Did MLB’s Clean Ball Policy Fuel the Hitting and Scoring Surge of the 1920s?

Some baseball analysts have pointed to baseball’s adoption of its “Clean Ball Policy” following the August 1920 beaning death of Cleveland SS Ray Chapman as a major catalyst of the surge in hitting and scoring that marked the turning point demarcating the end of baseball’s “Deadball Era” and the dawn of the modern game. If this hypothesis is true, this offensive surge should show up mainly in the later innings, when scuffed and dirty balls would have previously remained in play; the new policy would have had little or no impact in the early innings, when fresh baseballs were always in play. I tested this hypothesis by comparing total runs per 9 IP, ERA, OPS, HR per 9 IP, and SO per 9 IP in 1921–29 versus 1912–20 in the AL and NL for each inning and for three-inning groupings (1–3, 4–6, 7–9). Highly significant improvements (increased scoring, OPS and HR and decreased SO) were observed in every inning in each league after 1920. While these improvements tended to be somewhat smaller in the first three innings than later in the game, the differences between inning categories were of borderline statistical significance at best. Thus, MLB’s adoption of the Clean Ball Policy seems to have played only a minor role in the offensive resurgence of the 1920s.

INTRODUCTION

The 1920 season marked a watershed in the history of major league baseball (MLB). It was the year when Kenesaw Mountain Landis was named Baseball Commissioner in the aftermath of the gambling scandal that forever tainted the 1919 World Series.1 This season was also noteworthy for the founding of the Negro National League, the first (retroactively) recognized major league for non-white players.2 The 1920 season also marked Babe Ruth’s debut as a Yankee and as a full-time outfielder; he hit an astounding 54 HR, nearly doubling his previous MLB record of 29 HR set in 1919.3

On a somber note, 1920 was also the season in which Cleveland’s beloved 29-year-old star shortstop, Ray Chapman, was hit in the head and killed by a submarine fastball thrown by Yankee Carl Mays in the top of the fifth inning on August 16.4 It was an overcast and rainy afternoon, and the ball was damp and discolored. Chapman, who habitually crowded the plate, never saw the pitch coming until it was too late. Although MLB disregarded calls for mandatory batting helmets until many years later, it took immediate steps to require prompt removal of scuffed and discolored baseballs from play—the “Clean Ball Policy.”

Most importantly—and perhaps not coincidentally—the 1920 season is widely regarded as the inflection point marking MLB’s transition from the nearly two-decade scoring slump known as the “Deadball Era” to the free-swinging modern game.5 The term “deadball” is largely a misnomer, since it attributes the scarcity of runs to the characteristics of the ball used in the early twentieth century, which in fact was the same as the ball used in the mid-1890s, when the league-wide ERA was at its all-time high.6 The decline in scoring around the turn of the century was due mostly to a combination of the influx of stronger, harder throwing pitchers, the popularization of the spitball, and improvement in fielding gloves, and was greatly accelerated by the adoption of the foul-strike rule in 1901 (NL) and 1903 (AL).7 MLB did try to stimulate scoring starting in 1910 by substituting cork for the hard rubber core to make baseballs lighter and livelier, but this change brought only a modest and transient increase in scoring in 1911–1913.8

Although the first signs of MLB’s offensive resurgence of the 1920s actually appeared as early as 1918 or 1919, the hypothesis that the “Clean Ball Policy” adopted after the August 1920 death of Ray Chapman might have fueled this resurgence has recently gained currency.9 After all, it stands to reason that balls that have been discolored, scuffed, softened, and/or misshapen by repeated use may become harder to see and to hit. Fortunately, it is quite easy to test this hypothesis by comparing inning-by-inning splits of run scoring and other offensive indices before and after 1920. Since each game started with a fresh baseball even before 1920, the implementation of the Clean Ball Policy could have affected scoring only after balls would otherwise have been in play long enough to become scuffed or dirty. Thus, we should see little or no impact on scoring in the first inning but an ever widening impact as one progresses from the second through the ninth (and extra) innings. We will now compare and contrast the actual inning-by-inning splits in runs and other related performance measures in the nine years immediately preceding (1912–20) and following (1921–29) MLB’s implementation of the Clean Ball Policy.

 

A roughed-up baseball from the Deadball Era. (Library of Congress)

 

METHODS

All data used in this article were taken from Baseball Reference.10 Inning-by-inning splits were calculated using their Stathead tool.11 These splits are only available starting in 1912 and are unavailable for the Negro Leagues. Although splits are available for the Federal League, we confined our analysis to the American (AL) and National (NL) Leagues to keep the number of teams constant from year to year. Since the Clean Ball Policy was not in effect through most of the 1920 season, we classified 1920 as a pre-Clean Ball Policy season in our grouped analyses.

