FREE hit counter and Internet traffic statistics from freestats.com

Saturday, September 20, 2008

As Time Goes By

Today we'll run another tidbit from the errata of It Ain't Over 'Til It's Over: The Baseball Prospectus Pennant Race Book...




During which decade did baseball fans enjoy the best pennant races? Bill James, in The New Bill James Historical Abstract says unequivocally that the 1940s was "The Best Decade Ever for Pennant Races". Our compilation of Race Score by decade agrees.


Decade Aggregate Races Avg
1900s 375.5 12 31.3
1910s 234.9 9 26.1
1920s 423.6 12 35.3
1930s 226.1 13 17.4
1940s 390.9 11 35.5
1950s 371.5 12 31.0
1960s 385.1 12 32.1
1970s 310.0 17 18.2
1980s 354.1 20 17.7
1990s 247.0 16 15.4
2000s 431.5 27 16.0


The 1940s pulled out the highest average Race Score although the 1920s came in a close second and actually included one more race. Interestingly, the 1930s included 13 races, the highest percentage at 65% of any decade, and nine of those were in the NL with only 1931 excluded. However, many of the races were only marginal with the 1934 NL race won by the Cardinals on the strength of a 33-12 record down the stretch taking the highest score at 29.9 and ranking 43rd.

James ranks the races of the 1940s and so here is his list alongside our ranking.


James Year Lg Score Rank Teams Winner
1 1940 AL 46.0 16 3 Detroit Tigers (90-64)
2 1944 AL 21.0 80 2 St. Louis Browns (89-65)
3 1948 AL 72.0 3 3 Cleveland Indians (97-58)
4 1946 NL 32.9 35 2 St. Louis Cardinals (98-58)
5 1949 NL 37.0 30 2 Brooklyn Dodgers (97-57)
6 1949 AL 37.0 29 2 New York Yankees (97-57)
7 1942 NL 50.9 15 2 St. Louis Cardinals (106-48)
8 1941 NL 36.5 31 2 Brooklyn Dodgers (100-54)
9 1945 AL 18.0 86 2 Detroit Tigers (88-65)
10 1945 NL 29.9 43 2 Chicago Cubs (98-56)
1947 NL 9.8 2 Brooklyn Dodgers (94-60)


The only race from the 1940s that James doesn't include is the 1947 NL race which ranks 121st on our list in which the Dodgers overcame the Braves at midseason and held off the Cardinals, winning by a margin of five games. As James notes, the NL races of the 1940s were dominated by the Dodgers and Cardinals while in the AL the races were more diverse.

This compilation by decade also reinforces the notion that modern races garner lower race scores overall as not only the average Race Score has declined but also the number of races that have positive scores has fallen from around 57% before divisional play to 44% after.

But since our Race Score gives extra weight to races with multiple teams with good records, this trend can also be attributed to an increasing competitive balance over time. As shown in the graph below for the AL from 1901-2005, the standard deviation in winning percentage has noticeably declined over time (albeit with a number of bumps along the way and a small upturn in the past five years) as the dotted linear trend line indicates. As more teams are bunched closer together, it is statistically less likely that two or more teams will break away from the pack and therefore score very highly in the Race Score metric.



Just why competitive balance has generally increased with time is another story. The most accepted notion, popularized by the late paleontologist and baseball fan Stephen Jay Gould in the context of the disappearance of the .400 hitter , wrests upon two pillars. First, as knowledge about how to play the game has improved and become standardized it has become more difficult for players and hence teams to take advantage of their less skilled competitors. Second, the general level of play has increased due to better athletes produced through a larger population from which the best players are chosen, better diet and training, and better technology, all of which moves the game closer to the limits of human ability providing less space for variation. In the end that leaves great players and great teams, in Gould's words*, less "space for taking advantage of the suboptimality of others".


* The 1996 book Full House: The Spread of Excellence from Plato to Darwin by Stephen Jay Gould contains an extended discussion of Gould's argument. Also see my column "Schrodinger's Bat: The Myth of the Golden Age".

