ball, score as many runs as you can, keep the other team from scoring however you can, and let the numbers be just that…numbers.

    • BobDD says:

      Well I do believe that more numbers are needed because those stats you give could be part of a very good player who makes fewer outs (OB%), fields a difficult position well, steals bases well and has no record of injuries. However those same stats you use could be a player who rarely walks, so he has lower OB% making too many outs, could be a bad fielder, maybe is on the disabled list twice a year and so on. There is room for that player you described with three stats to be good or bad. That is my basic love of stats, but if it doesn’t turn you on so be it.

      Having disagreed on the merits, to me the basic argument here is whether to limit myself to as few statistics/information as possible or to have as much information as can be. Of course then, I’d have to apply it correctly. I have no arguments against any of your information – none at all. But it sounds like you are saying my additional information is at least unnecessary (or maybe even bad?).

    • Andrew Daull says:

      Bad? No. Stats are just numbers, and that goes for both traditional and saber. Stats can be good. Stats can be bad. Stats can be twisted any way to make whatever argument you’d like. I just feel that viewing the game through “saber-lenses” is nothing more than holding a mirror to what is actually going on, and defines a player by his stats, as if those stats were part of his genetic makeup, and those stats are what “allows” him to be good.

      Take a player who has a +7 WAR.
      I see a player who is good and helps his team win games. His WAR is therefore high because of his output. In my view, WAR is secondary. His +7 rating is a product of a mathematical formula devised to quantify his “goodness,” and therefore not necessary for me to judge him as a player.
      A sabermetrician sees a +7 WAR, so therefore this player is deemed as good. WAR is primary, and ability is secondary.

      Both of us are watching the same player, and both of us accept him as an indispensable part of the team. So what advantage does the “new” way have over the “old” way?

      If we’re talking about who we should sign from a pool of a bunch of 0.0 WAR players, then who does it matter who we sign? Ultimately, we’d probably both go for the better pinch hitter, the guy who Ks less, the guy with the most pop, or the better defender, whichever our team needs. Again, I’m not seeing the advantage of the “new” way over the “old” way.

    • Andrew Daull says:

      “Of course then, I’d have to apply it correctly.”

      Exactly. Unfortunately, I’ve yet to see it applied “correctly,” and therefore “better,” like I’ve heard so many say sabermetrics are.
      You can sign a player to the T of what sabermetrics say is best, but that player is still subject to the chaos theory ruling over every game (see the 2012 and 2013 Anaheim Angels)

    • Ross says:

      I’d been meaning to get into these topics for a while, but the good discussion here finally triggered me into writing. Just put up two new posts, one on the value of stolen bases and the other on the value of utilizing low-priced platoons over expensive everyday players. I think anyone who’s enjoying this discussion here might enjoy it:

      I’d love to hear anyone’s feedback and any additional thoughts. I think that the league could actually be stealing more bases, as league-wide success rate is above the break-even threshold (and stealing third, in particular, is very easy for a smart base runner). As for platoons, I’ll just say here that teams should be using them a lot more.

    • Robert says:

      “Player A is a 25 year old with 3 years experience with yearly slashes of .290/18/86, .268/24/97 and .281/19/88.

      We see very solid rookie season, an uptick in power and maybe some good hitting with RISP in year two, and a swing more towards the mean in year three.
      An eyeball projection would tell you that for the next 4 or 5 years, until age or injuries start affecting him, he will hit within that range.”

      Well, 25 year olds have usually not peaked, which sabermetrics went a long way toward proving. So, if you’re intending on this hypothetical player being your everyday right fielder by paying him assuming 5 years of that output, you’ll be disappointed when you don’t sign him. Or, if his high HR/RBI season included an abnormal BARISP or HR%/FB, sabermetrics could keep you from overprojecting…maybe you don’t use those number and you get your everyday right fielder by overpaying for him.

      The idea that someone (anyone) could put an optimum performing team on the field for the optimum price using the eye test is silly. GMs should, and the good ones do, want every piece of data available to them to project future performance. And since baseball has been played for hundreds of years over thousands of players and hundreds of thousands of games, sabermetrics is an objectively useful tool in doing that. That fact that it’s not perfect is hardly a criticism at all.