The outcome parameters we examined were total runs (R), earned runs (ER), home runs (HR), and strikeouts (SO), each normalized per 9 IP per team, and on-base plus slugging percentage (OPS). There were slight discrepancies in R, HR, and SO between the Baseball Reference pitching and batting databases presumably due to missing data; we used the numbers from the pitching stats.

Analyses of variance were performed using Microsoft Excel’s Data Analysis Tools add-in. Two-way ANOVA models with replicates were used. The main factors were the Inning-by-Inning Split (10 categories—first to ninth innings and extra innings), and Period (2 categories—1912–20 and 1921–29), and their interaction (whether the change in outcome between the two time periods differs by Inning Split). Yearly totals from each of the two leagues and nine seasons provided the 18 replicates (2*9) for each Inning*Period category. This ANOVA had 359 (2*10*18-1) degrees of freedom. The alpha level was set at 0.05. We also repeated these analyses after collapsing the inning splits into three categories—1–3, 4–6, 7–9—and omitting the Extra Innings data. In these ANOVAs, each Inning Split*Period category had 54 replicates (3*2*9) and there were 323 (2*3*54-1) degrees of freedom.

RESULTS

Figures 1–4 show the temporal trends for scoring, OPS, HR, and SO for each league from 1910–30.12

 

To address the extent (if any) to which this offensive surge was fueled by MLB’s adoption of the Clean Ball Policy after the August 1920 death of Ray Chapman, inning-by-inning splits in run scoring, OPS, HR and SO were compared in the nine seasons following this tragedy (1921–29) versus the nine seasons preceding this tragedy (1912–20).13 If the Clean Ball Policy played a significant role in the offensive surge, one would expect to see no increases in offensive metrics in the first inning and ever growing increases in these metrics as one progressed through innings 2–9 and into extra innings. This is not what happened.

 

 

Offensive production clearly rose after 1920 across the board (Table 1). Teams scored roughly one more run (and one more ER) per 9 IP in 1921–29 than in 1912–20 for every inning split category. OPS also rose consistently, HR rates increased two- to three-fold, and strikeout rates dropped. Note also that scoring is consistently highest in the first inning (when a team’s best hitters come to the plate) and lowest in the second inning (when the bottom of the order usually comes to the plate) in both time periods. This effect is smoothed out when three-inning groupings were analyzed (bottom three rows). Offensive production follows no particular pattern thereafter except for a drop-off in extra innings.

The key question for our present purpose is whether the increase in offensive production after introduction of the Clean Ball Policy in late 1920 was significantly greater in the later innings (when this policy should in theory have had the most impact) than in the earlier innings (when this policy should in theory have had little or no impact).

Table 2 shows the increases in scoring, OPS, and HR rate and the decrease in SO rate between 1912–20 and 1921–29 for each inning and for three three-inning groupings (which exclude extra innings).

 

 

Clearly, there was substantial improvement in every offensive indicator – more total and earned runs, higher OPS, more home runs, and fewer strikeouts – across the board, even in the first inning, before baseballs had much chance to become dirty or scuffed. Thus, the adoption of the Clean Ball Policy was clearly not the primary cause of the offensive renaissance of the 1920s. That being said, the changes – especially in OPS – appear to be greater in innings 4–9 and in extra innings than in innings 1–3. Two-way ANOVA was performed to address the statistical significance of these more subtle differences (Table 3).

 

We did the ANOVA in two ways: 1) by individual inning, with the composite of extra innings as the 10th category, and 2) by three-inning grouping, excluding extra innings. Clearly, the average differences between the nine year periods preceding and following 1920 were overwhelmingly significant for every offensive indicator. The offensive differences by inning were also highly significant when considered inning by inning (10 categories) but not when collapsed into three-inning groupings, in which the highest and lowest scoring innings (the first and second) offset each other in the Innings 1–3 category. Most importantly for our hypothesis, the interaction between Inning Split and Time Period (which tests the whether the offensive metrics improved more in the later than the early innings) was statistically significant only for HR (P=0.047) in the inning-by-inning analysis and for HR (P=0.012) and OPS (P=0.022) in the analysis of three-inning groupings. Thus, while the Clean Ball Policy may have had a marginally stimulatory impact on offense, it was not the major driver of the transformation from the “deadball” to the “live ball” era.

