# I Don’t Care If I Ever Get Back: Marathons Lasting 20 or More Innings

Baseball is thankfully free of artiﬁcial boundaries of time which conﬁne other sports. This freedom helps to shape the unique magical charm that is an evening at the ballpark. Fans never know whether it will be a two-hour squeaker or whether they may be enchanted until past sunrise by the ﬁrst-ever wild 12-hour 46-inning slugfest.

In the bottom of the seventh, baseball fans worldwide stand up to sing Albert von Tilzer’s music and Jack Norworth’s lyrics for the 1908 baseball anthem, “Take Me Out to the Ball Game.” When games go into extra innings, the song is sung again, with much more meaning, in the bottom of the 14th, bottom of the 21st, bottom of the 28th, etc. Each time it ends with “I don’t care if I ever get back.”

For 41 years I have been researching baseball games lasting 20 or more innings, ﬁnishing after 1:00 a.m. local time, and taking more than six hours, ever since my father and I attended a 26-inning twinight doubleheader at Forbes Field August 9, 1963. After a long rain delay, the opener went 15 innings, and Roberto Clemente’s RBI ended the nightcap in the 11th at 2:30 a.m. The next day, we discovered no one at KDKA Radio or any Pittsburgh newspaper could answer the question, “Is that the longest-ever night of baseball?”

For purposes of this article, a marathon is deﬁned as a game lasting 20 or more innings. In my research I have discovered 341 marathons. These games are hard to ﬁnd. Leagues either keep no records, or keep track only of their longest game; only the Texas League keeps records on all marathons.

Nobody has ever before explored such questions as: What is the probability a game will go x number of innings? How often should we expect a marathon of 20 or more innings, or 40 or more innings? What is the probability the 26-inning major league record will be broken this year? Which is “rarer,” the 26-inning major league record, 33-inning minor league record, or 45-inning “other” category record? In this article we’ll answer these questions.

##### JOE DiMAGGIO’S HITTING STREAK

There are important parallels between this research on the probability that marathons will occur, and research concerning the probability a batter may ever break Joe DiMaggio’s 56-game hitting streak, the prime example of a baseball event totally defying the laws of probability. The late Stephen Jay Gould once stated that DiMaggio’s streak was the “most extraordinary thing that has ever happened in sports.” I hope these parallels can be further explored in the future.

##### LONGEST MARATHONS NEVER PLAYED

The record for longest game has been increasing ever since the ﬁrst baseball game was played back in the late 1700’s or early 1800s. But the record has been cluttered with games that were never played! At the Delaware County Fairgrounds in Manchester, Iowa, in September 1925, or so the story goes, the Cascade Reds and Ryan Shamrocks started a game they never dreamed would last two years! The supposed “54-inning” game continued over six days through 16 innings, another 15 innings, a rainout, a snowout, another 14 innings, and ﬁnally another nine innings. Cascade ﬁnally won 9-8 in September 1926. This was thought to be an incredible all-time baseball record which would never be broken.

However, as I tracked down microﬁlm records, I discovered that, as is so often the case, not all is as it seems! Thanks to SABR member R. J. Lesch, the mystery was unraveled as we obtained accounts from the Cascade Pioneer, Dubuque Telegraph-Herald, and Cedar Rapids Gazette. Box scores proved the supposed 54- inning game was actually four separate games! So how were generations of baseball historians misled? Local newspapers clearly indicated four separate games were played. But when Maury White’s column in the May 21, 1984 Des Moines Register mistakenly referred to a “54-inning game,” the seed was planted for the “great myth.”

I uncovered another supposed six-day-long contest which never happened. From 1976 through 2004, a 28-inning game between Highland and Grifﬁth High Schools in Indiana was listed by the National Federation of High School Associations as the longest high school game ever played. But there was never any ﬁnal score given, and it was always listed simply as “1976,” with no month and no date.

Despite many inquiries, nobody could tell me the date of the game or the ﬁnal score. David Zandstra of the Highland Historical Society ﬁnally solved the puzzle by ﬁnding a scrapbook containing articles from the Lake Suburban Sun Journal and Gary Post-Tribune. The ﬁrst three days were rainouts, 10-inning and 11-inning tie games were played on the fourth and ﬁfth days, and Grifﬁth took a 5-2 victory in seven innings on the sixth day.

I found three other “marathons” which were never played: a 28-inning 1948 amateur game in South Dakota, a 22-inning 1932 Mississippi Valley League game in Iowa, and a 20-inning 1904 Cotton States League game in Mississippi.

##### LONGEST MASSACHUSETTS-RULES MARATHON

During the 19th century many games were played under Massachusetts Rules, requiring the winning team to score a minimum number of runs or “tallies,” sometimes 25, sometimes 50, sometimes 100. Such games established records for innings played which have never been equaled. But these records must be considered differently than other records, since one inning was not three outs, but rather one.

In 1860, at the Agricultural Fairgrounds in Worcester, the Medway Unions and Upton Excelsiors played the longest game ever in the history of baseball: 172 innings over seven days! Final score: Upton 50, Medway 29, after 21:50 game time.

##### LONGEST FICTIONAL MARATHON

The Iowa Baseball Confederacy by W. P. Kinsella is the greatest baseball novel ever written. The Chicago Cubs came to Big Inning, Iowa, July 4, 1908, to play an all-star team from the minor league Iowa Baseball Confederacy. After the game was adjourned for the evening, still tied, 5,000 fans packed the park the next day, having no idea the game would eventually continue through driving rainstorms for 40 days! At dawn on August 12, a Confederacy pinch-hit homer in the bottom of the 2,614th ended the game. Final score: Confederacy 12, Cubs 11.

##### LONGEST SPOOF MARATHONS

In 1884, the Denver Opinion printed a spoof about a 39-inning 1873 game between two Portland, Oregon, newspapers, the Oregonian and Bulletin. The game began at 12:30 p.m., and when darkness arrived, lanterns and locomotive headlights were found to light the ﬁeld. At 1:00 a.m., in the 39th, a long ﬂy to right was booted for a four-base error. Final score: Bulletin 1, Oregonian 0. Asked about his error, the right ﬁelder stated moonlight had been shining in his face and he couldn’t see the ball.

In 1907, the spoof innings record was broken in Munchausen, Pennsylvania, as the Lyerhelms and Fakenhursts played to a 50-inning scoreless tie. The “Liars” and “Fakers” played in a town named after Baron von Munchausen (1720-97), the famous German storyteller whose tall tales were so outrageous that the medical condition for compulsive lying, Munchausen’s Syndrome, is named after him.

The next year, the record was broken again as the Washington Post told of a game at Jones County Eye, Ear, and Tongue Inﬁrmary. At dusk, a farmer drove his horse-drawn wagon to a nearby university’s chemistry lab to obtain phosphorus, which was smeared on the baseball to allow play to continue. At dawn, the Lightfoot Lilies took a 1-0 lead over the Ringtail Roarers in the top of the 57th on a mammoth homer by Bull Thompson. In the bottom of the inning, Bruiser Brown was at bat with a full count, two runners on, and only one out. Then, just as the pitcher picked the runner off second, Bruiser swung at a ﬁreﬂy ﬂitting near the plate for strike three and a double play, ending the game.

The last spoof occurred in October 2003, authored by ESPN. com. The Cubs, waiting since 1908, and Red Sox, then waiting since 1918, met in the 2003 World Series to determine “whose curse is worse?” The Red Sox led 1-0 in Game Seven as the Cubs batted in the ninth at Fenway. With Fate desperately seeking to allow neither team to win, and the Cubs down to their last strike, Sammy Sosa hit a home run off the CITGO sign, tying the game. As dawn approached, a meteor struck the earth, ﬂoods covered the globe, and a dust cloud encircled the planet. Chaos reigned, and Game Seven was suspended in the 28th inning. Final score: Cubs 34 Red Sox 34, with both teams still waiting to win a Series. Little then did the Red Sox know that they would wait only one more year!