Tuesday, September 16, 2008

The Wheel of Change

With the firing of Ned Yost yesterday with less than two weeks to go in the regular season, I thought it would be interesting to continue the errata from It Ain't Over 'Til It's Over: The Baseball Prospectus Pennant Race Book with this tidbit on managerial changes and pennant winners.



After dropping a 4-2 contest to Brooklyn at Ebbets Field on Tuesday August 2nd 1932, the Cubs under manager Rogers Hornsby were sitting at 53-46 in second place staring up at the Pirates who held a 5-game lead. As in his three previous managerial jobs Hornsby rubbed the powers that be the wrong way and William Veeck (father of the more famous Bill Veeck) ousted him and his $8,000 per month salary "for the best interests of the club" as the Cubs traveled on to Philadelphia. In his place he ensconced first baseman Charlie Grimm. After an off-day the team beat the Phillies 12-1 with the Pirates dropping a doubleheader to the Dodgers to shrink the lead to three and a half games. Under Grimm the Cubs went a sizzling 37-18 the rest of the way including a 14-game winning streak from the 20th of August through the 3rd of September. The Pirates meanwhile struggled to a 27-26 finish thereby propelling the Cubs to the pennant by a comfortable 4-game margin and coming in 132nd in our rankings.

The distinction of the 1932 season was that it was the first time in the modern era that a team changed managers in mid season and went on to win the pennant. Perhaps inspired by the move owner P.K. Wrigley in 1938, with his team in third place with a record of 45-36 5.5 games out, fired Grimm at the end of July and replaced him with catcher Gaby Hartnett. In no small part the hiring of Grimm as the needed sparkplug was based on the results of a study performed by a University of Illinois professor Wrigley hired to psychoanalyze the team. While team chemistry is often derided, perhaps the professor was onto something. Down the stretch the Cubs posted a 44-27 record including a 10-game winning streak from September 22-28. And of course famously it was manager Hartnett who homered in the bottom of the ninth on September 28th with darkness threatening to give the Cubs a 6-5 win over the Pirates and the league lead which they would not relinquish. That 1938 race ranked 70th in our list and so in the span of six years the Cubs had twice replaced their manager well into the season and both times it paid dividends.

Although these first two occurrences were wildly successful, it would take another 40 years and the advent of divisional play before the fateful summer of 1978 would see a similar occurrence. With the Yankees sitting at 52-43 in fourth place and behind the Red Sox by 10.5 games on the morning of July 25th, Bob Lemon would replace Billy Martin (Dick Howser managed one game in the interim) who resigned after disparaging comments referencing star Reggie Jackson and owner George Steinbrenner. Lemon is credited with calming the stormy sea and the team went on a 47-20 tear to tie for the AL East title and then…well, the rest as they say is history.

It wasn't long after, that three teams in the fateful summer of 1981 would make the post season after having made a managerial change.

  • The Kansas City Royals slumped badly in the first half after their 1980 World Series appearance posting a 20-30 record under manager Jim Frey. After opening the second half 10-10 Frey was dismissed in favor of Dick Howser and the Royals went 20-13 the rest of the way winning the second half Western Division title.


  • The Montreal Expos were in a similar situation posting a 30-25 record in the first half under Dick Williams. A 14-12 record to begin the second half led to his ouster and replacement by Jim Fanning who guided the team to a 16-11 finish and the Expos only post season appearance.


  • The Yankees led by Gene Michael won the AL East first half title with a record of 34-22. However, a 14-12 start to the second half led to Michael's replacement by the miracle worker of 1978, Bob Lemon. This time though, the Yanks did not respond and went 11-14 the rest of the way before picking themselves back up and beating the Brewers and the A's enroute to the World Series. This was the only time in history that a post season team's replacement manager had a worse record than the manager being replaced.


  • But none of these were, statistically speaking anyway, the biggest turnarounds correlated with managerial changes by post season teams. That honor goes to the 1989 Toronto Blue Jays. After enduring a 12-24 (.333) start under Jimmy Williams, General Manager Pat Gillick hired Cito Gaston on May 31st as the interim manager. That interim title was quickly forgotten as the Jays reeled off a 77-49 (.611) record with the help of acquisitions Lee Mazzilli and Mookie Wilson from the Mets leading to a 20-9 August that saw them pull into a first place tie with the surprising Orioles as the month closed. After holding a slim lead most of the month of September, the Jays hooked up with the Orioles in a three game series at the new Sky Dome (opened in June and host to a new Major League attendance record of almost 3.4 million fans) on the season's final weekend with the Orioles one game back. The Blue Jays took the first two games of the series 2-1 and 4-3 to seal the deal and come in 129th in our rankings. The difference in winning percentage after the managerial change of .278 was the largest in history by a wide margin.

