How Bases on Balls were Scored: 1864–1888

This article was written by Richard Hershberger

This article was published in Spring 2018 Baseball Research Journal

In 1876 a base on balls was charged against a batter’s average, then 11 years later in 1887 it is credited to the average. These anomalies were the manifestation of a decades-long discussion on how to think about the base on balls. This discussion only affected averages in those two years, but continued beneath the surface in the intervening years, finally arriving at the modern conclusion in 1888. includes “A note about statistics from 1887 and 1876,” signed in 2001 by Sean Forman. In it, the inimitable site founder explained:

1887 – During this year, Major League Baseball and the guides reporting on that season decided that a walk was equivalent to a hit. This was the only season in which this was true. Total Baseball and other encyclopedias have at times decided to honor, and at other times decided not to honor this scorekeeping method of the time and compute a player’s batting average with walks as hits and at bats for that season. I had originally decided to go along, but now have changed my mind and have changed back to the standard method of computing batting averages and other stats.

1876 – Additionally, TB has at times recognized the convention that a walk was a charged at bat in 1876, so a player with four walks in four plate appearances will be credited with an 0 for 4 batting. I have decided to backtrack on this as well and restore the stats to as they were previously.[fn]Sean Forman, “A note about statistics from 1887 and 1876,” December 2, 2001,[/fn]

Forman’s note raises an important question: What were they thinking? In 1876 a base on balls is charged against a batter’s average, then 11 years later it is credited to the average, with the modern practice followed in between. This seems inscrutable.

These anomalies were the manifestation of a decades-long discussion on how to think about the base on balls. This discussion only affected averages those two years, but continued beneath the surface in the intervening years, finally arriving at the modern conclusion in 1888. The subject was closed for decades afterward, until it was reopened in the form of on-base percentage.


Three individuals have roles to play in this story. The first is Henry Chadwick, the Hall of Fame journalist. He rose to prominence during the Civil War and spent the later 1860s and ’70s as the premier baseball journalist and writer of numerous baseball guide books. He took a particular interest in scoring and statistics. The use of the letter K to denote strikeouts is a direct survival of his scoring method.[fn]Henry Chadwick, Beadle’s Dime Base-Ball Player (1862), (New York: Beadle & Co., 1862). Chadwick’s reason for using K rather than S is unclear. The explanation typically offered is that he needed S for some other purpose, usually either a sacrifice or a single. Neither stands up to scrutiny. Sacrifices weren’t scored until decades later. Avoiding confusion with singles is more plausible, but the early scoring method recorded how players were put out, not, for the most part, how they avoided this fate. “H.R.” for home runs was the exception, but no provision was made for recording singles, doubles or triples, and the letter S was not used in the early system. The only explanation Chadwick offers is that “we use the first letter in the words Home, Fly, and Tip, and the last in Bound, Foul and Struck, and the first three letters of the alphabet for the first three bases.”[/fn]

Even as his influence waned in the 1870s, his opinions on scoring were widely respected.

The second individual is Nicholas Young. He was a major presence from the 1870s into the 20th century. In 1871, as secretary of the Olympic Club of Washington, he proposed what turned into the founding meeting of the National Association of Professional Base Ball Players, the first professional organization.[fn]“The Coming Season: The Professionals Ready for Business,” Sunday Mercury (New York), March 5, 1871.[/fn] When the National League was founded in 1876 he was appointed league secretary. He added the office of National League president in 1885 and remained the NL president and secretary until his retirement in 1902. His duties as secretary under the NL constitution included the preparation of “a tabular statement of the games won and lost by each club.”[fn]National League Constitution for 1876, Article XII, Sec. 5.[/fn] The strict requirement was merely for the final standings, but Young chose to interpret his mandate broadly by preparing and publishing a collection of statistical analyses. His opinions of what statistics to compile were accepted as the de facto standard.[fn]Young’s statistical tables were published by many newspapers. They can also be found in the annual Spalding Guides starting with 1877.[/fn]

