Questioning batting average, 96 years ago

A modest proposal
Predicting the 2012 Bullpen

If you’ve read this site long enough, then you’re probably familiar with the idea of linear weights and wOBA. If not, then I suggest checking out Joe’s primer. In a post at the FanGraphs Community blog yesterday, Sam Menzin presented an article from the 1915 edition of Baseball Magazine (pdf link), in which author F.C. Lane questions the idea of batting average and its accuracy. Allow me to excerpt…

Lane opens his discussion with a question: “Suppose you asked a close personal friend how much change he had in his pocket and he replied, ‘Twelve coins,’ would you think you had learned much about the precise state of his exchequer?” He goes on to compare two mens’ respective financial situations: Man A, with “twelve coins” consisting of a combination of quarters, nickels, and dimes; and Man B, with twelve silver dollars. Saying both men have equal financial means is equivalent to the system of tracking batting averages, he explains. “One batter, we may say, made twelve singles, three or four of them of the scratchiest possible variety. The other also made twelve hits, but all of them were good ringing drives, clean cut and decisive, three of them were doubles, one a triple, and one a home run…Is there no way to separate the dimes from the nickels and give each its proper value?” Sound familiar?


This issue was not solely unique to Lane’s inquisitiveness. John Heydler, secretary and future president of the National League, added, “that the system of giving as much credit to singles as to home runs is inaccurate to that extent. But it has never seemed practicable to use any other system. How, for instance, are you going to give the comparative values of home runs and singles?”

Lane goes on to use an example of two players, one with a higher batting average and lots of singles and another with a lower batting average but lots of extra base hits. He compared each players’ hit rates (singles, doubles, triples, homers) to the league average, which is essentially an early version of wOBA and wRC+. It’s very fascinating stuff, a nearly hundred-year old article questioning the merits of a statistic still valued so highly today. I suggest clicking the links above and reading both articles, Lane’s and Menzin’s. I really can’t recommend it enough, it’s amazing stuff.

Full Disclosure: Our own Larry Koestler edited the post for Sam. Not that that means anything, just figured I’d mention it.

A modest proposal
Predicting the 2012 Bullpen
  • CP

    would you think you had learned much about the precise state of his exchequer?”

    Love these quotes from long ago. Exchequer?

  • Soriano Is A Liar

    Super interesting. One thing that always surprises me is how fans of traditional stats are so attached to, for example, ERA, with its convoluted rules, but they don’t want to get behind another stat that attempts to quantify pitching skill, such as FIP or RA, merely because it’s new, and they trust the old stat. Very cool to see people were curious about the effectiveness of ‘traditional’ stats over a hundred years ago.

    • nsalem

      I question the efficiency of the new stats almost to the extent that I question the old. Your kids may very well be laughing at FIP and RA in the same way that you laugh at Wins and RBI’s. It’s inevitable.

      • Soriano Is A Liar

        Very good point. I just think it’s funny that some people are will to question the new stats but blindly accept the old. If you’re going to question the new set, I think you kind of have to be skeptical about the old set too.

      • Plank

        You question the new metrics because you don’t think they are representative of baseball skill or because you think there are better metrics available?

        Those are two very different reasons for not liking the new metrics.

        RBIs and Wins are indicative of skill over a long time period too, they just aren’t very good from year to year.

        • Ted Nelson

          I read those are being the same reason, not very different.

          *They are not as representative of baseball skill as better (perfect) metrics*

          • Plank

            What are better metrics?

            I’m saying disregarding a metric because you think it isn’t perfect because there is some theoretical metric out there uninvented yet is different from disregarding a metric because there is a metric that is available and is shown to be better.

        • nsalem

          I like many of the new metrics and their availability many are more appealing to me then others. I sometimesI question stats such as HR/9 or FIP because they don’t and maybe can’t take into account the situations. Should a solo Home Run given up when the score is 6-0 count the same as a 2 run Home Run yielded when the score is 2-0. Was the walk given up to a pitcher in a one run game or was it a lefty pitching around an Albert Puljos type with a man on second and a left handed batter on deck who does not hit left handers well. I would imagine FIP penalizes types like Luis Tiant, Jimmy Hunter and El Duque because of their craftiness and their willingness to take one step back to go two steps forwards. It’s an old meme which probably can be argued till the cows come home. I view most stats (Including FIP) to be both informative and interesting, but because their are so many variables in baseball they might not tell the whole story 100% of the time. Maybe my objections are to those who think it does.

          • Plank

            Should a solo Home Run given up when the score is 6-0 count the same as a 2 run Home Run yielded when the score is 2-0.

            Over the course of a season or longer? Yes. Over the course of a game or a two game stretch? No.

            Those variations you are talking about, which are important in individual games, get evened out over a long enough time frame.

            • nsalem

              Do you have proof of that? I don’t know of any. If you do please let me know. I have to run now, but would be happy to continue discussion at the next open thread.

              • Plank

                I don’t have any proof. Someone has probably done that study. What leads you to believe some players give up home runs disproportionately in certain situations?

  • FIPster Doofus

    “Practicable.” “Exchequer.”


  • Tyler

    And this is why I love RAB so much…well done.

  • Jose M. Vazquez..

    Under The Norman kings of England the Exchequer meant the treasury of the state. By the time the article was written in the early 20th century it meant the funds in your pocket or whatever valuables you had.