##### FIRST-EVER GAMES OF X INNINGS

The ﬁrst-ever game to last more than nine innings may have been the 10-inning game on November 18, 1845, at Elysian Fields. William Wheaton’s Team beat William Tucker’s Team 51-42. The 21 Rule called for the team scoring 21 runs ﬁrst to be declared the winner, so these teams must have been tied at the end of several innings, or must have both exploded offensively in the 10th inning.

The record for longest game stood at 10 innings for seven years, until June 24 or 27, 1852, when the Gothams took 16 innings to beat the Knickerbockers 21-16 at Red House Ground in New York City. The ﬁrst-ever extra-inning game under the new nine-inning rules may have been in Brooklyn on June 6, 1865. The Gothams scored in the ninth to tie Enterprise at 17. In the 13th, Enterprise scored one, but the Gothams scored two to win, 19-18. The ﬁrst-ever games taking 1-45 innings, and last-ever games taking 20-45 innings are given in an appendix on the website.

##### BALLPARKS AND CITIES IN MARATHONS: THE BEE HIVE

The park which hosted the most marathons is Braves Field in Boston, now known as BU’s Nickerson Field—appropriate since it hosted the longest major league game ever played, 26 innings. Braves Field has seen ﬁve marathons, two between April 17, 1936, and April 23, 1941, when the Braves were ofﬁcially named the Bees and the park was known as the Bee Hive.

Old Comiskey hosted four marathons: an American Giants game in the Negro American League and three White Sox AL games, including the longest AL game ever played, 25 innings. Wrigley has hosted three. Other current major league parks that have hosted marathons include RFK and Shea with two apiece, and Yankee Stadium, Fenway, Angels Stadium, the Metrodome, McAfee Coliseum, and Pro Player Stadium with one apiece. The other 20 have never hosted a marathon.

##### LONGEST MAJOR LEAGUE MARATHON

At old Comiskey on May 8-9, 1984, the White Sox downed the Brewers 7-6 in 25 innings. Suspended after 17 innings at 1:05 a.m. by the AL curfew, the game was won the next evening in the 25th by Harold Baines’ homer, which just barely cleared the bullpen fence in center. The Sox scored two in the ninth, and three in the 21st to keep the game tied, and would have won in the 23rd except Dave Stegman was ruled out for coach’s interference when third base coach Jim Leyland helped him to his feet after Stegman tripped rounding third. This game is rich in “might-have-been’s”: Had it been an NL game with no curfew, it would have ended at 3:42 a.m. Had it been played between 1910-48 or 1976-80, when old Comiskey had no inner fence in center, Baines’ drive would have been caught, and they might have broken the major league record of 26 innings. Had the game been the nightcap of the foggy September 24, 1971, Astros at Padres twinight doubleheader, which began at 12:01 a.m., it would have ﬁnished at 8:07 a.m., and the last few innings could have been covered live by The Today Show.

On September 11, 1974, Ken Reitz’s homer for the Cards with two outs in the ninth tied up the Mets at Shea. Only a thousand fans remained to see Bake McBride score all the way from ﬁrst in the 25th when Mets pitcher Hank Webb’s pickoff throw to ﬁrst was wild, and the relay to the plate was dropped by catcher Ron Hodges to give the Cards a 4-3 win at 3:13 a.m. As home plate umpire Ed Sudol ruled McBride safe at the plate, he couldn’t help but remember he had also been behind the plate during two other Mets marathon losses: 23 innings in 1964, and 24 innings in 1968. Amazingly, the ﬁrst base umpire had called a balk on the pickoff. Under a rule just revoked, McBride would have had to return to second. He might never have scored, and the game might never have ended!

Because an April 22, 1871, game between Washington’s Olympics and Nationals was later thrown out (along with the Nationals), the honor of setting the ﬁrst major league innings record thus went “after-the-fact” to the Fort Wayne Kekiongas and Forest City’s of Cleveland, who met May 4, 1871, in Fort Wayne, Indiana. Although this is considered the ﬁrst major league game ever played, if one believes history cannot and should not be altered, then it is really the second.

There is no dispute whatsoever as to the longest major league game ever played. On May 1, 1920, the Boston Braves hosted the Brooklyn Robins at Braves Field. The game took 26 innings and lasted 3 hours 50 minutes, but ended as a 1-1 tie when called at 6:50 p.m. due to darkness by Umpire Barry McCormick. Two thousand fans saw starting pitchers Joe Oeschger of the Braves and Leon Cadore of the Robins go the whole way. This could never happen now, with the emphasis on relief pitchers, but starters commonly pitched entire marathons in the early 20th century.

The Robins’ next two games were against the Phils and then vs. the Braves again. They took 13 and 19 innings respectively to lose both. So in just three games, they played 58 innings, losing two and tying one, a three-game record for futility which will probably never be equaled. You would think the Braves and Robins must hold the record for combined innings for two separate games when the ﬁrst was tied and had to be replayed in its entirety: by adding nine innings played later to ﬁnally have the Braves win once and for all what began May 1, to the 26 innings played May 1, you have 35 innings. But you would be wrong! The A’s and Tigers hold this record: 40 innings. On July 21, 1945, they played a 1-1 tie in 24 innings at Shibe Park. When they met two months later to ﬁnally come to a decision, the A’s took 16 innings to win.

##### MOST INTERESTING EVENTS DURING MY RESEARCH

The research process involved in discovering marathons has taken me to the Hall of Fame Libraries in Cooperstown and Tokyo, and just about everywhere in between, including hundreds of ballparks, SABR meetings, and libraries. The most unusual discovery was the 21-inning 1939 marathon in Wisconsin in which the hometown Clintonville Four-Wheel-Drive (FWD) Truckers defeated Two Rivers Polar Bears 1-0. This game was discovered in the June 29, 1939, Daily Independent of Helena, Montana!

I have been very fortunate to experience many fun times while conducting research at every major league park, roughly half the existing minor league ﬁelds, and many overseas diamonds in Latin America, Europe, Africa, Asia, and Australia. I especially treasure my “Croix de Candlestick,” watching the Yakult Swallows play through a monsoon in beautiful ancient Meiji Jingu Stadium in Tokyo without the slightest thought of a rain delay; taking the subway to Yankee Stadium after hearing on the radio at midnight that a rain-delayed game was only in the ﬁfth; and cheering when the St. Paul Saints, trailing in the bottom of the ninth, down to their last out, scored seven consecutive runs, climaxed by a walk-off grand slam, to win the Northern League championship over the Schaumberg Flyers September 19, 2004, the only season-ending walk-off grand slam ever hit in the history of baseball!

##### LONGEST MINOR LEAGUE MARATHON

At 4:07 a.m. on Easter morning, April 19, 1981, just 51 minutes before sunrise, 17 freezing and very fortunate souls huddled in the 28-degree pre-dawn chill of Pawtucket, Rhode Island’s McCoy Stadium, having just watched their beloved Paw Sox fail to break a 2-2 tie with the Rochester Red Wings in the bottom of the 32nd. When the umpires suspended the game, these brave 17 fans could look back on 8 hours, 7 minutes of baseball, preceded by a half-hour power failure delay.

The game resumed June 23, and the mercury had risen to 80 degrees. McCoy was packed to capacity, and because the major leagues were on strike, the eyes of the entire baseball world were focused on Pawtucket. The Paw Sox won 3-2 in the 33rd. Final totals of 8 hours 55 minutes elapsed time and 8 hours 25 minutes game time are modern baseball records, and 33 innings is still the all-time professional record. Momentoes of this historic game are now buried in a time capsule beneath the ﬁeld, where they join the ﬁve-ton truck that in 1942 sank without a trace into the swampy outﬁeld while McCoy was being built by the WPA.