The third individual is less well known today: Alfred Wright. He was the baseball editor for the Philadelphia Sunday Mercury in the late 1860s and ’70s, as well as the official scorer of the Athletic Club of Philadelphia. He also was said to have a remarkable memory for records, able to “give scores and dates of games played between first-class clubs since the game has become popular,” making him baseball’s earliest known trivia maven.[fn]Sunday Republic (Philadelphia), June 25, 1871.[/fn] He kept the most detailed record of any club scorer, was forward thinking in what records should be kept and how they should be analyzed, and used his newspaper position to put his ideas into practical form. Close examination of the Sunday Mercury’s annual compilation of the club’s records in the late 1860s and early ’70s shows innovations that would later become standard.


Baseball statistics arose in the 1850s. The basic offensive calculation was always understood to be an average: a fraction with a numerator and a denominator. This was widely understood because that discussion had already taken place in the context of cricket. Early baseball statistics were directly borrowed from cricket. The question was what values would constitute the numerator and denominator?

The early box scores recorded just two items for each player: hands lost (i.e. outs) and runs. Both were used for early averages. Beadle’s Dime Base-Ball Player for 1862 includes statistics for players in prominent clubs. The tables include games played, hands lost and runs, along with averages of hands lost per game and runs per game (recording the quotient as an integer and a remainder, just as schoolchildren are taught division to this day).[fn]Chadwick, Beadle’s Dime Base-Ball Player (1862).[/fn] Beadle’s would publish both averages through its 1869 edition, but in the meantime two developments made this obsolete: Outs per game dropped out of widespread use as its limitations became obvious, such as penalizing a player put out on a force play, and it became clear that runs per game was not a useful measure.

The original runs per game standard was copied directly from cricket. The structure of cricket is beyond the scope of this article, but suffice it to say that runs per game is a much more reasonable standard in cricket than it is in baseball. Among other differences, cricket lacks the concept of the base hit: that is, an intermediate goal for the batter beyond mere survival that does not necessarily result in an immediate run.

An aside: Baseball statistics could have explored outs per game. The modern insight behind OBP is that the batter’s goal is to not make an out. If he and enough of his teammates successfully avoid making outs, the scoring of runs will take care of itself. This insight would have lent itself to the outs-per-game average. Add the concept of the fielder’s choice and a version of outs per game might have formed the basis for offensive averages, and the rest of this story would have been very different. But Chadwick took this in a different direction, abandoning outs per game and instead adopting the runs-per-game average by changing the numerator.

His first experiment was with what we now call slugging average. He realized that runs scored was a flawed standard because a player’s score depended on the actions of his teammates. He offered a better standard:

The true estimate of a batting score . . . is the number of bases made on hits. Thus, a player making his 1st base twice, his 2d once and his 3d once, and getting home but once, thereby being left four times and scoring but one run, makes a better score than the player who makes his 1st base four times by his hits, and yet gets home every time by the good batting of the players following him. The score of batting never tells the truth, as one man may be credited with six runs who get his base twice or three times on miss-catches or wild throws, while another player may be credited with but one run and may have made his base each time by clean hits and yet have been left.[fn][Chadwick?] “To Correspondents,” New York Clipper, November 24, 1866.[/fn]

Chadwick soon dropped total bases made in favor of “first base hits.” He did not mean singles, as this is sometimes misinterpreted. Rather, it was the number of times that the batter hit the ball and reached at least first base, without reference to any additional bases. “First base hits” came to be shortened to “base hits” or simply “hits.” It wasn’t quite the modern rule since the fielder’s choice was not yet factored in, but it was the same idea as the modern base hit. He favored the first base hit over total bases made because the fielding of the day was too erratic to reliably assign credit:

There can be no mistake about the question of a batsman’s making his first base, that is, whether by effective batting, or by errors in the field, such as muffing a ball, dropping a fly ball, or throwing badly to the bases, whereas a man may reach his second or third base, or even get home, through errors of judgment in the out-field in throwing the ball to the wrong man, or in not properly estimating the height of the ball — errors which do not come under the same category as those by which a batsman makes his first base.[fn]Chadwick, “The True Test of Batting,” Ball Player’s Chronicle, September 19, 1867.[/fn]

In other words, too many extra-base hits resulted from what could unkindly be called “Little League plays,” even if they didn’t involve errors in the strict sense of an errant throw or a muffed catch. Modern players train in best practices of where fielders should go in any situation and where the ball should be thrown. The more thoughtful players were working out what these best practices were, but the play of even the best clubs in 1867 was erratic by modern standards. Rather than attempting to assign a value to such hits, it was better to stick to the more straightforward question of whether the batter reached first on a hit or an error.

The new scheme carried with it a new role for errors. Scoring errors was not itself new. Scorers had intermittently noted fielders’ missed catches since the late 1850s. These early examples had, however, no broader significance. They were reportorial harrumphs directed at the erring players. Chadwick instituted his new regime of considering bases made, whether first or total bases, by adding a column to the box score for first base hits (initially abbreviated B and later 1B) in addition to the old runs and outs. It was apparent that in some cases the batter made his base(s) not through his own efforts but due to fielding errors. It didn’t seem right to credit the batter for these errors. As the New York Clipper advised in 1869, the “first base hits” column included only those made on “clean hits”:

In making up a score at the close of the match the record should be as follows:–Name of player, total number of times the first base was made by clean hits, total bases so made, left on bases after clean hits, and the number of times the first base has been made on errors, which include called balls, wild throws, dropped fly balls, muffed balls and bases made by the ball thrown to other bases to put out players forced off by poor hits.[fn]“On Scoring In Base Ball,” New York Clipper, February 20, 1869. For early notes of missed catches, see, e.g., the summary to the box score of the game of August 10, 1859: [Chadwick], “Knickerbocker vs. Empire” New York Clipper, August 27, 1859. The first base hit column debuted in the box score for the Athletic of Philadelphia vs. the Union of Morrisania game of August 19, 1867: [Chadwick], “The Athletic Club Matches in Brooklyn,” Ball Players Chronicle, August 22, 1867.[/fn]

This scheme was in fact more elaborate than as yet appeared in practice, but even in the simpler form used, there were enough data to calculate the batting average. What mattered for that was to not assign undeserved credit to the batter. First base reached on clean hits filled this need. Those reached on errors did not enter into the calculation.

This was a quietly radical change in the role of the scorer. He no longer merely recorded events. He now sat in judgment of the players, with his decisions affecting subsequent assessments of their abilities. This was a huge new responsibility. Chadwick’s decision to favor first base hits over total hits was an attempt to minimize this new role, only slightly lessening the impact of the change.

Chadwick started including first base hits in box scores and other scorers started copying him. By 1869 the practice was widespread enough that Chadwick had the data to calculate batting averages in the 1870 edition of Beadle’s as first base hits per game.[fn]Chadwick, Beadle’s Dime Base-Ball Player (1870).[/fn] He also inspired Alfred Wright to experiment with more elaborate record keeping. Wright’s box scores and end-of-season compilations had previously included only the conventional items. Then, in the November 22, 1868, issue of the Philadelphia Sunday Mercury, he came out with a season record of the Athletic Club that was downright baroque. It included as separate items the various ways a player could be put out, the corresponding ways that a fielder could put a player out, and various miscellaneous data such as times left on base and home runs.[fn]Sunday Mercury (Philadelphia), November 22, 1868. The ways a player could be put out were: fly ball; foul fly ball; foul bound ball (i.e. foul ball caught on the first bounce, which at that time was an out); at first base; at second base; at third base; at home base; run out between bases; and struck out.[/fn] No other club published anything like this as early as 1868. Wright didn’t seem to have any clear idea of what to do with all this information. The only analysis was the conventional averages of outs per game and runs per game. But he clearly sensed that there was a use for these additional data and he kept playing with them.