  • Will (the other one)

    After reading this article, is anybody else convinced that “Baseball Magazine” is an elaborate hoax conceived and written by Dave Cameron?


  • Evan

    Batting average as used today may not be the best means of deciding who to give a 200 million dollar contract, but it does have its uses. If I’m down by 1 run in the 7th, 8th or 9th inning, bases loaded, I want the guy with the highest average possible. I want Derek Jeter over Prince Fielder. I’m not even too concerned about OBP because it’s unlikely that the pitcher will issue a walk in that situation.

    I agree that average is given too much importance when dealing out contracts and determining league awards, but it is still a very useful stat when used properly.

    • RetroRob

      Batting average is a useful statistic; it’s just that it’s not as useful or meaningful as many people think it is. Certainly has more relevance than RBIs or pitcher wins. If BA didn’t exist today as a statisitc, the sabermetric community would invent it!

    • whozat

      But you’re using batting average here as a proxy for contact rate, which is a stat we have. Why not just make your decision based on that?

      Also, you’re acting like walks are something pitchers give out by choice. The plate discipline of the hitter plays a part as well.

  • rfwarrior

    Serious question. Can someone please explain to me why advance stats always push walks as a highly rated stat. I noticed in the wOBA calculations (as well as other stats I’ve seen), walks are just a notch below singles.

    How is it possible, that a walk, is anywhere near as valuable as a single? Singles can result in an RBI with a man on 2nd, or 3rd in nearly any situation. They can allow a runner to advance from 2nd to 3rd or even 1st to 3rd.

    Walks (while they do have the advantage of negating a double-play) only advance runners in force situations, and only result in a run when the bases are juiced.

    How do they come up with such close “weights” for a walk and a single? Seriously, please do explain.

    /sorry for the super-long rant

    • Plank

      Do you mean for hitting stats? I don’t know exactly, but walks are much more likely to maintain a similar rate from year to year, so a player with a high walk rate is more valuable than one without a high walk rate going forward if they have similar numbers.

    • whozat

      I think it’s mostly because of a couple things:
      1) the relative frequency of situations where the single drives a run in (man on 3, fast man on 2, men on 2/3, men on 1/3, men on 1/2/3) vs the ones where it doesn’t (bases empty, man on 1st, slow man on 2nd, men on 1/2).
      2) the fact that not making an out gives the next guy a greater chance to drive in more runs than he would have had otherwise, by putting someone on base and possibly advancing a runner into scoring position.

      I’m sure The Book ( has more detail.

    • Mickey Mantle’s Outstanding Experience

      A single’s run value is 25% higher than a walk’s, so it’s not that close.

      But the reason is that situations that a walk is as good as a single are much more frequent than situations where the single is a lot better.

      56% of PAs come with the bases empty, and 18% come with a runner on 1st (the run value of a single is higher here, but not that much higher because the runner is only able to go 1st to 3rd less than 30% of the time.)

      Compare that to situations you mentioned where a single is much more valuable than a walk – Man on 2nd: 9%, Man on 3rd:3%, men on 2nd and 3rd:2%.

      In the 3/4 of PAs, a walk is as good as a single. In the remaining 1/4, a single might be a little better (advancing a runner from 1st to 3rd) or a lot better (runners on 2nd and 3rd), but when you average it out, you get the rates in the wOBA formula.

  • Kevin

    Although this may be a fallacy of the golden mean, I think it’s best to look at as many different statistical formulations as practicable and see if they agree. Especially with advanced statistics like the different WAR formulas where a single aspect of the game is given so many different weights that there can be vast swings.

  • LiterallyFigurative

    I think you just have to take as much data as possible when it comes to making decisions on players and the mythical “who’s better” arguments. The eye test is great, but decieving, due to where you are playing and who’s on your team, who’s on base, etc.

    Some folks on here down the value of wins, but praise HR/9 and K/9, as examples. The ability to win games as a pitcher is dependent on your team, but some guys are better at finding their way through the 5th and 6th innings with the lead intact than others. Some guys give up a decent amount of Hr’s, but will give up solo Hr’s more than the other guy who gives up maybe 50% fewer hr’s, but have men on base.

    High K guys tend to be high BB guys, high fly ball % guys, and high pitch count guys. He may strike out 6 guys in 5 innings, but it took him 105 pitches to do so, exposing the middle relief guys and affecting how the manager gets through the final innings. The second guy may have a higher contact rate, but a softer contact quality, allowing him to pitch 6 2/3rd innings or 7.

    I just think the dismissal of “traditional” stats is a gross overreach by the new-stat proponents, just as the ignorance of the new stats by old-heads is nuts as well. You have to see the whole picture, not just the black and white sections, or not just the 3D sections either.

    • Mickey Mantle’s Outstanding Experience

      Some guys give up a decent amount of Hr’s, but will give up solo Hr’s more than the other guy who gives up maybe 50% fewer hr’s, but have men on base.

      Can you provide an example of who you mean?

      High K guys tend to be high BB guys, high fly ball % guys, and high pitch count guys. He may strike out 6 guys in 5 innings, but it took him 105 pitches to do so, exposing the middle relief guys and affecting how the manager gets through the final innings.

      Again can you provide an example? Who are these pitchers that are only completing 5 innings or so a start and are praised as great pitchers by advanced stats?