The longest doubleheader ever played was a North Carolina State League twinbill on July 5, 1915. The Raleigh Capitals downed the Durham Bulls 3-2 in 14 innings in the a.m. game at Raleigh. Then they bussed over to Durham for the p.m. game, which was called a 2-2 tie after 21 innings, making a total of 35 innings for the day. A list of all doubleheader marathons going 28 or more innings can be found on the web site.

I have come across 14 occasions involving three or more games in one day. Amazingly, 12 were sweeps. The odds on that must be extremely low! Nine of 11 tripleheaders were sweeps. In Brooklyn, September 1, 1890, the Dodgers swept three from Pittsburgh, 10-9, 3-2, and 8-4, a total of 27 innings. In Baltimore, September 7, 1896, the Orioles swept three from Louisville, 4- 3, 9-1, and 12-1 in eight innings, a total of 26 innings. And in Pittsburgh, October 2, 1920, the Pirates lost two out of three to Cincinnati, losing the ﬁrst two, 13-4 and 7-3, and winning the third 6-0 in six innings, a total of 24 innings. There have been eight minor league tripleheaders; seven were sweeps.

There has never been a major league quadrupleheader or longer. Both minor league quadrupleheaders were sweeps. There has been just one minor league sextupleheader, again a sweep! In Manchester, NH, September 4, 1899, in the New England League, the host Manchesters swept six from the Portland Phenoms by 14-7, 12-8, 12-2, 8-4, 9-1, and a 9-0 forfeit. Portland walked off the ﬁeld and forfeited after two innings in the sixth game to protest the ump’s decision to eject one of their players, but the ﬁrst ﬁve games lasted nine innings each, a total of 47 innings for the sextupleheader.

LONGEST MARATHONS WHICH BROKE NO RECORDS

Two 27-inning games, although tied for third longest minor league game ever, have received no attention because they were played after the Pawtucket 33-inning marathon in 1981 and therefore broke no records. The first was a three-day-long thriller at MacArthur Stadium in Syracuse. On June 19, 1985, the Pawtucket Paw Sox and Syracuse Chiefs played 22 innings before getting suspended. After a 13-minute rain delay in the 23rd, and another of 50 minutes in the 24th, the game was suspended due to rain after 232 innings. The Paw Sox finally won 3-1 on the third night. Total game time: 7:07. What makes this game even more amazing is it was the second time in just over a week the Chiefs had lost a three-day marathon! That Chiefs-Clippers game in Columbus was suspended after 20 innings, rained out the next night, and the Chiefs finally lost 8-7 in the 21st on the third night.

Three years later, on June 24, 1988, in Burlington, North Carolina, the Bluefield Orioles (Baby Birds) came to town for an Appalachian League game against the hometown Indians. When the Orioles finally won 3-2, the game had taken 8 hours 16 minutes, it was 3:27 a.m., and the crowd of 2,204 fans had dwindled to just either 84 or 50. Why the uncertainty on the remaining crowd at the end? Apparently, the sportswriters were sleepy because they had remarkably different accounts. Craig Holt of the Burlington Times-News counted 84 fans at the end, but wrote incorrectly that the finish had been at 3:37 a.m. rather than at 3:27 a.m. Dale Mullins of the Bluefield Daily Telegraph counted 50 fans at the end, and got the finish time correct as 3:27 a.m., but wrote incorrectly that the game had taken place in Raleigh rather than in Burlington.

I agree with Baby Birds manager Glenn Gulliver, who told reporters afterward, “I do not favor suspending games, no matter how long they take. That would wreck everything. You want to go until somebody wins.” Amen! Suspension will always be a bad idea that wrecks everything. Let the ballplayers play on to a conclusion!

##### LONGEST MARATHON

On May 24, 1942, just 11 days before the Battle of Midway (June 4-7) began, Taiyo and Nagoya of the Japanese Federation fought to a 28-inning 4-4 tie, setting a new organized baseball record. Seventeen years later, on May 2, 1959, Nippon Pharmaceuticals defeated Kurashiki Rayon 2-1 in 29 innings. So Japan had some experience with marathons. But when play began on September 20, 1983, in the title game of the 38th annual Emperor’s Cup Nan-shiki Tournament at Ibaraki-Mito Prefectural Stadium in Mito, Japan, nobody had the least idea what lay ahead!

The game between Light Manufacturing and Tanaka Hospital began at 8:50 a.m. The local Mito Band was to play after the game, and was asked to be ready at 11:00 a.m. As noon came and went, the teams were locked in a scoreless tie. After the 25th, plate umpire Choshu told the teams to take a 30-minute break. The players refused. Choshu joined his fellow umps for a short six-minute break, and then the game moved on. In the 35th, both teams pushed across one run, so the marathon continued. Finally, the game concluded at 5:15 p.m. after Light Manufacturing scored in the top of the 45th to win 2-1. Final totals: 1,029 pitches and 8 hours 19 minutes of baseball! Including the six-minute delay in the 26th, the game lasted 8:25. Excluding Massachusetts Rules games and games planned as marathons to raise funds for charity, this is the longest game by innings in the history of baseball. The Mito Band ﬁnally got to play after waiting around for over six hours.

##### HOW MANY GAMES GO INTO EXTRA INNINGS?

How many games go into extra innings? My detailed research indicates less than 6% went into extra innings in the 1800s, when late afternoon starting times resulted in many games being called due to darkness before ever having the opportunity to go into extra innings. Between 9% and 10% have gone into extra innings in the 1900s and 2000s. The average since 2000 is down to 8.2%. The number of extra-inning games has been declining since the 1960s when it was 9.9%; 1970s 9.7%, 1980s 9.7%, 1990s 9.0%, 2000s 8.2%. The recent decline can be attributed to higher scoring games.

Based on all this research, the best empirical data for percentage of extra-inning games is: 5.68% in the 1800s, 9.16% in the 1900s through 1948 day-ball era when all or most games were played in daytime, and 9.41% during the night-ball era from 1949 through now. Our theoretical model estimates 10.01% of all games will go into extra innings.

##### VISITING TEAM ADVANTAGE IN MARATHON-LAND?

Of all the 341 marathons found so far, 46 have been in the major leagues, 173 in the minor leagues, and 122 in the “other” category (42 school games, 38 amateur games, 38 international games, and 4 women’s games). One of the most surprising and unexplained facts about marathons is that a statistically signiﬁcant majority (57%) have been won by the visiting team. Is there a hidden “visiting team advantage” lurking somewhere in the Kingdom of Marathons?

##### LONGEST MARATHON RAIN DELAYS

If one loves long marathon baseball games, what better to accompany them than a long rain delay! Old Comiskey holds the record for the longest rain delay ever: 7 hours 23 minutes on August 12, 1990. The game never even got started. Finally called off at 8:58 p.m., the game was played ﬁve days later in Texas. The Rangers won 1-0 in 13 innings, as Nolan Ryan had 15 strikeouts and gave up only three hits pitching the ﬁrst 10 innings for Texas. The minor league rain delay record is held by Yogi Berra Stadium in Little Falls, NJ. The start of the August 14, 2000, Northern League game was delayed 7 hours 6 minutes, as the Catskill Cougars beat the NJ Jackals 6-1, ending at 11:06 p.m.

The longest rain delay in a minor league marathon was 1:46 in the 11th inning on September 7, 1990, at Greer Stadium in Nashville. The Omaha Royals defeated the Sounds 8-7 in 20 innings at 3:50 a.m. Old Comiskey saw the longest rain delay in a major league marathon, when the 14th was delayed 17 minutes in a 21-inning game May 26, 1973.

The longest rain delay during one at-bat is 1:52 on July 22, 1994, at the Vet. In the bottom of the fourth, Mickey Morandini took one pitch from Scott Sanders at 9:12 p.m. before a 38-minute rain delay. He took a second pitch from Sanders at 9:50 p.m., before a second rain delay of 1:14 at 9:51 p.m. Then he completed his 1:53 at-bat by doubling to left off Jeff Tabaka at 11:05 p.m. The Padres finally won 7-4 at 1:56 a.m.