The breakout year for using first base hits to calculate batting average was 1869. Wright’s end-of-season compilation in the November 28, 1869, Philadelphia Sunday Mercury develops the data some more. He includes, in addition to the conventional averages, “Average Times Bases on Clean Hits” and “Average Number of Total Bases on Clean Hits.” In other words: batting average and slugging average (which he had fewer qualms about than did Chadwick). His data also included a novel entry: “Total Number of Times at the Bat.” Other top clubs copied Wright’s extensive data collection and his leap to a new form of batting average.[fn]The Atlantic Club’s averages are in “Review of the Season of 1869. Its Victories and Defeats,” New York Clipper, December 11, 1869; and the Unions of Lansingburgh’s in “Review of the Season of 1869. Its Victories and Defeats,” New York Clipper, December 18, 1869.[/fn]

Wright in 1869 still used games played as the denominator but scorers now had the tools necessary for the modern denominator. This leap was suggested in 1871 by Hervie Dobson, a correspondent for the New York Clipper:

According to a man’s chances, so should his record be. Every time he goes to the bat he either has an out, a run, or is left on his base. If he does not go out he makes his base, either by his own merit or by an error of some fielder. Now his merit column is found in “times first base on clean hits,” and his average is found by dividing his total “times first base on clean hits” by his total number of times he went to the bat. Then what is true of one player is true of all, no matter what the striking order.[fn]“The Professional Club Secretaries’ Meeting,” New York Clipper March 11, 1871.[/fn]

No one did anything with the idea at first and it is not clear that anyone even took notice of Dobson’s proposal. The leap was finally made in 1874 by several clubs, using batting average in its modern form.[fn]The Athletics in the Philadelphia Sunday Mercury, December 6, 1874; the Bostons in the Philadelphia Sunday Mercury, November 29, 1874; the Hartfords in “The Hartford Club Record,” New York Clipper, December 12, 1874.[/fn] Chadwick still had to use the older form of averages in the Beadle’s Dime Base-Ball Player. Not every scorer was up to date, so complete data weren’t available, but the advantage of using at-bats rather than games played was obvious and the trend favoring this advanced statistic was clear.

By 1876 every club in the new National League was calculating batting averages as first base hits per at-bat. This was made quasi-official by Nicholas Young. He issued an annual league book, beginning in 1877. This was mostly intended for internal use. It included not only the playing rules but also business documents such as the league constitution and minutes of its meetings. And he appended a statistical record of the previous year.[fn]Constitution and Playing Rules of the National League of Professional Base Ball Clubs, (Chicago: A. G. Spalding & Bro., 1877). And following years.[/fn] These statistics were disseminated in the sporting press, making them the consensus standard.


Nicholas YoungHow did bases on balls fit into these developments? Poorly. Bases on balls were not an original feature of baseball. The early game was conceived as being about fielding and throwing and running. Batting was merely how the process got started. The pitcher’s role was to deliver the ball where the batter could hit it. The batter’s role was to hit it, at which point the fun began. As baseball grew more competitive, clubs started gaming the system. Pitchers started throwing the ball as fast as they could, often sacrificing control for speed. The rules required the pitcher to deliver the ball “for the striker” but originally with no penalty for failure. The base on balls was added in 1864 to address this absence. The umpire was empowered to call a ball should the pitcher, after due warning, persist in pitching unfair balls. Should the pitcher refuse to be recalled to his duty, the umpire could call a second and then a third ball. Upon the umpire calling three balls, the batter was given first base.[fn]For a good summary of these developments, see: Peter Morris, A Game of Inches: The Stories Behind the Innovations That Shaped Baseball, (Chicago: Ivan R. Dee, 2006).[/fn]