    • Andrew Daull says:

      “That fact that it’s not perfect is hardly a criticism at all.”
      That’s what I’ve been trying to tell sabermetricians for a long time. Maybe I’m just jilted from countless discussions that I have both had and seen that just quickly devolved to “mom’s basement” and “stick in the mud” name-calling. Luckily, that hasn’t happened in this comment thread, and there has been some great back and forth by both sides on the merits (and demerits) of sabermetrics.
      Anyways, in my experience, sabermetricians are quick to devalue most “traditional” statistics as flawed and/or archaic, a fact I will agree with to some degree. Of course a batting average is slightly affected by a fielder’s speed, overall ability, or an odd scoring here and there. Naturally, a pitcher’s ERA can be affected by the same factors. Of course, maybe the hitter is just good (or bad), and the same with that pitcher.
      However, to point out the flaws of sabermetrics often turns into a holy war. In my experiences, a solid majority of sabermetricians view those stats as concrete and flawless. When I say that I don’t see how sabermetrics are better or offer an advantage, I don’t mean to discredit them. I simply view them as an alternate way to manage (or in our case, view) the game we both love.

      In my 25 year old player example, let’s assume he hasn’t peaked yet and ends up hitting .300+/25+/100+ for the next 4 or 5 years.
      Your studies tell you he hasn’t peaked because the metrics say most 25 year olds haven’t peaked yet due to a myriad of other mathematical formulas..
      My studies tell me he hasn’t peaked because perhaps a couple more years of experience will have him playing at his highest level of ability. Or maybe there are maturity or work ethic issues, a la, the career arc or Delmon Young. Maybe he’s an extreme case, but metrics couldn’t accurately predict what was expected of him and what he has produced. However, to take a good look at his maturity level and work ethic 5 years ago would have told you this guy is probably never going to reach his full potential.
      Anyway, I digress back to our 25 year old player. His BARISP in year two is something I would look at. We agree there. But looking at those overall numbers, you can automatically assume he was perhaps overswinging, or if his BARISP was high, maybe he developed better plate discipline in those situations. But that’s something I would make a decision on based off of watching video and watching his batting approach. If we both did that, (for argument’s sake) we both see that with the bases empty he was overswinging, but with RISP, he choked up on the bat a little bit, closed his stance and moved a little closer to the plate. We both see a better situational approach. I would simply be content with seeing that better approach and the things that resulted from that. So would you, but you choose to attach more numbers to it.
      After his return more toward the mean in year 3, both of us are confident that this guy has what it takes and can be a good player consistently. You’re just putting more numbers to what we both see. This is why I believe sabermetrics are just the quantification of our eyesight and analytical baseball minds, and don’t offer a true value over traditional observations. It’s the same thing, except I’m just using words and you’re just using numbers to make the same point.

  20. Ross says:

    This is a really good discussion. Obviously Joe’s point wasn’t about whether or not SB attempts are good, but rather than both sides of the argument should become informed about the other side’s views so they can have a legitimate discussion (hilariously, many former players also offer almost nothing of value in “old school” thinking either, as so many guys just played the game without putting much thought into what they were doing).

    What’s funny about this particular topic is that it’s one of the oldest “Moneyball” themes and generally a pretty outdated concept. For example, I’ve seen a few commenters saying that teams shouldn’t try to steal so much – what teams in MLB actually attempt steals with any kind of frequency anymore? It’s actually to a point where teams should attempt to steal much more, based on run expectancy projections.

    Even the A’s themselves, still the most misunderstood team in the league by those who think Moneyball was Billy Beane’s lifelong manifesto rather than Michael Lewis’s description of the organization’s thought process, have updated their thinking on stolen bases. The point of Moneyball and sabermetrics in general is that constant innovation is necessary in a competitive marketplace, especially one where resources aren’t allocated evenly.

    In 2002 (the year Moneyball depicts), the A’s went just 46-for-66 (69.6%) in team steals (Eric Chavez led the team with 8!). Even with such a low number, they probably could have stolen less, given the poor success rate. They had a .339 team OBP that year (5th in AL), though, and scored 800 runs (8th). They hit a lot of home runs so preserving outs was the smart move, as a 2-run bomb was always in play (205 team HR was 4th in AL).