    All of the races already mentioned and a few more where post season teams have made managerial moves are shown in the table below and sorted by change in winning percentage.


    Year Team Lg Manager W L Pct Replaced By W L Pct Change
    1989 Toronto AL Jimmy Williams 12 24 0.333 Cito Gaston 77 49 0.611 0.278
    2003 Florida NL Jeff Torborg 16 22 0.421 Jack McKeon 75 49 0.605 0.184
    1981 Kansas City AL Jim Frey 30 40 0.429 Dick Howser 20 13 0.606 0.177
    1978 New York AL Billy Martin 52 43 0.547 Bob Lemon 48 20 0.706 0.159
    2004 Houston NL Jimmy Williams 44 44 0.500 Phil Garner 48 26 0.649 0.149
    1932 Chicago NL Rogers Hornsby 53 46 0.535 Charlie Grimm 37 18 0.673 0.137
    1982 Milwaukee AL Buck Rodgers 23 24 0.489 Harvey Kuenn 72 43 0.626 0.137
    1983 PhiladelphiaNL Pat Corrales 43 42 0.506 Paul Owens 47 30 0.610 0.105
    1988 Boston AL John McNamara 43 42 0.506 Joe Morgan 46 31 0.597 0.092
    1938 Chicago NL Charlie Grimm 45 36 0.556 Gabby Hartnett 44 27 0.620 0.064
    1981 Montreal NL Dick Williams 44 37 0.543 Jim Fanning 16 11 0.593 0.049
    1996 Los Angeles NL Tommy Lasorda 41 35 0.539 Bill Russell 49 37 0.570 0.030
    1981 New York AL Gene Michael 48 34 0.585 Bob Lemon 11 14 0.440 -0.145


    A few notes:

  • Fifteen years after his replacement by Gaston, Jimmy Williams was once again shoved aside in favor of Phil Garner who led the Astros to a playoff appearance in 2004 making Williams the only manager to capture such a "distinction".


  • "Trader Jack" McKeon captures the second biggest turnaround with the Marlins 75-49 finish on the way to their second World Championship. McKeon was no stranger to big turnarounds. On May 23, 1978 the A's were leading the AL West by two games with a record of 24-15 when manager Bobby Winkles, deciding he'd had enough of Charlie Finley, stepped down. McKeon replaced Winkles, who ironically had replaced him the previous season, and the A's went on to post a 45-78 record good for the largest drop in winning percentage after a managerial change at -.250.


  • The replacement in 1982 of Buck Rodgers by Harvey Kuenn was told in colorful detail by Daniel Okrent in his classic 9 Innings. Mike Caldwell, Ted Simmons, and Rollie Fingers were among the most vocal of Rodgers critics. In fact, a public tirade by Fingers after he wasn't brought in against a lefty in the ninth inning of a May 31st loss sealed the coffin.


  • From an analysts perspective the thing to note is that except in the cases of the first three teams listed in the table, all the rest were respectable to good teams who simply played better once their new managers were in place. The aggregate winning percentage of these thirteen before the change was .512 while after it skyrocketed to .616. In other words, these teams were already in a position to succeed.

    Aside from these teams there have been 276 others since 1900 (not counting the 1961-62 Cubs whose famous "college of coaches" experiment failed) that have employed multiple managers (with the 1937 Tigers and 1968 White Sox employing five managers each). Obviously the vast majority of managerial changes engender no such turnaround. Even so, considering only the 52 teams who already had a .500 record or greater when their first manager was replaced, we find that roughly 20% (13 of the now 65) of the teams equipped to win, went on to post season play after changing managers. Most front offices would take those odds. Knowing when to pull the trigger, on the other hand, is the tough part.