Dynamic Duos: Are two superstars enough?

Last August, I wrote about Matt Kemp, Clayton Kershaw and the 2011 Dodgers season. At the time of that post, the Dodgers sat in fourth place of the NL West, 11.5 games behind eventual division winner Arizona. Los Angeles was ten games below .500 despite having two stars with WAR's above 5. The Dodgers would finish the season above .500, winning 82 games, but still 11.5 games behind Arizona. Kershaw won the Cy Young and Kemp should have won the NL MVP. Kemp and Kershaw formed baseball's most dynamic duo of teammates, but their ability was not enough to make the Dodgers a playoff team. My post on the Dodgers' superstars was short (and incorrect on some points), but it concluded with an interesting (but quite brash) point:

"This is part of what makes baseball so great, a team needs a solid 25-man roster to win. In basketball, having two superstar players would make a team a legit contender, but that is not the case in baseball. Kershaw can only pitch in 1 out of every 5 games, and Kemp cannot play ever position or hit in each spot in the order; thus, having two superstars with no one around them is just an incredible waste of talent and value out of two of baseball's brightest stars."

The sample size backing that conclusion was unbelievably small. One season from one team and two players is not enough to back such an overarching concept, despite the fact that the conclusion may make some sense abstractly.

It seems so obvious that over the course of a 162-game season, in which major league rosters fluctuate so much that anywhere from 40 to 50 players will appear in a game for each of franchise, that a team who get the majority of their value from just two players would not be a contender. The idea made a ton of sense, but I had so little evidence to back that claim. Thus, I decided to consider the greatest duo of superstars in baseball history.

Babe Ruth and Lou Gehrig are baseball's quintessential 1-2 punch. Ruth and Gehrig were so dominant that they finished one and two in terms of rWAR, twice. In 1927 and 1930 the Yankees' stars led the world in rWAR, but those were two very different seasons for their franchise.

In 1927, Ruth and Gehrig combined for an incredible total WAR of 23.6, but they had some serious help. The '27 Yankees were nicknamed Murderer's Row for a reason, that team had a total WAR of 64.9 (third all-time), winning 110 games and a World Series. That team had so many great players that Ruth and Gehrig's unbelievable contributions were only 36.3% of the team's value.

In 1930, Gehrig and Ruth combined for less WAR, but still a phenomenal 19.4 wins. That Yankees team did not win the World Series like the '27 club, finishing in third place of the AL, with just 86 wins and a team WAR of 39.1. Ruth and Gehrig had less support that season, especially from a pitching staff that boasted a paltry WAR of just 1.3. Their contributions were worth almost half (49.6%) of the team's total value, but not enough to lead the Bronx Bombers to a championship.

These results make sense; the Yankees had a better overall team in 1927 than '30 and in turn achieved more success. While the results make sense in that aspect, they were also slightly surprising.

Baseball was a different game eighty years ago; which puts a twist on comparing a duo like Kershaw and Kemp to Gehrig and Ruth.

Here's a brief look at differences between the eras:

1930	2012
16 Franchises	30 Franchises
154 game season	162 game season
Color Barrier	No Color Barrier
No DH	DH-rule
> 80% of IP by starters	<70% of IP by starters
Rosters made up of ~40% pitchers	Rosters made up of ~50% pitchers
~20% of Relief IP by primary Relievers	~70% of Relief IP by primary Relievers

**Some of this data courtesy of Colin Wyers' chapter on relief usage in BP's Extra Innings

The increases in games, teams, designated hitters, talent pool and relief pitchers have allowed more individuals to contribute on professional baseball teams than ever before. Today's rosters are specialized with a vast number of players bringing value (or hurting) their team, and accumulating WAR. The climate during the time of Ruth and Gehrig seems to be more suitable for a team with baseball's two best players to be completely dominant. But the 1930 clubs seems to lend credence to the notion that even then having two superstars wasn't enough to lift a ballclub to the championship level.

The sample size here is still incredibly small (only three teams considered), so I decided to look into teams with two superstars (dynamic duos), in the two eras. I first defined a "superstar" as any player who was better (higher WAR) than half of the league's best player. Thus, for the sample that stretched from 1926-39, a superstar was any player in the top-8 (16 teams), in terms of WAR, and for the 1998-2011 sample, a superstar was a player who finished in the top-15 (30 teams).