    Last season, the A’s had a pretty putrid .310 team OBP (12th in AL) and hit just .238 (13th). They still had good power (6th in HR), but with such a poor team BA and the most strikeouts in the AL, it was never a safe bet to expect too much. They stole 122 bags on 154 attempts (79%), allowing for 713 runs scored (8th in AL), which far exceeds their projection based on the poor team OBP.

  21. Ross says:

    Continued from above…

    Fact is, stealing a lot of bases at a high success rate is quantitatively destined to create a lot of scoring opportunities. Especially now, with HR down across the league and unproductive outs (strikeouts, mostly) way up, aggressive baserunning is a huge area of advantage for teams.

    Last season, the entire league stole at a 74% success rate (above the break even point). Obviously, part of this is scouting and understanding smart situations to run and part is awareness of sabermetrics. But even with a high league success rate, only 3229 total bases were stolen (108 per team). I think that number could be much higher and wouldn’t even an extra 10 well-placed steals for a team like the Angels or Rays – who each missed out on the postseason by just a few games – maybe turn a couple close losses in their favor? I don’t have any evidence to back that up, but common sense tells me there are tons of one-run games where an extra runner in scoring position would help.

    And – not just first to second, but completely overlooked anymore is stealing third, which is actually much easier to do for a smart baserunner and opens up so many possibilities. There’s such a stigma attached to the possibility of making an out at third that nobody ever tries this, even if there are guys who would be successful 90% of the time.

    This is actually the case with a lot of aspects of the game. Nobody swings at the first pitch because there’s a stigma attached to making an out on the first pitch. Well what if it’s a fastball right down the middle? I’m not saying it’s right or wrong to swing at the first pitch, but it’s definitely right to be ready in case it’s a great pitch. With so many strikeout pitchers, you can’t afford to give away free strikes. Same thing with trying to run up the SP’s pitch count to knock him out of the game. I mean, that’s fine if the guy is, say, Max Scherzer, who is a very good pitcher, not necessarily the biggest workhorse, and isn’t always spot on with his control. But it’s reasonable that getting him out after five or six innings gives you a shot at lesser bullpen arms.

    But what about Cliff Lee? It’s foolish for guys to try working the count against a guy who throws nothing but strikes and can throw 120-130 pitches without breaking a sweat. Your best move against him is to go up there ready to hit and try knocking him out of the game with the lumber. Yet teams still sit around and watch hittable pitches go by.

    Same with James Shields. Ok, so maybe he runs up a high pitch count and leaves after 6 or 7. So here come Collins, Herrera and Holland. Congratulations! So much of this seems to be based on the old reasoning that relief pitchers are relievers because they weren’t good enough to be starters. This is true in player development terms, but in real life terms those guys as relievers are throwing upper 90’s and striking everyone out. Maybe Aroldis Chapman wouldn’t cut it as a starter, but I’d rather face Bronson Arroyo in the 9th.

    Anyhow, that’s a lot of rambling to say that, yes, it would be nice of the “old school” would at least become informed about what they’re supposedly arguing against.

  22. Ross says:

    And yes, a few thoughts I had quickly became 945 words…sorry about that but I hope to hear anyone else’s thoughts on what I said

  23. Ro says:

    Joe’s a good writer, but people like him speak to the sabermetric skeptics in such a condescending way. How to win converts by implying that the old school is stupid. That’s not a good way to sell your latest WAR+adjustedforBABIP stat of the week.

    • tomemos says:

      What is Hudler being, if not condescending? Does he really think that viewers won’t understand the upside of the stolen base?

    • Wilbur says:

      I doubt that he is trying to win converts.

      And tomemos is correct. No one is more condescending than the Joe Morgans and Ken Harrelsons of this world.

    • Robert says:

      In the early days of sabermetrics, the condescension was basically always cast TOWARDS the “geeks in their mom’s basement”. The geeks eventually returned the attitude, once they were largely proven right.

  24. KHAZAD says:

    The numbers for the percentage of time you must be successful in a stolen base attempt are based on long term studies of how many runs or how often a team scores in different base/out situations in actual games. I have run some of the numbers for fun, even including the fact that about 5% of successful steals end up being for 2 bases because of errors.

