Marynell Meadors of the Atlanta Dream said something to the effect that the goal of the Dream was to win the old fashioned way. The Dream would crash the baskets and get a lot of fouls. In effect, the Dream would swap the new version of the 3-point play - the long range shot behind the 3-point arc - for the older version which was making the basket and getting the foul.
If there's anyone who knows about old-fashioned basketball, it's Marynell Meadors. She graduated from Hillsboro High School in Nashville, Tennessee in 1961 and attended Middle Tennessee State University from 1962 to 1965. Meadors coached at Tennessee Tech before Pat Summitt even started coaching. (Her record at Tennessee Tech was 363-139 from 1970 to 1986). Put it this way: when Meadors wrote in her high school yearbook that her goal was to be a women's basketball coach (true story), Nancy Lieberman was a toddler.
The 3-point line has been a part of women's basketball since the NCAA 1987-88 season. And for the last 20 years it seems, basketball fans have been either praising a team's 3-point prowess or bemoaning that their favorite players can't hit from beyond the arc. "We went 2-for-6 from beyond the 3-point line! How are we going to win if we can't hit the three?" The question then becomes "Can the Atlanta Dream - the team coached by Marynell Meadors - go very far without hitting the 3-pointer?"
This question begs two other questions:
1) Has there been a team in the WNBA like the 2010 Atlanta Dream? If so, how well did it do? Did it go to the playoffs? Did it win a championship?
2) Do teams have to be good at hitting the 3-pointer in order to win championships?
Questions like this have been consuming the time of basketball statheads since writers like Dean Oliver and John Hollinger began applying statistical tools to basketball statistics. However, basketball is a fluid game where players contribute to wins in ways that aren't very well reflected in a basketball box score. Baseball is the sport that has yielded many of its secrets to statistical metrics due to the fact that its components can be neatly isolated - batter versus pitcher. How do we do this for basketball?
Take the first question: what does it mean in basketball to be a team that is like the 2010 Atlanta Dream? How close does a team have to be? There's no team that's going to be exactly like the 2010 Dream. So how do we go about solving this puzzle.
One step would simply be to expand the stats of the 2010 Atlanta Dream out to a full season. Nine games have been played, so we multiply various stats by (34/9), enough to "blow them up" to a full 34-game season. Field goal percentage, free throw percentage, and 3-point percentage should remain the same - if you blow up shots made and shots attempted by the same multiple, the ratio is the same. This gives us a 2010 Atlanta Dream that we can put on an equal basis with other teams.
Now we define what "like" is. Some elementary work has been done with something called a "similarity score". One can simply take the season stats of two given teams and compare them stat by stat. The idea is that you start with a number - say 1000 - and take away points for any difference in field goals, or in turnovers, or in steals, etc. At the end, you should get a number that is less than 1000...the only way you could get a 1000 after the comparison is if the two teams had equal stats. The higher the final result, the more alike the results are.
I've created a similarity metric for players based on the one at databaseBasketball.com for players. We now take the players metric and apply it to our hypothetical Dream team. Here are the teams in WNBA history most similar to the 2010 Atlanta Dream (so far) in descending order.
Teams Most Similar to 2010 Atlanta Dream (so far)
1. 2009 Atlanta Dream
2. 2008 Minnesota Lynx
3. 2006 Washington Mystics
4. 2006 Connecticut Sun
5. 2007 Connecticut Sun
6. 2007 Detroit Shock
7. 2007 Seattle Storm
8. 2008 Detroit Shock
9. 2009 Minnesota Lynx
10. 2009 Detroit Shock
When introducing a metric, the very first test is the "smell test" - is there anything that would cause you to reject the results out of hand? The first part of the list looks promising: one would expect the closest team to the 2010 Dream to be the 2009 model - both teams have similar principal players. The 2007 Sun had Erika de Souza on the team. The 2007 Storm had Iziane Castro Marques on the squad.
But how similar are these teams exactly? Very similar? Somewhat similar? Not at all similiar? The appearance of the Detroit Shock multiple times - the Shock were known to be a good rebounding team - is very promising.
How did these teams do in the end? The average record of these teams is about 19-15. All in all, eight of these teams went to the playoffs, and two of those teams went to the WNBA Finals. Those teams aren't exactly like the 2010 Dream, but it's nice company to be in.
What about the teams that didn't go? The 2008 Lynx finished with a losing record - barely - at 16-18. Only the 2009 Lynx had a bad record at 14-20 and even the 2009 Lynx had Nicky Anosike in the post.
Clearly, teams "like" the 2010 Atlanta Dream can win ballgames, and even finish the season wearing championship rings. Of course, none of these teams are as like the 2010 Atlanta Dream as we'd like them to be. So maybe answering the second question should be answered - can a team be a winner if they can't shoot the three?
We need some way to distinguish three point shooting teams from teams that don't shoot the three. Two definitions come to mind:
a) A 3-point shooting team could be one that shoots a lot of threes, or
b) A 3-point shooting team might not take many threes, but could be deadly accurate with the threes that it does make.
So we need to combine 3-point attempts with 3-point accuracy. Multiplying the two gives us a very simple stat: 3-point shots made. Since not every WNBA team played a 34-game regular season schedule, we can equalize all teams by looking at "3-pointers made per game".
Who are the leaders as of 2010?
WNBA Leading Teams in 3-Pointers Per Game
1. 2007 Phoenix Mercury
2. 2006 Phoenix Mercury
3. 2008 New York Liberty
4. 2009 Phoenix Mercury
5. 2009 San Antonio Silver Stars
6. 2009 New York Liberty
7. 2007 New York Liberty
8. 2009 Minnesota Lynx
9. 2008 Phoenix Mercury
10. 2008 Indiana Fever
Are you sensing a pattern here? If you have someone who can hit the 3-pointer, your team is going to go to the top of this list.
How well did these teams do? They averaged a record of 17-17. (More like 17.4-16.6.) The good Phoenix Mercury teams were evened out by the bad Liberty and Silver Stars teams. Five had losing records and only one - the 2007 Mercury - even made it to a WNBA Finals.
Okay. Maybe those ten teams aren't the best examples. How about overall? Will having good 3-point shooting at least give you a budge?
There's one test left. That test is a simple linear correlation. What is the correlation between a team's 3-pointers per game and its final end-of-season record? The number returned in this statistical test is between a -1 (strictly reverse: one always goes up, other always goes down) and a +1 (both go up or down in lockstep order). A correlation of zero implies disconnection.
The final answer: 3-pointers made per game correlates to winning percentage on the order of 0.15. This is what we'd call a "weak" correlation. Yes, it probably helps if your team can hit threes...but it's not necessary. The connection between the two (or between the two and some other unknown thing) isn't so strong that one should conclude, "Hey, I'll get Becky Hammon and Diana Taurasi and a WNBA championship will be an absolute lock!" The number says, "hey, if hitting the three is part of having a winning team, it's only a small part of it." Perhaps, a part so small that you could simply do without it if you had another way to win.
For example, the correlation between winning percentage and effective field goal shooting percentage is much stronger: 0.54, at least a medium correlation. Or say, between winning percentage and turnovers per game: -0.28, a case where you'd expect a negative correlation.
So can you win the old-fashioned way? The numbers suggest that the answer is "yes".