The 1926-1939 Sample:

Dynamic Duo (Year)	Team	rWAR	Team rWAR	% of Team	Wins	Season Result
Wes Ferrell/Lefty Grove (1935)	Red Sox	18.8	25.2	74.60%	78	4th in AL
Mel Ott/Carl Hubbell (1932)	Giants	14.7	24.4	60.25%	72	6th in NL
Lefty Grove/Wes Ferrell (1936)	Red Sox	17.8	33.6	52.98%	75	6th in AL
Tommy Thomas/Ted Lyons (1927)	White Sox	15.2	29.8	51.01%	70	5th in AL
Babe Ruth/Lou Gehrig (1930)	Yankees	19.4	39.1	49.62%	86	3rd in AL
Babe Ruth/Lou Gehrig (1926)	Yankees	17.5	38.0	46.05%	91	Lost World Series
Mel Ott/Carl Hubbell (1936)	Giants	17.0	37.4	45.45%	92	Lost World Series
Charlie Gehringer/Hank Greenberg (1937)	Tigers	14.8	33.8	43.79%	89	2nd in AL
Babe Ruth/Lou Gehrig (1928)	Yankees	18.8	44	42.73%	101	Won World Series
Jimmie Foxx/Lefty Grove (1932)	Athletics	18.6	44.9	41.43%	94	2nd in AL
Lefty Grove/Jimmie Foxx (1939)	Red Sox	13.5	32.9	41.03%	89	2nd in AL
Paul Waner/Burleigh Grimes (1928)	Pirates	13.2	32.8	40.24%	85	4th in NL
Babe Ruth/Tony Lazzeri (1929)	Yankees	14.9	38.4	38.80%	88	2nd in AL
Lou Gehrig/Lefty Gomez (1934)	Yankees	17.5	45.4	38.55%	94	2nd in AL
Lefty Grove/Al Simmons (1930)	Athletics	16.4	44.1	37.19%	102	Won World Series
Babe Ruth/Lou Gehrig (1931)	Yankees	18.6	50.9	36.54%	94	2nd in AL
Babe Ruth/Lou Gehrig (1927)	Yankees	23.6	64.9	36.36%	110	Won World Series
Mel Ott/Carl Hubbell (1934)	Giants	14.1	39.4	35.79%	93	2nd in NL
Jimmie Foxx/Joe Cronin (1938)	Red Sox	13.6	39.1	34.78%	88	2nd in AL
Left Gomez/Joe DiMaggio (1937)	Yankees	16.8	49.3	34.08%	102	Won World Series
Charlie Gehringer/Hank Greenberg (1935)	Tigers	15.0	45.1	33.26%	93	Won World Series
Lefty Grove/Al Simmons (1931)	Athletics	16.3	50.5	32.28%	107	Lost World Series
Babe Ruth/Lou Gehrig (1932)	Yankees	15.7	49.9	31.46%	107	Won World Series
Al Simmons/Jimmie Foxx (1929)	Athletics	14.9	51.0	29.22%	104	Won World Series

The 1998-2001 Sample:

Dynamic Duo (Year)	Team	WARP	Team WARP	% of Team	Wins	Playoff Result
Clayton Kershaw/Matt Kemp (2011)	Dodgers	14.5	23.2	62.50%	82	Failed to Reach
Darin Erstad/Troy Glaus (2000)	Angels	16.4	28.3	58.09%	82	Failed to Reach
Albert Pujols/Scott Rolen (2006)	Cardinals	13.6	27.4	49.82%	83	Won World Series
John Olerud/Mike Piazza (1998)	Mets	14.9	33.7	44.21%	88	Failed to Reach
Albert Pujols/Ryan Ludwick (2008)	Cardinals	16.3	37.7	43.19%	86	Failed to Reach
Randy Johnson/Luis Gonzalez (2001)	Diamondbacks	17.3	40.4	42.86%	92	Won World Series
Albert Pujols/Matt Holliday (2010)	Cardinals	14.0	33.0	42.52%	86	Failed to Reach
Barry Bonds/Rich Aurilla (2001)	Giants	18.1	44.7	40.47%	90	Failed to Reach
Barry Bonds/Jeff Kent (2002)	Giants	17.7	43.9	40.30%	95	Lost World Series
Albert Pujols/Scott Rolen (2004)	Cardinals	16.8	41.9	40.11%	105	Lost World Series
Miguel Tejada/Brian Roberts (2005)	Orioles	12.1	30.6	39.62%	74	Failed to Reach
Joe Mauer/Johan Santana (2006)	Twins	12.2	31.3	39.16%	96	Lost in ALDS
Nomar Garciaparra/Pedro Martinez(1999)	Red Sox	16.8	43.2	38.89%	94	Lost in ALCS
Alex Rodriguez/Derek Jeter (2005)	Yankees	14.9	39.5	37.71%	95	Lost in ALDS
Ben Zobrist/Evan Longoria (2009)	Rays	13.6	37.0	36.80%	84	Failed to Reach
David Wright/Jose Reyes (2008)	Mets	12.8	35.3	36.21%	89	Failed to Reach
Albert Pujols/Edgar Renteria (2003)	Cardinals	14.7	40.7	36.10%	85	Failed to Reach
Chase Utley/Jayson Werth (2009)	Phillies	12.2	34.7	35.21%	93	Lost World Series
Troy Tulowitzki/Matt Holliday (2007)	Rockies	12.3	35.4	34.84%	90	Lost World Series
Albert Pujols/Jim Edmonds (2005)	Cardinals	13.8	39.5	34.83%	100	Lost in NLCS
Alex Rodriguez/Jorge Posada (2007)	Yankees	14.5	42.4	34.23%	94	Lost in ALDS
Chase Utley/Jimmy Rollins (2007)	Phillies	13.0	39.5	32.81%	89	Lost in NLDS
Alex Rodriguez/Ken Griffey Jr. (1998)	Mariners	12.3	38.6	31.87%	76	Failed to Reach
Jim Thome/Jermaine Dye (2006)	White Sox	11.1	35.5	31.22%	90	Failed to Reach
Gary Sheffield/Marcus Giles (2003)	Braves	14.0	45.1	31.13%	101	Lost in NLDS
Curt Schilling/Randy Johnson (2002)	Diamondbacks	12.6	41.0	30.68%	98	Lost in NLDS
Nomar Garciaparra/Pedro Martinez(2002)	Red Sox	12.7	41.5	30.63%	93	Failed to Reach
Barry Bonds/Jeff Kent (2000)	Giants	15.6	51.1	30.63%	97	Lost in NLDS
Derek Jeter/Bernie Williams (1999)	Yankees	13	43.9	29.61%	98	Won World Series
Kevin Youkillis/Jason Bay (2009)	Red Sox	12.0	45.2	26.64%	95	Lost in ALDS
Nomar Garciaparra/Pedro Martinez(2003)	Red Sox	12.5	49.6	25.14%	95	Lost in ALCS