The longest rain delay during one inning is 5:00 on June 9, 1980, also at the Vet. Steve Carlton, pitching a no-hitter at the time, had to wait 5:15 between pitches. His last pitch of the top of the fourth was delivered at 8:19 p.m. His first pitch of the top of the fifth was delivered at 1:34 a.m. After two long rain delays in the bottom of the fourth, the first for 1:28 and the second for 3:32, the Giants finally won 3-1 at 3:11 a.m.

The longest rain delay during one game is 5:54 on July 2, 1993, again at the Vet. There were three rain delays: 1:10 at the start, 1:56 in the bottom of the fourth, and 2:48 in the top of the sixth. The Padres won 5-2 at 1:03 a.m., but the Phils came back to win the nightcap 6-5 in 10 innings at 4:40 a.m.

The longest rain delay during one game in the AL is 5:04 on September 19, 2000 at Camden Yards in the day portion of a day-night doubleheader. There were two rain delays: 2:43 at the start, and 2:21 in the top of the eighth. The A’s finally won 7-4 at 10:36 p.m. The PA announcer told the crowd that the night game would begin promptly in 20 minutes. Six minutes later, however, he announced the night game had been postponed.

##### EMPIRICAL RESULTS FOR MAJOR LEAGUE MARATHONS

Since major league baseball began in 1871, there have been 46 major league marathons out of a total of 197,446 games played: 26 NL, 18 AL, one Negro NL, and one Negro AL. The empirical probability therefore that a major league game will take 20 or more innings is 46 / 197,446 = 1 / 4,294, or 0.02329%. This indicates a marathon should take place every 4,294 games.

In 2004, with each team playing 81 home games, there were 2,430 regular season games, plus 34 (it could have been anywhere from 24 to 41) post-season games, so there were 2,464 games. Assuming 2,464 major league games per season, a major league marathon taking 20 or more innings can be expected to come along roughly every 4,294 / 2,464 = 1.74 seasons. So we should expect a marathon a little more often than once every other season. What have we actually experienced recently? Over the past 10 years we should expect to have seen 6 major league marathons, but we actually have had only one, with an 8-year drought 1995-2002. The Cards defeated the Marlins 7-6 in 20 innings in Miami, April 27, 2003.

##### EXTRA-INNING AND MARATHON SCORING RECORDS

The highest-scoring major league extra-inning game is the 18- inning A’s 18-17 win over the Indians in Cleveland on July 10, 1932. Jack Burnett got nine hits, and Eddie Rommel relieved in the second for the A’s and went the rest of the way for the win, still the longest-ever relief effort in major league history.

Most runs scored by both teams in extra innings of a major league game is 13. On June 15, 1929, at Forbes Field, the Giants and Pirates were tied at 11 after nine, both scored one in the 11th, the Giants scored 8 in the 14th while the Pirates scored only 3: New York 20, Pittsburgh 15. On July 4, 1985, at Atlanta Stadium, the Mets and Braves were tied at 8 after 9, both scored two in the 13th, both scored one in the 18th, the Mets scored 5 in the 19th, while the Braves scored only 2: New York 16, Atlanta 13 at 3:55 a.m. The post-game Fourth of July fireworks were faithfully carried live back to New York viewers from 4:01 a.m. to 4:12 a.m. by the Mets WOR-TV broadcast team, which included Tim McCarver, who was no doubt recalling his lack of lightning speed at 3:15 a.m. in Philadelphia a decade ago September 25, 1975, which allowed Rusty Staub of the Mets to throw him out at the plate, thus ending another game that had been threatening to go until dawn!

The Rangers scored the most runs ever in one extra inning in a major league game on July 3, 1983, in Oakland, when they defeated the A’s 16-4 with 12 runs in the 15th. Note how had this game been played in Texas, the Rangers could not have scored more than four runs in the 15th.

The record for highest-scoring marathon was set August 1 and 8, 1932, when the K of P’s A Team and K of P’s B Team battled through two weekends and 22 innings in Elyria, Ohio, and we still don’t know yet who won! On August 1, the two teams were tied 19-19 when the game was called on account of darkness. The game continued August 8 and was won by one of the teams in the 22nd, but we don’t know the final score, or how many innings had been played when the game was suspended, or which team won.

We could also count games played under Massachusetts Rules. Then our highest-scoring marathon would be Upton’s 100-56 victory over Medway in Worcester, October 11-12, 1859. And if we count planned marathons, the record is the African Gray Birds’ 127-110 win over the Red-Eyed Nites in the Women’s Marathon 24 Hours for Africa in Tucson, October 18-19, 2003.

##### DATABASES ON RUN PRODUCTION PER INNING

To predict how many games will go into extra innings, and how many extra innings they will last, one must first determine how many runs are scored in each inning. Dr. Darren Glass, professor of mathematics at Columbia University, and I used two different databases to create a theoretical model for extra innings. The first is my complete database for runs scored per inning for all games that have ever gone 20 or more innings and have a box score, with 5,006 innings in 194 games.

The second database consists of Dave Smith’s Retrosheet data for innings 1-19 and my database for innings 20-45. Table 2 of Smith’s 2004 SABR convention presentation entitled “Coming from Behind: Patterns of Scoring and Relation to Winning” took data for 73 seasons (1901, 1904, 1909-10, 1912-13, 1918, 1936, 1938-42, and 1944-2003), with 2,259,116 innings in 122,906 games. Combined, these two database constitute the very best available data to predict how many runs teams will score in any given inning of any given game. The first database applies only to those games going 20 or more innings, while the second database applies to all games. These two databases are given in Appendix 2. Six very interesting facts arise out of these two databases.

First, run production is significantly higher in the first inning than in any other inning. Whereas teams score 0.487 runs per inning, the visiting team scores 0.514 and the home team scores 0.607 runs in the first inning. Higher scoring in the first inning is expected because this is the only inning in which teams are assured their best batters, at the top of the order, will all bat together.

Second, the often cited “home field advantage” is true for innings 1-8, with the home team scoring an average of 0.048 runs more per inning than the visiting team. As cited above, this home field advantage is by far most prominent in the first inning, when the home team scores on average 0.093 runs more than the visiting team.

Third, what appears to be a “visiting team advantage” exists from the ninth inning on, with the home team scoring an aver- age of 0.051 runs less than the visiting team in innings 9-45. After some reflection, this is reasonable because after the home team scores enough runs to win the game from the bottom of the ninth inning on, the game is over and so the home team stops batting. This is not really a case of a “visiting team advantage.” A very interesting variation shows this difference is much less in the ninth inning (0.035 runs) than in innings 10-45 (0.124 runs). I have no explanation for this.

Fourth, run production is much lower in extra innings than in innings 1-9. This makes sense because games are usually low scoring. High-scoring games rarely go into extra innings, but there are many 1-0 and 2-1 extra-inning games. It is statistically much less likely for the Philadelphia A’s 49-33 victory over the Troy Haymakers on June 28, 1871 (highest-scoring major league game ever) or the Cubs 26-23 win over the Phils on August 25, 1922 (highest-scoring NL game ever), to remain tied and go into extra innings.

Fifth, my data limited to only games going 20 or more innings shows dramatically lower run production in innings 10-19 than does Smith’s data based on all games. For example, Smith’s data shows runs per inning of 0.392 in the 15th inning, vs. my data showing 0.031 in the 15th. Smith’s data for innings 10 and 12-17 is more than ten times higher than my data for these same innings. Smith’s data averages out to 0.387 runs per inning for innings 10-19. My data, on the other hand, averages out to only 0.027 runs per inning for innings 10-19. We can conclude marathons lasting 20 or more innings produce 93% fewer runs during innings 10-19 than do extra-inning games in general.