 

DISCUSSION

MLB’s abrupt and profound scoring surge circa 1920 fundamentally reshaped how baseball was played, as the emphasis shifted from contact hitting and “small-ball” strategies designed to eke out one precious run at a time to the “go-big-or-go-home” paradigm of swinging for the fences and playing for the big inning. Although many players (especially disgruntled pitchers) attributed the offensive resurgence to the surreptitious introduction of a “rabbit ball,” the only known significant modification of the baseball itself—the change from hard rubber to cork in the core—had occurred 10 years earlier.14 Shifting the focus to what pitchers did to the ball once it was in the game, i.e., the application of saliva and foreign substances and the practice of deliberately nicking and scuffing balls to produce more deceptive movement, some have looked to the banning of the spitball in 1920 to explain the offensive surge. But this hypothesis cannot bear serious scrutiny, since the 17 spitball pitchers who were grandfathered under this ban continued to throw spitballs legally long after 1920 and did not stop until Burleigh Grimes retired after the 1934 season.15 An analysis I published in 2021 showed that home run rates (as a percentage of batters faced) increased at least as much for these 17 grandfathered spitballers (0.44 in 1919 to 0.67% in 1920) as for the non-spitball pitchers (0.56 to 0.68%).16

The hypothesis that the 1920 Clean Ball Policy requiring the prompt removal of scuffed and dirty balls from play drove the offensive surge of the 1920s has more face validity than the spitball hypothesis. However, the present analysis suggests that this policy was a minor contributor at best. While the post-1920 increases in scoring, OPS, and HR and decrease in SO were systematically smaller in the first three innings than the later innings, and some of these differences met (barely) the alpha =.05 criterion for nominal statistical significance (Tables 2–3), none had a P-value <0.01, which would have provided more confidence in the significance of the results, given the large number (10) of ANOVAs performed.

The offensive resurgence of the 1920s, which was marked by an unprecedented emphasis on home runs, is probably best understood as a shift away from an outmoded small-ball paradigm, which was well suited for the Baltimore Orioles of the 1890s but could not keep pace with the advances in pitching and fielding and with changes in the scoring environment over the ensuing 25 years.17 When Willie Keeler was “hitting ’em where they ain’t” for the 1897 Baltimore Orioles, the rules favored a contact first approach.18 With the foul strike rule still four years in the future, Keeler could choke up on the bat and foul off as many pitches as he pleased without getting into a two-strike count; he struck out only 5 times in 618 PA all year and hit .424.19 When he finally put a fair ball in play he enjoyed the advantage of a rock-hard infield and primitive fielding gloves that allowed 2.5 errors per team per game—much more than the 1.48 errors per team per game in 1920.20 The spitball had not yet been invented, and only a few pitchers, like Amos Rusie and Cy Young, threw seriously hard. Keeler’s contact first approach was perfect for that environment.

But that environment changed rapidly as the new century began. The adoption of the foul strike rule in 1901 (NL) and 1903 (AL) curtailed the ability of hitters to foul off pitches indefinitely and brought about an immediate 50–60% jump in strikeout rates.21 Flame throwers like Walter Johnson and Rube Waddell, and practitioners of new novelty pitches like the spitball (Ed Walsh, Jack Chesbro) and the fadeaway (Christy Mathewson) made the lives of hitters more difficult. As fielding gloves improved, errors declined and a higher percentage of balls in play became outs. While a few superstars like Honus Wagner, Nap Lajoie, Ty Cobb, Tris Speaker, Eddie Collins, and Joe Jackson continued to thrive using the old contact first approach, MLB-wide batting averages plummeted, falling as low as .239 in 1908.22 As runs became scarcer, hitters and managers doubled down on the 1890s approach, relying on strategies like the sacrifice bunt and hit-and-run and aggressive base running to eke out every possible run from a dwindling number of runners on base.

When Babe Ruth began spending time in the outfield in 1918, he brought an entirely new approach, swinging as hard as he could on every pitch and hitting home runs at an unprecedented pace. Baseball was slow to recognize the superiority of Ruth’s approach for producing runs, which flew in the face of tradition, probably because it was instinctual, not strategic. As Ruth’s HR totals climbed from 11 in 1918 to 29 in 1919 to 54 in 1920, he was more or less alone.23 However, his growing popularity and impact on attendance could not be ignored for long. When deadball star Rogers Hornsby, who had never hit more than 9 HR in any of his first six seasons (1915–20) suddenly hit 21 HR in 1921 and surpassed Ruth with 42 HR in 1922, the new paradigm took hold for good.24 A baker’s dozen other AL and NL sluggers—Ken Williams, Tillie Walker, Cy Williams, Bob Meusel, Lou Gehrig, Hack Wilson, Jim Bottomley, Chuck Klein, Mel Ott, Lefty O’Doul, Al Simmons, Jimmie Foxx, and Don Hurst—would join Ruth and Hornsby in the 30+ HR club over the course of the 1920s.25 Although many baseball traditionalists decried the change from the beauty and “science” of the deadball game to the brute power of the modern game, most fans disagreed, and the only science that mattered favored the approach that put more runs on the board. MLB had entered its modern era, and there was no going back. 