Baseball Prospectus' WARP was used for 1998-2011 and Baseball-Reference's WAR was used for 1926-39.

The Results:

In the 28 MLB seasons that I sampled, there were 55 teams that boasted two superstar players (roughly two teams per season). Six of those 55 teams had the majority of their WAR come from their two superstars. All six of those teams (which included the 2011 Dodgers of the original premise) failed to reach the postseason.

The two teams to have 50+ percent of their WAR come from two stars in the 1998-2011 sample, both finished with 82 wins and outside of the postseason. The four teams from the 1926-39 sample, all finished with less than 80 wins, and finished between fourth and sixth in their respective divisions. Those four teams would not have been playoff teams even under the much different (larger) playoff structure that the teams in the 1998-2011 sample played under.

It seems clear that a majority of contributions coming from just two players is not enough to push a team to the playoffs, let alone to make a team a World Series champion. The percentage of team WAR coming from a pair of superstars for the World Series teams in the sample is also interesting:

	1998-2011	1926-39	Total
Avg. % for World Series Qualifers	39.62%	36.81%	37.63%
Avg. % for World Series Winners	40.76%	34.90%	36.65%
Highest % for World Series Qualifier	49.82%	46.05%	49.82%
Highest % for World Series Winner	49.82%	42.73%	49.82%

These numbers are interesting and useful, but are slightly skewed by an outlier. As shown above, the average percentage of team WAR that comes from two stars, for a World Series winner, is in the mid-30's (36.65%). However, one World Series winner had a percentage of WAR coming from a duo of stars much higher than the mid-30's average, almost 50% of that team's WAR in fact.

The 2006 St. Louis Cardinals are one of the strangest teams in baseball history and quite possibly the worst team to ever win the World Series. That '06 Cardinals team only won 83 games and had the lowest winning percentage of any World Series winner. Their pythagorean wins, 82, and total team WARP of 27.4, show that this club may have been even worse than their 83-78 record. The Cardinals had the fifth most wins in the NL that season; which would have left them well outside the World Series under the playoff format that was in place at the time of Ruth and Gehrig. That '06 championship was a product of circumstance, divisional and playoff structure, a hot streak and a fair amount of luck. Thus, I decided to also post the percentages for World Series winners with the '06 Cardinals excluded:

	1998-2011	1926-39	Total
Avg. % for World Series Qualifers	37.02%	36.81%	36.83%
Avg. % for World Series Winners	36.24%	34.90%	35.33%
Highest % for World Series Qualifier	42.86%	46.05%	46.05%
Highest % for World Series Winner	42.86%	42.73%	42.86%

These numbers make even more sense. Having two superstars is a great thing for a team to have, but it may also not be enough to turn a team into a competitor. My original conclusion, based on the 2011 Dodgers, was that a team needs a solid 25-man roster to win and that having two superstar players does not automatically vault a team to the status of a legit contender.

I think these numbers back this claim. Regardless of era or playoff structure, the opinion rings true; a dynamic duo is just not enough to rely on to be a championship team, no matter how incredible those two players are. Maybe the sample is still too small, but over a large chunk of time, in two very different baseball eras, a team who got the majority over their contributions from just two stars was not a championship contender. A championship team with two stars would still need at least 60-70% of their wins to come from other players.

A team that boasts an MVP and Cy Young winner (or two Cy Young winners, or two MVP candidates) is not guaranteed anything. Unless a franchise can surround their superstars with above replacement level, serviceable major leaguers, a World Series will elude their grasps.

All statistics come courtesy of Baseball Prospectus and Baseball-Reference.

For more from Glenn you can follow him on twitter @Baseballs_Econ