Sixth, other than higher scoring in the first inning, all scoring variations, both between visiting and home teams, and also between different innings, are so minor they may be ignored statistically. An average game involves each team scoring 0.487 runs per inning. Assuming the visiting and home teams each win roughly half the time, there will be an average of 82 innings (17 half-innings) per game in a game that does not go into extra innings, or a total of 0.487 x 17 = 8.28 runs per game. Variations such as the home team scoring 0.048 more runs per inning in innings 1-8, while interesting, are statistically insignificant. When evaluating the possibility of whether an average game, involving between 8 and 9 runs, will go into extra innings, this 0.048 runs per inning is only 0.048 / 8.28 = 0.0058, a statistically irrelevant half of one percent of the total runs in the game.

##### MATHEMATICAL MODEL OF EXTRA INNINGS

Based upon this data, Dr. Glass and I constructed our mathematical theory of extra innings. To establish our model, we made some deﬁnitions. Let n = the number of innings in an extra-inning game. Let P (n) = the probability an extra-inning game will last n innings. Let T = the probability a game is tied after nine innings = the probability a game will go into extra innings. We previously determined this empirically to be 5.68% in the 1800s, 9.16% in 1900-48, and 9.41% in 1949-2003. Our theoretical model predicts this should be 10.01%. For predictions about the future, we use 10.01%, or 0.1001, for T.

Let k = the probability both teams will score the same number of runs in one inning. This is where we use the extensive databases mentioned earlier. My database results in the value of k being 0.5841. We checked additional databases (see the Bibliography on the web site) from which it is possible to obtain an empirical value for k, including Lindsey’s 0.5696 for all innings in 1958, Lindsey’s 0.5895 for just extra innings in 1958, Lindsey’s 0.5552 for all innings in 1959, Lindsey’s 0.5479 for just extra innings in 1959, and Woolner’s 0.5606 for 1980-98.

Having made the required deﬁnitions, now let’s separate an extra-inning baseball game into three separate events. The ﬁrst part is the beginning, or ﬁrst nine innings. The second part is the middle, from the tenth inning through the next to the last inning, the (n – 1)th inning. The third part is the last inning, the nth inning. Since these three events are statistically independent, P (n) = the probability of a game going n innings will be the probability of the ﬁrst part occurring times the probability of the second part occurring times the probability of the third part occurring

The probability of the ﬁrst part of a future game occurring, as we have already discussed, is empirically T = 0.0941, and theoretically T = 0.103. There are most likely several factors contributing to the discrepancy between our predicted theoretical value and the actual empirical data. In our estimation, the biggest one is that to build our model we assumed both teams are average, whereas in the real world on team may be above average and the other team may be below average, which would decrease the probability of a game going into extra innings. For details, see our article in the upcoming issue of By the Numbers published by the SABR Statistical Analysis Committee. For predictions regarding future games, we use T = 0.103.

The probability of the second part of the game occurring is k times itself for as many times as there are innings in the middle part of the game, from the tenth inning through the (n – 1)th inning. The number of innings in this middle part of the game is (n – 1) – 9 = (n – 10). We have to multiply k times itself (n – 10) times, so the probability of the second part of the game occurring is k (n – 10).

The probability of the third part of the game occurring is trickier than the previous two probabilities. If k = the probability both teams score the same number of runs in an inning, then (1k) has to be the probability both teams do not score the same number of runs in an inning, because the combined probability that they do and they do not score the same number of runs in an inning must add up to one. Now all we need to do to get our theoretical model is to string together the three different probabilities of our three independent events, and multiply them together:

P (n) = T k (n – 10) (1 – k)

Getting the correct value of k is crucial to success of our model. Data gathered by Kevin Woolner, as well as common sense, suggests offensive powerhouse teams have a different scoring distribution than teams scoring few runs. One of the key features of our model is it takes as an input the average number of runs per inning that each team scores. The model agrees with Woolner’s data showing the probability two teams score the same number of runs per inning goes down as offensive production increases, since one team is more likely to have a “breakout” inning and score a high number of runs. This also indicates the number of extra-inning games and their lengths should increase during eras when scoring is low, such as the Deadball Era. We assume both teams score 0.487 runs per inning, based on Smith’s data, the best available empirical data. This choice yields a value of k of 0.5601.

Our theoretical model can be used to predict several things. First, it predicts the theoretical probability a game will take x number of innings to play. Second, it predicts the theoretical probability a game of x number of innings will take place in the next y number of years. To accomplish this, we must make some decisions. Do we count ties, fake ties, thrown-out games, forfeits, playoffs, World Series? I have done so. How many major league games have ever been played? I have calculated this number through 2004 as 197,446. This includes all games in the Negro Leagues, which I count as major leagues. How many minor league games have ever been played? Using the Sumner and Johnson/Wolff books as my guideline (see the Bibliography on the web site), I have calculated this number through 2004 as 1,405,188. How many games in the “other” category have been played, including school games, amateur games, international games, and women’s games? I have estimated this number as ten times the number of minor league games, or 14,051,880.

Thus, my estimate of the total number of baseball games ever played through 2004, at all levels and at all locations worldwide, is 197,446 + 1,405,188 + 14,051,880 = 15,654,514. It should be noted many “other” category games are scheduled for only six or seven innings, and some minor league games are scheduled for only seven innings.

I believe 100% of all major league marathons taking 20 or more innings have been discovered. The only possible exception to this is that there could be one or more Negro League marathons yet to be discovered. For the purposes of this research, however, I assume all major league marathons have been found. This is deﬁnitely not the case for games in the minor leagues and the “other” category, due to incomplete records and lack of media coverage. Therefore, statistical probabilities for marathons developed in this article for all games at all levels are based on major league data.

Certain aspects of baseball strategy affecting the length of an extra-inning game are not included in our theoretical model. These aspects include such managerial strategies as going for a tie at home and a win on the road, frequency of using relief pitchers, cold vs. warm weather which can decrease or increase runs scored per inning, temperatures getting colder as a night game progresses, and eras such as the Deadball Era when offensive production has been signiﬁcantly different.

Using the model, we calculated the chances of major league games going x innings. The ﬁt between theoretical and actual data seems very good. For example, our theoretical model predicts the probability of a major league game going exactly 10 innings should be 4.46%; actual data indicates it is 4.12%. Our model predicted reality to within 0.34%. Similarly, our model predicts the probability of a major league game going 16 innings to be 0.129%; actual data shows it is 0.133%. Our model predicts the probability of a major league game going 22 innings to be 0.00373%; actual data shows it is 0.00405%. The ﬁt between our model and the real world is surprisingly good! Theoretical probabilities of a major league game going x innings are compared to empirical data in Appendix 3.

##### THE RELATIVE “RARITY” OF RECORD-LONG GAMES

So how “rare” are record-long marathons? Our theoretical model predicts the 26-inning major league record game is not as rare as empirical data would indicate, but the 33-inning minor league record game and 45-inning “other” category record game are significantly more rare than empirical data would indicate.

The record for a major league game is 26 innings in Boston. According to our theoretical model for the period 1871 (beginning of the major leagues) through 2004, we should expect 53 major league marathons to have been played. In fact, there have been only 46. We should expect a 48% chance to experience a major league marathon in any given season. We should expect 0.85 major league games, or almost one, to have gone 27 or more innings by now. In fact, we have not yet had such a game in 134 years of major league play. We should expect a 10% chance to see a major league game of 27 innings or more in any given decade. Since we have been waiting almost thirteen and a half decades now, it is not at all unrealistic to expect we should very soon have a major league game go 27 or more innings. The 26-inning Boston game should be not so rare. So far, so good. Our model is realistic, and fairly consistent with empirical data.