Such was the theory. In practice, the innovation was widely considered strange, umpires were reluctant to embrace their expanded role, and the new rule was mostly ignored. Chadwick was an early advocate of the rule and devoted much ink to wheedling umpires to enforcing it. For two years, the rule was nearly a dead letter. Finally, in 1866, Chadwick’s efforts paid off and some umpires began calling balls. The rule was still thought unnatural and some batters were reluctant to take first base, so walks remained rare.[fn]See, for example, accounts of the Union of Morrisania vs. Surprise May 19, 1866, game in the New York Sunday Mercury, May 27, 1866; and Eureka of Newark vs. Union of Morrisania June 12, 1866, in the New York Sunday Mercury, June 17, 1866.[/fn] Chadwick, in his role as chairman of the rules committee, made the language of the rule ever more imperative on the umpire, with limited success. The rules makers finally took the extreme measure of changing the rule to nine balls for a walk in 1875. At that point, calling a ball was a small enough penalty that umpires were willing to do it. This was followed by a gradual reduction in the number of balls needed for a walk in reaction to improvements in pitching—overhand delivery and curveballs. The modern number was finally reached with the 1889 season.[fn]Richard Hershberger, “When Did Umpires Start Calling Balls and Strikes?” Protoball, July 2014.[/fn] Pitchers gave up under one walk per game through most of the 1870s, with the number gradually rising in the 1880s to the modern level of about three a game.

Batting average in its primitive form of runs per game predated bases on balls. Even once bases on balls entered the game their rarity resulted in little pressure to consider how they should be integrated into batting average. Tacit consensus took the place of discussion. The one thing that everyone agreed about concerning walks was that they were errors by the pitcher. His job, after all, was to put the ball over the plate. Failure in this simple task was considered much like any other throwing error. The only way they were treated differently was that they were not usually tabulated in the main box score, but relegated to the summary.[fn]Chadwick, “On Scoring In Base Ball.”[/fn]

This finally brings us to the 1876 rule, which counted walks as errors and therefore hurt the batters’ averages. It is not strictly true that this was the rule, as true official scoring rules had not yet been enacted. Nor was it strictly speaking a change from earlier practice. Walks had never counted in the numerator of the batting average. When the denominator was games played, walks were a lost opportunity for a hit in the same way as were errors. The change was that when the denominator was games played, the occasional base on balls was barely noticeable in the average. With the denominator now at-bats, the effect of bases on balls on the batter’s average was more apparent.


The National League introduced official scoring rules for 1877, defining the data official scorers were required to report to the league. These data were to be tabulated in seven columns. The first column listed for each player “the number of times he has been at bat during the game.” This codified the “at bat” statistic that was already established practice, but with one change: “Any time or times where the player has been sent to base on called balls shall not be included in this column.” This was a remarkable conceptual leap, transforming “at bat” from a straightforward description into a term of art. The change was enacted with remarkably little discussion. It seems to have helped that few understood it. Lewis Meacham, a baseball reporter present at the meeting, wrote:

One of the most amusing features of the session was the utter and blank amazement and despair with which the most practical of the Managers listened to the above [scoring rules], without the dimmest idea what it was all about.[fn]”Pastimes: Convention of Base-Ball Managers at Cleveland,” Chicago Daily Tribune, December 10, 1876.[/fn]

Meacham does not tell us who suggested the idea, but Nicholas Young is a good guess, being present, in a position to make the proposal, and being the individual most directly interested in the topic. (It certainly was not Chadwick, who was not present and who was on poor terms with the National League in its early years.) Not until a month later did anyone think to explain the new rule, giving the reasons we would expect:

One of the papers which object to the League system of scoring, asks why a “time at bat” should not be given when a player is sent to base on called balls. The answer is simple: Because, if the umpiring is correct, he didn’t have his proper opportunity to hit the ball, and inasmuch as the time at bat has a tendency to lessen the player’s score, it would be unfair to give it to him unless you also gave him a chance to increase his score of clean hits.[fn]”Hop-Scotch: The Curling Tournament at Lincoln Park, with an Outline of the Game,” Chicago Daily Tribune, January 14, 1877. Note that the Tribune often covered a variety of sports in a single column, like a modern “Digest” or “Roundup” column, and stacked many headlines on top of the piece. Citations for such columns are for the top headline.[/fn]

This new definition of the “at bat” set the standard for considering bases on balls. They were omitted from batting average entirely, neither contributing to nor deducting from a batter’s average. It did not follow, however, that bases on balls were no longer regarded as errors. Rather, they were counted as errors that were not included in the batting average. The question over the next decade was how exactly to treat them when assessing pitchers as fielders. Were bases on balls errors the same as any other, so far as the pitcher was concerned, or were they a category distinct from other errors?