    The success rate needed to increase the number of runs the team is likely to score by stealing 2nd base: With no outs, 72.87%, with one out, 72.35%, with two outs, 70%.

    However, in a late game situation, when your goal is to make sure you get one run, (for example needing one run to tie the game late, or the winning run in a tie game, or an insurance run when you are up one) the success rate needed is less, as the goal is different: 56.51% with no outs, 58.70% with one out, and 58.19% with two outs

    Going from second to third, to increase the number of runs: 76.57% with no outs, 68.27% with one out, 82.66% with two outs.

    To try and get one run in late game situation: 67.59% with no outs, 55.77% with one out, 74.92% with two outs.

    In the first and third situation mentioned in Joe’s story, it should only be attempted with a good base runner on third as part of, or at least the threat of, a double steal. Using a straight steal of second, the rates needed to increase the number of runs are pretty high: 80% with no outs, 75.85% with one out, and 79.82% with two outs. It is, of course, a little easier to steal second with the threat of a double steal, as sometimes the catcher will not make a throw, but there is no sense doing a straight steal in a late game situation, as the main increase in your chances of scoring one run is an error: In a straight steal, you need a 97.49% success rate with no outs, 86.23% with one out, 91.14% with two outs.

    Just for fun, as many of us had the childhood dream of stealing home to win the game, (Anyone else remember the old Mike Piazza ESPN commercial?) It is rarely attempted as a straight steal, because the success rate is so low. But if you ever find yourself at third with two outs in a tie game in the bottom of the ninth (or extras) you only need to be successful 27% of the time to make it “worth” trying to be the hero.

  25. mickey says:

    Mercy! I’ve read through 99 comments about sabermetrics and stolen bases, and not one mention of TWTW! The Will To Win leads to The Will To Steal. And then you Put It On The Board. Yeesss!

  26. Mark Daniel says:

    The ability to steal bases is important on the micro level, not on the macro level. Look at any season and try to correlate stolen bases to runs scored. They don’t correlate. Sometimes they correlate negatively. But on the micro level, the ability to steal a base can be quite important. As Khazad said above, down 1 run in the late innings, against an ace reliever, with a weak hitter at the plate – you might just want to move that runner over to 2nd so the weak hitter has at least a small chance of driving that guy in. The Red Sox glorious 2004 season probably would have ended disastrously if it weren’t for a single stolen base. Statistically, this single stolen base was perhaps unimportant when looking at the season as as whole, but in reality this stolen base was hugely important.

  27. news says:

    Internet Online System is very usefull and Billions of Peoples use Internet and Many of the peoples want to make money online from home and make online business for a life time
    We have the best Online earning System with Data Entry Jobs, Online Jobs, Network Marketing, SMS Earning, and Much More Online Work From home, Skype : Jobz.Corner

  28. Sarwat Tariq says:

    World’s Best Current Affairs Website which you Get Any thing you want, Most Popular Pictures, Loll Pictures, Funny Pictures, Funny Videos, Current Affairs Videos, Entertainment News and Videos, Entertainment Pictures, Hot Girls Affairs and all hot Current Affairs in the World

  29. Sarwat Tariq says:

    Glorious Live design philosophy is based on versatility and glamour with a clear focus on creating outfits to make each woman feel beautiful in whatever she wears. We also provide custom made design as per your requirement.

  30. whoman69 says:

    Sabermetrics is often fools gold. They also say strikeouts are not a negative. Bill James, king of the saber guys said in 1988 in his baseball abstract, “George Sisler is probably the only player other than Lou Gehrig who can reasonably be considered the greatest first basemen ever in terms of peak value.” Sisler was ranked 3rd in peak value and 7th in career value. In his 2003 follow up James writes, “(Sisler was) Perhaps the most over-rated player in baseball history.” He was rated the 24th best first baseman. Did George Sisler have a second career 15 years after he died that changed his numbers? Saber guys says stolen bases aren’t worth it unless they can steal first 70%, then 75%, then 80%. If the numbers are so conclusive, why are they ever changing? By that latter standard only 43 men in the history of baseball should have attempted to steal.

Leave a Reply

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