The record for a minor league game is 33 innings. According to our theoretical model for the period 1877 (beginning of the minor leagues) through 2004, we should expect 379 minor league marathons. If the percentage of minor league games which are marathons is the same as for the major leagues, or 0.02329%, then we should expect 327 minor league marathons. In fact, I have discovered 173 through 2004, or 46% of what the model predicts, and 53% of the number to be expected if the percentage of games that are marathons is the same in the minors as in the majors. I had predicted I would find about 50% of minor league marathons.

We should expect 6.1 minor league games to have gone 27 or more innings. In fact, we have had six such games, further indication our model is doing a good job of predicting reality. We should expect 0.078 minor league games to have gone 33 innings. In fact, we have had one such game, which means our actual count is 13 times the expected count. So the 33-inning game may be very rare indeed.

We should expect a 99.0% chance we will have a minor league marathon in any given season, a 0.12% chance we will have a minor league game of 34 or more innings in any given season, a 1.18% chance of seeing a minor league game of 34 innings or more in any given decade, and an 8.5% chance of seeing a minor league game of 34 innings or more in a lifetime of 75 years.

The record for an “other” category game is 45 innings. We should expect 0.00000320 “other” category games to have gone 45 innings. We should expect 0.00000399 “other” category games to have gone 45 or more innings. In fact, we have had one game of 45 or more innings, which means our actual count of one is 312,500 times the expected count. So the 45-inning Mito game may be extremely rare indeed!

There is a 50% chance we will see a major league game go 27 innings or more in the next 55 years. There is a 95% chance we will see a major league game go 27 innings or more in the next 236 seasons. So the 84-year old 26-inning major league record, while rare, is not so rare that we should not expect to see it possibly broken someday soon.

There is a 50% chance we will see a minor league game go 34 innings or longer in the next 450 years. There is a 95% chance we will see a minor league game go 34 innings or more in the next 1,947 years. So the 23-year-old 33-inning minor league record may be very rare, and although it could be broken at any time, we should not expect to see it broken anytime soon.

There is a 50% chance we will see an “other” category game go 46 innings or more in the next 47,350 years. There is a 95% chance we will see an “other” category game go 46 innings or more in the next 204,800 years, or just two-thirds the age of mankind (about 300,000 years old). So the 21-year-old 45-inning “other” category record seems to be extremely rare indeed, and although it could be broken at any time, it is very conceivable it may never be broken.

##### EPILOGUE

As long as there is a “hot stove league,” baseball fans will argue whether somewhere back in the murky uncharted depths of undocumented baseball history there may perhaps be a game that lasted longer than 45 innings. It is deﬁnitely possible. After all, I did 40 years and 11 months of research on this subject before I found the 45-inning game. Of course, had Hank Webb’s pickoff not been wild, or had Ron Hodges not dropped the ball, or had Ed Sudol ruled the sliding Bake McBride out at the plate, or even better, had the rules on simultaneous balks and wild pickoff throws not been changed recently and Bake McBride had been required to return to second in the top of the 25th at Shea on September 11, 1974, the Cards and Mets might still be playing! With nine hours of rest daily, ﬁve months off for winter, and three hours per nine innings, or 45 innings daily and 9,450 innings annually, they would now be in the 31st year and roughly the 288,225th inning of that game. And every seven innings the fans would still be stretching and singing,