The 1877 rules did not explicitly address this question. Errors had their own column in the tabulation, defined as “each misplay which allows the striker or base runner to make one or more bases when perfect play would have insured his being put out.” Young interpreted that in his record for the year as treating bases on balls as any other error and noted this on the blank score sheets distributed to official scorers.[fn]“Pastimes: Opening Game of the Chicago Ball Season,” Chicago Daily Tribune, April 22, 1877.[/fn] Young’s season record for 1877 did not include a separate category for bases on balls, but it nonetheless pointed the way. He separated wild pitches from fielding errors in the pitchers’ records and passed balls from fielding errors in those of the catchers.

This logic was extended in the 1880 rules to bases on balls. The rule that year approached errors from the opposite direction. There was no error column, but it was implied. The fielding record had previously listed outs made, assists, and errors. Now it had outs made, assists, and chances offered. Chances offered was simply the sum of a fielder’s outs, assists, and errors. Errors could still be determined, but now required arithmetic. The definition of chances offered explicitly excluded called balls, passed balls, and wild pitches. These three came to be known as “battery errors” and the 1880 rules recognized that they were a different sort of error from fielding errors, and counted them separately. They were still recorded but were to be relegated to the summary rather than tabulated in the box score proper.[fn]Constitution and Playing Rules, 1880.[/fn]

This was a controversial change. Many reporters and scorers didn’t understand the point. The result was a back and forth debate, with one faction controlling the rules one year and the other the next. Bases on balls were put back in the error column for 1883. Wild pitches and passed balls joined them for 1885, only to be removed again for 1886. The decade of 1877–86 turns out to be less modern than it superficially appears. It was modern in that there was a consensus that bases on balls should not count against the batter’s average, but there was no agreement on how they should be assessed against the pitcher.


The 1887 season would see something entirely new: bases on balls scored as hits. This was a notable year in the development of the rules. The National League and American Association determined for the first time to appoint a joint committee to amend the rules for both organizations. Even more remarkably, they invited the players to submit suggestions to their meeting in the fall of 1886. Most remarkable of all, they took the suggestions seriously.[fn]“The Joint Rules Committee Appointed,” Sporting Life, June 23, 1886; “A Great Week. Three Important Meetings to be Held,” Sporting Life, November 17, 1886.[/fn]

The nascent Brotherhood of Professional Ball Players had been proclaimed the previous August. The invitation to suggest rule amendments was a golden opportunity, both for its own sake and as good publicity. They held a meeting that November to draft a set of proposals. The most lasting idea to come out of this meeting was the modern pitching delivery rule, specifically the set position. Scoring a base on balls as a hit came out of the same meeting:

A base hit should be scored for a base on balls; this would reward the batter and properly punish the pitcher; it would remove the cause of much internal dissension between the captain and players and would avoid the tiresome exhibition of pitchers giving bases on balls to save their records.[fn]J. F. B., “The Players: The Meeting of Their Executive Council,” Sporting Life, November 17, 1886.[/fn]

There was a new idea here: that working the count for a walk merited reward. The idea of working the count was not new, but earlier it had not met with approval, as seen in this report from 1870:

When [Forest City] had bravely crept ahead of the Brooklynites [i.e. the Atlantics], the latter resorted to the mean, barefaced dodge of waiting—taking advantage of the reprehensible leniency of the umpire—to worry the pitcher or get a base on . . . balls.[fn]“The National Game. Forest City, of Rockford, Ill., vs. Atlantic, of Brooklyn—The Atlantics Whipped,” New York Herald, June 1, 1870.[/fn]