DAVID J. GORDON, MD, PHD, is a retired biomedical scientist and longtime Cubs fan, who joined SABR in 2016 and has a keen interest in baseball history and in metrics that can be applied across historical eras. Since 2016, he has authored nine BRJ papers and three books—Baseball Generations, a historical account of the best players of each of seven eras spanning 1871–2019; The American Cardiovascular Pandemic: A 100-Year History, a chronicle of how advances in cardiovascular science have reversed the twentieth century tide of heart attack mortality; and Baseball’s Shooting Stars, which tells the stories of 30 players who had a single Hall of Fame quality season that they could never replicate. More recently, he has collaborated with John Contois on a book about MLB’s greatest hitting pitchers, which should be available in bookstores in 2026.

 

Notes

1. BR Bullpen, “Black Sox Scandal,” https://www.baseball-reference.com/bullpen/Black_Sox_Scandal.

2. BR Bullpen, “Negro Leagues,” https://www.baseball-reference.com/bullpen/Negro_Leagues.

3. Baseball Reference, “Babe Ruth,” https://www.baseball-reference.com/players/r/ruthba01.shtml.

4. Don Jensen, “Ray Chapman,” SABR BioProject, https://sabr.org/bioproj/person/ray-chapman/.

5. BR Bullpen, “Deadball Era,” https://www.baseball-reference.com/bullpen/Deadball_Era.

6. David J. Gordon, “The Rise and Fall of the Deadball Era,” SABR Baseball Research Journal, Vol. 47, No. 2 (Fall 2018), 92–102.

7. BR Bullpen, “Foul Strike Rule,” https://www.baseball-reference.com/bullpen/Foul_strike_rule.

8. Zachary D Rymer, “The evolution of the baseball from the Dead-Ball Era through today,” Bleacher Report, June 18, 2013, http://bleacherreport.com/articles/1676509-the-evolution-of-the-baseball-from-the-dead-ball-era-through-today.

9. Travis Sawchick, “Baseball’s Beanball Death 103 Years Ago Changed Baseball Forever,” The Score, 2023, https://www.thescore.com/mlb/news/2693942. Spencer Rickles, “The Pitch That Killed Ray Chapman And Changed Baseball Forever,” ATL Braves Country, April 15, 2025, https://atlbravescountry.com/pitch-that-killed-ray-chapman-and-changed-baseball-forever/. BR Bullpen, “Chapman Beaning,” https://www.baseball-reference.com/bullpen/Chapman_beaning#Aftermath.

10. Baseball Reference, https://www.baseball-reference.com/.

11. Stathead Baseball, https://www.sports-reference.com/stathead/baseball/.

12. Baseball Reference, Year-by-Year Pitching and Batting Totals: AL Pitching https://www.baseball-reference.com/leagues/AL/pitch.shtml, NL Pitching https://www.baseball-reference.com/leagues/NL/pitch.shtml, AL Batting https://www.baseball-reference.com/leagues/AL/bat.shtml, NL Batting https://www.baseball-reference.com/leagues/NL/bat.shtml.

13. Stathead Baseball, Team Pitching Split Finder: 1912–1929 by Year, Inning, and League, https://www.sports-reference.com/stathead/baseball/split_finder.cgi?request=1&order_by_asc=1&order_by=year_game&year_min=1912&year_max=1929&split_1=situa%3Ainnng&class=team&type=p&sr_pitching_splits_output=view_all&combine_lg=Y.

14. Rymer.

15. BR Bullpen, “Spitball,” https://www.baseball-reference.com/bullpen/Spitball.

16. David J. Gordon, Baseball Generations (South Orange, New Jersey: Summer Game Books, 2021), Table 4.2, 41.

17. Gordon, “The Rise and Fall of the Deadball Era.”

18. Doug Skipper, “Willie Keeler,” SABR BioProject, https://sabr.org/bioproj/person/Willie-Keeler/.

19. Baseball Reference, “Willie Keeler,” https://www.baseball-reference.com/players/k/keelewi01.shtml.

20. Baseball Reference, Major League Fielding Year-By-Year Averages, https://www.baseball-reference.com/leagues/majors/field.shtml.

21. Gordon, “The Rise and Fall of the Deadball Era.”

22. Baseball Reference, Major League Batting Year-By-Year Averages, https://www.baseball-reference.com/leagues/majors/bat.shtml.

23. Baseball Reference, “Babe Ruth.”

24. Baseball Reference, “Rogers Hornsby,” https://www.baseball-reference.com/players/h/hornsro01.shtml.

25. Baseball Reference, Year-By-Year Top-Tens Leaders & Records for Home Runs, https://www.baseball-reference.com/leaders/HR_top_ten.shtml.