##### Appendix 1. CHRONOLOGICAL LIST OF EVERY MARATHON OF 20 OR MORE INNINGS
 101 Ag. Fair Gds. Boston, MA 9/1859 AMA Medway Unions 100 Holliston Winthrops 71, susp. after 66 innings 80 Ag. Grounds Worcester, MA 10/11/1859 10/12/1859 AMA Upton Excelsiors 100 Medway Unions 56 (11:02), suspended after 6:02 30 Janesville, WI 8/10/1860 AMA Croft’s Team 50 Hogan’s Team 34 172 Ag. Grounds Worcester, MA 9/25/1860 9/26/1860 9/27/1860 9/28/1860 10/1/1860 10/4/1860 10/5/1860 Upton Excelsiors 50 Medway Unions 29 (21:50), dinner/rain delay 0:20 bottom 13th on 9/25, susp. on 9/25 after 2:40 and 12.5 innings, rain delay 2:30 bottom 13th on 9/26, susp. on 9/26 after 5:40 and 34.5 innings, lunch delay 0:20 on 9/27, susp. on 9/27 after 12:50 and 83.5 innings, susp. on 9/28 after 17:20 and 136.5 innings, scheduled to resume 10/1 in Springﬁeld but did not, rain delay top 173rd on 10/4, susp. on 10/4 after 21:50 and 172 innings, rain delay top 173rd on 10/5 24 Holmes Field Cambridge, MA 5/11/1877 IA/ ICA Manchester Pros 0 Harvard Crimson 0 (3:30) 22 11th St. Grounds Tacoma, WA 5/16/1891 PNW Tacoma Daisies 6 Seattle Blues 5 (3:35) 25 Militia Grounds Devils Lake, ND 7/18/1891 RRV Grand Forks Forkers 0 Fargo Graingrowers 0 (4:10) 20 League Park (I) Cincinnati, OH 6/30/1892 NL Colts 7 Reds 7 (3:20)
 21 Lake View Park Peoria, IL 6/26/1898 WA Peoria Blackbirds 8 St. Joseph Saints 4 20 Old Fair Grounds Springﬁeld, MI 7/19/02 MOV Springﬁeld Reds 2 Nevada Lunatics 1 (5:00) 20 Driving Park Kingston, NY 8/10/03 HUD Hudson Marines 2 Kingston Colonials 2 23 Goodwater Grove Stockton, CA 7/2/05 CAS Stockton Millers 1 Lodi Crushers 0 (3:26) 20 Huntington Gds. Boston, MA 7/4/05 AL Athletics 4 Americans 2 (3:31) 20 Huntingdon Gds. Phila., PA 8/24/05 NL Cubs 2 Phillies 1 (4:00) 20 Electric Park Monessen, PA 1906 AMA Glassport Athletic 4 Monessen East Ends 3 24 Huntington Gds. Boston, MA 9/1/06 AL Athletics 4 Americans 1 (4:47) 20 Joplin, MO 9/4/06 WA Joplin Miners 0 Webb City Gold Bugs 0 (3:15) 20 Washington Park Lowell, MA 4/27/07 NEL Haverhill Hustlers 1 Lowell Tigers 1 (3:25) 23 Athletic Park Hutchinson, KS 5/29/07 WA OK City Mets 2 Hutchinson Salt Packers 1 (3:10) 28 Franklin Field Boston, MA 6/8/07 GRM Pierce School (Dorchester) 4 Bennet School (Brighton) 3 (5:50) 21 Jacksonville, IL 6/26/07 IAS Jacksonville Lunatics 3 Burlington Pathﬁnders 2 (3:05) 30 Brookside Park Cleveland, OH 7/4/07 AMA Brooklyn Athletic 4 East End All Stars 1 (5:50) 36 Rec. Park (I) Columbus, OH 7/5/07 AMA Heintz Victors 2 Columbus Selects 2 (3:50) 23 Green Bay, WI 7/14/07 WIS Green Bay Orphans 2 La Crosse Badgers 1 (4:05) 20 Webb Park Marion, OH 7/20/07 OPL Mansﬁeld Pioneers 2 Marion Moguls 1 (2:18) 22 Ringwood Park Clinton, IA 7/25/07 III Peoria Distillers 3 Clinton Infants 0 (3:10) 21 Johnstown, PA 8/8/07 TRI Johnstown Johnnies 4 Reading Pretzels 3 (3:45) 24 Newark, OH 8/23/07 OPL Sharon Giants 3 Newark Newks 2 22 Schaller, IA 1908 AMA Sac City 3 Schaller 3 20 Aberdeen, WA 5/6/08 NWN Butte Miners 3 Aberdeen Black Cats 3 (3:45) 23 N. Corry Fgds. Corry, PA 6/25/08 AMA Falconer 3 Corry 1 (3:50) 20 n/a 6/30/08 SMA Saginaw Wa-was 5 Jackson Convicts 4 21 Vincennes, IL 7/18/08 EIL Charleston Evangelists 4 Vincennes Alices 3 21 Athletic Gds. Sheboygan, WI 7/19/08 LAL Sheboygan Chairmakers 1 Mil. White Sox 0 (3:40) 20 Vaughn Ballpark Portland, OR 8/2/08 PCL SF Seals 6 Portland Beavers 5 (3:40) 23 Fond du Lac, WI 8/4/08 WI Oshkosh Indians 4 Fond du Lac Cubs 2 (3:40) 20 Jacksonville, IL 9/3/08 CA Jacksonville Lunatics 4 Ottumwa Packers 1 (2:50) 26 Bloomington Gds. Bloomington, IL 5/31/09 III Decatur Commodores 2 Bloomington Bloomers 1 (4:20), rain delay 0:15 bottom 5th 24 Freeman’s Park San Fran., CA 6/8/09 PCL SF Seals 1 Oakland Oaks 0 (3:35) 26 Athletic Park Dixon, IL 6/25/09 SMP Dixon Browns 3 Muscatine Independents 2 (4:00) 21 McPherson, KS 7/27/09 KSS Lyons Lions 2 McPherson Merry Macks 1 (2:50) 22 Recreation Park Vancouver, BC 7/31/09 NWN Portland Colts 3 Vancouver Beavers 0 (3:08) 21 Athletic Grounds Sheboygan, WI 6/5/10 LAL Sheboygan Chairmakers 2 Port Washington 1 24 Clarksburg, WV 7/3/10 WVA Mannington Mountaineers 1 Clarksburg Bees 1 23 League Park San Antonio, TX 7/5/10 TX Waco Navigators 1 San Antonio Bronchos 1 (4:14) 21 Springbrook Pk. South Bend, IN 7/13/10 CTR S. Bend Bronchos 1 Zanesville Potters 0 (3:35) 20 Muscatine, IA 7/16/10 NA Muscatine Pearl Finders 2 Sterling Infants 1 20 Kirksville, MO 5/25/11 MOS Kirksville Osteopaths 2 Macon Athletics 1 24 Pottsville, PA 6/18/11 AMA Reading Ruth 1 Pottsville Alerts 0 (4:00) 21 Ironwood, MI 8/6/11 UPL Cary Empires 4 Ironwood Tigers 2 (4:05) 20 Panther Park Fort Worth, TX 8/19/11 TXO/ TX Cleburne Railroaders 0 Fort Worth Panthers 0 24 Buffalo Park Sacramento, CA 9/10/11 PCL Portland Beavers 1 Sacramento Sacts 1 (3:45) 21 Washington Pk. Los Angeles, CA 9/15/11 PCL Sacramento Sacts 4 Los Angeles Angels 4 (3:10) 20 n/a 1912 or 1913 HS East Berlin vs. New Berlin 20 n/a 6/19/13 IML Champaign Velvets 6 Kankakee Kanks 5
 20 Adrian, MI 7/17/13 SMA Battle Creek Crickets 1 Adrian Champs 1 20 West Side Pk. Jersey City, NJ 8/12/13 IL Tor. Maple Leafs 0 Jersey City Skeeters 0 (3:20) 21 Kankakee, IL 5/25/14 IML Streator Boosters 2 Kankakee Kanks 2 20 Washington Pk. Los Angeles, CA 5/27/14 PCL Oakland Oaks 4 Venice Tigers 2 (4:02) 20 Nicollet Pk. Minneapolis, MN 6/9/14 AA Minn. Millers 3 Louisville Colonels 2 (3:31) 23 Athletic Park KS 7/3/14 KSS Hutchinson Salt Packers 4 Great Bend Millers 3 23 Brewster Park New Haven, CT 7/14/14 EA Hartford Senators 2 N.H. White Wings 1 (3:55) 21 Forbes Field Pittsburgh, PA 7/17/14 NL Giants 3 Pirates 1 (3:42) 28 Delaware Cty, OH 7/18/14 AMA Columbus Champions 1 Delaware Stars 0 22 Sioux City, IA 7/19/14 WL Wichita Wolves 3 Sioux City Indians 2 (4:48) 20 Dugdale Pk. (II) Seattle, WA 7/19/14 NWN Spokane Indians 6 Seattle Giants 1 (3:03) 20 Charleston, WV 6/27/15 OHS Lexington Colts 5 Charleston Senators 2 (3:25) 22 Athletic Park Burlington, IA 6/27/15 CA Keokuk Indians 0 Burlington Pathﬁnders 0 (3:50) 23 Wehrle Park Newark, OH 7/4/15 AMA Knights of Columbus 2 Athletics 2 21 Durham, NC 7/5/15 NCS Raleigh Capitals 2 Durham Bulls 2 23 Norfolk, VA 8/14/15 VAL Suffolk Tigers 2 Norfolk Tars 2 (3:40) 22 n/a 9/4/15 VAL Norfolk Tars 3 Suffolk Tigers 3 22 Douglas Park Rock Isl., IL 7/9/16 III Hannibal Mules 8 Rock Island Islanders 3 (4:00) 20 Gulfview Park Galveston, TX 8/13/16 TX Waco Navigators 4 Galveston Pirates 1 21 Lincoln, NE 8/12/17 WL Lincoln Links 2 Joplin Miners 1 (3:38) 22 Ebbets Field Brooklyn, NY 8/22/17 NL Robins 6 Pirates 5 (4:15) 20 St. Joseph, MO 8/26/17 WL Des Moines Boosters 4 St. Joe Drummers 3 (4:00) 20 Panther Park Ft. Worth, TX 5/8/18 TX Shreveport Gassers 1 Fort Worth Panthers 1 20 St. Joseph, MO 6/1/18 WL Joplin Miners 3 St. Joseph Saints 1 21 Weeghman Park Chicago, IL 7/17/18 NL Cubs 2 Phillies 1 (4:00) 21 Braves Field Boston, MA 8/1/18 NL Pirates 2 Braves 0 20 Nat. Lg. Park (II) Philadelphia, PA 4/30/19 NL Robins 9 Phillies 9 (4:00) 23 Andrews Field Chattanooga, TN 6/13/19 SL Atlanta Crackers 2 Chatt. Lookouts 2 (3:40) 20 Sulphur Dell (I) Nashville, TN 7/12/19 SL Chattanooga Lookouts 6 Nashville Volunteers 5 20 Charlotte, NC 8/30/19 SAL Columbia Comers 5 Charlotte Hornets 5 21 Savin Road Park New Haven, CT 8/31/19 EL Worcester Boosters 4 New Haven Weissmen 3 26 Braves Field Boston, MA 5/1/20 NL Robins 1 Braves 1 (3:50) 20 Hamilton Grounds Hamilton, ON 6/17/20 MO London Cockneys 5 Hamilton Tigers 4 20 Hampden Pk. (II) Springﬁeld, MA 8/28/20 EL Albany Senators 5 Springﬁeld Hampdens 5 22 Washington Pk. Los Angeles, CA 4/10/21 PCL Seattle Rainiers 12 Los Angeles Angels 8 (4:46) 20 Hanlan’s Pt. (III) Toronto, ON 6/2/21 IL Buffalo Bisons 3 Toronto Maple Leafs 2 22 Stonewall Jackson Danville, VA 7/9/21 PML High Point Furniture Makers 7 Danville Tobac- conists 6 23 Rock Island, IL 8/4/22 MIV Ottumwa Cardinals 4 Rock Island Islanders 2 20 Schorling’s Park Chicago, IL 8/16/22 NNL Chi. American Giants 1 Bacharach Giants 0 (3:38) 20 Akron, OH 8/20/22 IND Firestone 0 General Tire 0 21 n/a 5/18/23 NYP Scranton Miners vs. Binghamton Triplets 20 Chadwick Park Albany, NY 5/30/24 EL Pittsﬁeld Hillies 9 Albany Senators 8 20 Brown College Providence, RI 6/7/24 NCA Providence Friars 1 Brown Bears 0 (4:00) 20 Ace Park Saginaw, MI 7/12/24 MO Flint Vehics 3 Saginaw Aces 2 23 n/a 1926 HS Pomona Red Devils 6 Fullerton Indians 5 25 Paciﬁc Junc., IA 8/14/26 AMA Folsom 7 Pleasant Valley 4 (5:00) 22 Braves Field Boston, MA 5/17/27 NL Cubs 4 Braves 3 (4:13) 22 Oakland Baseball Park Emeryville, CA 5/6/28 PCL Oakland Oaks 7 Sacramento Senators 6 (4:00) 20 Knauertown HS Diamond Knauertown, PA 6/2/28 TCL Boyertown 9 Warwick Athletic 8 (3:30)