The strategy was gradually accepted over the following years. When, in the Spring of 1886, Harry Wright coached his players to “take every chance at the bat by waiting for a good ball or secure the base on balls,” it was reported with no hint of disapproval, as it clearly was smart baseball.[fn]“The Local Championship: The Philadelphia Club an Easy Winner,” Sporting Life, April 21, 1886.[/fn] From there, it was a natural step to reward the batter for his smart play, and the joint rules committee accepted the players’ idea, counting a base on balls as both a hit and an at-bat.[fn]O. P. Caylor, “New Rules. Radical Changes by the Joint Committee. The Game to be Almost Revolutionized in Pitching and Batting,” Sporting Life, November 24, 1886.[/fn]

There was, of course, the inevitable opposition. The Boston baseball reporters, who were prone to collective contrarianism, were particularly vocal:

The great objection to this rule is the gross deception that it entails in the score. The score is arranged for the convenience of the public, that it may see as nearly as possible, and as correctly, just how the game was played and what each man did. If a man is credited in the score with making three hits, in order to see just what he actually did, the reader must consult the summary and subtract the number of times he got his base on balls to get at the real batting. Take that Athletic player who in one week this season led his nine in batting, having made four base hits according to the score, but in reality he didn’t make any hits in the week but got his base four times on balls. How ridiculous! It is a shame that the public is compelled to submit to such an imposition.[fn]Unidentified Boston correspondent, “Base Ball. All the News About the Players and the Clubs,” Times (Philadelphia), April 10, 1887.[/fn]

The critique was two-fold. The idea that a base on balls was the same accomplishment as a hit was disputed. Nor, it turned out, was there universal acceptance that working the count for a walk was desirable as this argument in Sporting Life makes clear:

That lying perjurer, the base-on-ball-base-hit, is seriously injuring the game with patrons. The invitation given to the batsman by the four-strike-five-ball business to tiresomely wait for a base on balls and be credited with a base hit is wearying spectators and prolonging the game and bringing censure on the umpire. It is making record players. In the last Athletic game one player, Davis, who seldom can hit a ball anyway, and that day couldn’t touch it with a boxing glove, had a batting (?) average of .600, while Griffin, a hard hitter and an emergency batter, too, who banged the ball all over the lot when bases were full, had a batting record for all this of .400, two hundred per cent. less than the man who tired out spectators, and yet readers of that score in other cities were probably saying to themselves—”What a slugger that Davis is.”[fn]Albert Mott [T. T. T., pseud.], “From Baltimore: Kilroy’s Wonderful Work Against the Athletics,” Sporting Life, April 27, 1887.[/fn]

The rule was kept in place through the season of 1887 despite a movement for its immediate repeal. There was a steady stream of complaints throughout the season leading up to the rules committee’s meeting in the fall. Chadwick (past his prime, but still influential in scoring matters) finally came out against the rule, arguing on the basis of the purity of the statistical record:

One result of the adoption of this exceptional rule has been to play havoc with the batting averages, which it has made utterly useless as a criterion of batting skill. It has also materially interfered with the value of the pitching averages. In fact, it has destroyed the usefulness of the averages in question as a basis of estimating the relative skill of batsmen and pitchers.[fn]Chadwick, “Chadwick’s Chat: The Old Man’s Estimation of the World’s Championship Series,” Sporting Life, October 26, 1887.[/fn]

Batting average had taken its modern form a decade earlier, and by 1887 what numbers constituted a good or a bad batting average were established in the public mind. Meddling with this was confusing, and just seemed wrong. The rule was changed back for 1888, with bases on balls counted as errors but not as at-bats.