 21 Comiskey Park (I) Chicago, IL 5/24/29 AL Tigers 6 White Sox 5 (3:31) 21 n/a 1930 NCA U. Miami Redskins def. U. Cincinnati Bearcats 20 Wrigley Field Chicago, IL 8/28/30 NL Cardinals 8 Cubs 7 (4:10) 20 IA between 5/14/31 & 5/31/31 MIV Cedar Rapids Bunnies def. Dubuque Tigers 22 Elyria, OH 8/1/1932 8/8/1932 AMA K of P’s Team A 19 K of P’s Team B 19, susp., completed on 8/8 with unknown result 25 Koshien Kyujyo Osaka, Japan 8/19/33 HS Chukyo-shogyo 1 Akashi-chugaku 0 (4:55) 21 Wauchula, FL 9/20/33 OBL Wauchula 7 Frostproof 4 (4:00) 21 Tech Field San Antonio, TX 6/2/35 TX San Antonio Missions 4 Dallas Steers 3 (3:50) 20 Albemarle Field York, PA 7/25/36 YAL Albemarle Black Cats 3 New Oxford 2 (3:35) 25 Honolulu Park Hilo, HI 2/20/38 HJL Papaikou 5 Shinmachi 5 (4:55) 20 Hook Ball Park Paducah, KY 7/8/38 KIT Fulton Eagles 14 Paducah Indians 9 (5:30) 21 Clintonville, WI 6/15/39 NSL Clintonville FWD Truckers 1 Two Rivers Polar Bears 0 (5:15) 20 La Grave Field (I) Fort Worth, TX 5/31/39 TX Fort Worth Cats 4 Oklahoma City Indians 3 (4:25) 23 Braves Field Boston, MA 6/27/39 NL Dodgers 2 Bees 2 (5:15) 20 Engel Stadium Chattanooga, TN 6/21/40 SA Little Rock Travelers 7 Chattanooga Lookouts 4 20 Braves Field Boston, MA 7/5/40 NL Dodgers 6 Bees 2 (5:19) 21 Texas Lg. Park Oklahoma City, OK 9/5/40 TX Dallas Rebels 2 Oklahoma City Indians 1 (3:58) 28 Korakuen Kyujyo Tokyo, Japan 5/24/42 JFD Taiyo 4 Nagoya 4 (3:47) 20 Swayne Field Toledo, OH 7/11/42 AA Louisville Colonels 6 Toledo Mud Hens 6 (4:25) 21 Oriole Park (III) Baltimore, MD 4/23/43 IL Toronto Maple Leafs 2 Baltimore Orioles 2 20 Cerveceria Tropical Havana, Cuba 12/2/43 CWL Cienfuegos Elefantes 6 Marianao Frailes Grises 5 (4:25) 21 Guam 1944/5 MIL Seabees vs. Rinkeydinks 20 South Bend, IN 7/14/45 MIL/ NCA Notre Dame Fighting Irish 1 Indianapolis Stout Fld Airmen 0 24 Shibe Park Philadelphia, PA 7/21/45 AL Tigers 1 Athletics 1 (4:48) 21 Artillery Park Wilkes-Barre, PA 5/10/46 EL Binghamton Triplets 5 Wilkes-Barre Barons 4 (4:35) 20 Comiskey Park (I) Chicago, IL 5/12/46 NAL Indianapolis Clowns 3 Chicago American Giants 3 21 Hartwell Field Mobile, AL 8/8/46 SA Atlanta Crackers 4 Mobile Bears 4 (4:30) 20 Lucky Beavers Stad. Portland, OR 5/7/47 PCL Portland Beavers 1 Sacramento Solons 0 (3:25) 20 n/a 5/15/47 HS Newtown HS 3 Bryant HS 2 21 Van Voorhis, PA 6/22/47 PCR Bentleyville 4 Van Voorhis 4 (5:00) 22 Horlick Field Racine, WI 7/31/47 AAG South Bend Blue Sox 4 Racine Belles 3 (4:30) 20 Thetford Mines, QB 9/7/47 QET Drummondville Forestiers Catholiques 3 St-Mau- rice de Thetford Mines Mineurs 3 (4:50) 21 Municipal Park Sanford, FL 5/27/48 FLS Sanford Giants 8 Palatka Azaleas 7 (5:15) 20 Playland Park South Bend, IN 9/10/48 AAG South Bend Blue Sox 3, Grand Rapids Chicks 2 23 Groton, NY 9/19/48 STR Homer Braves 0 Groton 0 (4:30) 22 Marsh Field Muskegon, MI 1949 AAG Muskegon Lassies 1, Rockford Peaches 0 (3:02) 26 Estadio Nacionale Managua 7/10/49 N1D Navarro Cubs 4 Escuelas Internacionales 3 (6:30) 20 Municipal Stadium San Jose, CA 8/7/49 CAL Ventura Yankees 4 San Jose Red Sox 4 (4:20) 23 Municipal Stadium Hagerstown, MD 9/3/49 INT York White Roses 3 Hagerstown Owls 2 (4:25) 20 Redbird Stadium Columbus, OH 9/3/49 AA Columbus Redbirds 4 Louisville Colonels 3 (3:58) 27 Managua late 40s N2D Manta Nica defeated Schumann (6:13) 20 7/13/50 BIG Wichita Falls Spudders 2 Sherman-Denison Twins 1 22 Red Wing Stadium Rochester, NY 8/13/50 IL Red Wings 3 Jersey City Giants 2 (5:15)

 Inning 2-Inning Datapoints Runs per 2-Inning 1 245,812 .561 2 245,812 .431 3 245,812 .491 4 245,812 .497 5 245,812 .491 6 245,446 .507 7 244,856 .493 8 244,088 .486 9 243,394 .433 10 23,026 .399 11 12,906 .398 12 7,220 .385 13 4,044 .396 14 2,282 .397 15 1,240 .392 16 704 .380 17 378 .384 18 202 .302 19 116 .440 20 362 .287 21 224 .299 22 142 .338 23 88 .239 24 60 .150 25 46 .130 26 32 .156 27 24 .250 28 18 .111 29 14 .143 30 10 .300 31 8 .250 32 8 .250 33 6 .167 34 4 .000 35 4 .500 36 4 .000 37 2 .000 38 2 .000 39 2 .000 40 2 .000 41 2 .000 42 2 .000 43 2 .000 44 2 .000 45 2 .500

For just games lasting 20 or more innings, tabulated by inning for innings 10-45 based on Lowry data for all games which have gone 20 or more innings and have a box score:

 10 362 .022 11 362 .044 12 362 .028 13 362 .022 14 362 .006 15 362 .033 16 362 .006 17 362 .017 18 362 .039 19 362 .072 20-45 see above see above