This proved unsatisfactory thanks to another innovation: the earned run. The idea of distinguishing between earned and unearned runs was not new. It went back at least to 1871. Here is an early discussion of what constituted earned and unearned runs:

The difference between runs which are earned and those which are not is as follows:—If the first striker is missed on the fly, the second gets his base on a wild throw, and the third by a ball muffed by the fielders, then the batting nine escape a whitewash, and if any runs are afterwards scored, no matter if obtained by base hits, no run is earned. But if the first striker makes a base hit, and the second makes a three base hit, and the next three are out on foul balls, one run is earned.[fn][Chadwick?], “To Correspondents,” New York Clipper, August 5, 1871.[/fn]

This clearly was not definitive, and the topic was regularly discussed in later years. This discussion, however, was unofficial. The official rules did not distinguish earned from unearned runs until 1888. Having an official definition to disagree with wonderfully focused sportswriters’ attention. The rule as implemented presented an especially irresistible target: An earned run was defined as one unaided by errors, with the explicit exception that “bases on balls though summarized as errors, shall be credited as factors in earned runs.”

This contradiction was met with widespread outrage. The whole point of an earned run is that it is achieved without benefitting from an error. It seemed obvious nonsense to turn around and include runs achieved from bases on balls. Sporting Life editorially characterized this as “ridiculous,” and was far from alone in this assessment.[fn][Francis Richter], “Results,” Sporting Life, November 23, 1887.[/fn]

The problem was a lack of clarity about the purpose of the distinction between earned and unearned runs, and difficulty integrating the ideology of earned runs with the ideology of bases on balls. Was earned runs an offensive statistic meant to determine whether to assign credit to the batting side? Or was it a defensive, pitching statistic meant to determine whether to assign blame to the pitcher? It seems to have been originally conceived as an offensive stat, but came to be seen also as a defensive stat by 1879, when Chadwick—using italics to stress his point—wrote: “There is but one true estimate of a pitcher’s excellence in playing the position, and that is in the number of earned runs charged against him.”[fn][Chadwick,] “The Professional Season of 1878,” New York Clipper, January 11, 1879.[/fn]

These two conceptions existed side by side through the 1880s, leading to the disconnect. If earned runs are an offensive statistic, and if a base on balls is an error by the pitcher rather than an accomplishment by the batter, then it would make no sense to credit the offense with an earned run off a base on balls. If earned runs are a defensive statistic designed to assess whether to assign blame to the pitcher, then it would make no sense to exclude runs obtained from the pitcher’s inability to throw strikes.

The decision to include bases on balls in earned runs was an implicit declaration that earned runs were a defensive statistic. This was not universally understood, much less accepted. Confusion and outrage were inevitable. Something had to give. The rules committee held a mail ballot and voted to remove bases on balls from the error column, relegating them to the summary alone. It took some time for everyone to internalize the system. As late as 1889 no less a person than Nicholas Young was confused, claiming in an interpretation entirely unsupported by the text of the rule that bases on balls were not counted in earned runs.[fn]R. M. Larner, “Washington Whispers,” Sporting Life May 15, 1889.[/fn] People eventually figured it out, and the rule has remained substantially unchanged since.


There is a temptation to regard the “traditional stats” as obvious, in contrast to modern “advanced stats.” Both sides in the advanced stats discussion tend to fall into this trap, treating “obvious” either as virtuous or as naive. In fact, the traditional statistics are anything but obvious, other than through long familiarity. They were the product of extensive discussions spanning decades, and involving close observers of the game. More recent discussions often reprise older debates. We have seen how Chadwick considered and rejected slugging average in the 1860s, and the 1887 batting average rule is immediately recognizable today as on-base average. They didn’t always reach the best conclusion—Chadwick was probably right to reject slugging average, given the conditions of the day, but in retrospect, they probably should have stuck with on-base average-style batting average. But, right or wrong, they were not merely settling for an obvious answer.

RICHARD HERSHBERGER writes on early baseball history. He has published in various SABR publications, and in “Base Ball: A Journal of the Early Game.” He is a paralegal in Maryland.