Is There Really Parity in the WNBA?

WNBA Analysis Archive
Storm News Archive
Kevin Pelton, | September 30, 2005
There are many annual traditions in the WNBA: Training camp opens, opening day, the playoffs, coaches complain about the playoff schedule, writers talk about the parity in the league.

As far as the last of those concerns, 2005's best example was provided by the Washington Post's Ivan Carter in previewing the season.

"As the WNBA prepares to tip off its ninth season tomorrow, 'parity' is the word that has been on the lips of many general managers and head coaches," wrote Carter.

How much do the numbers bear out this talk of parity? I decided to take a look.

The best measure of parity is probably the standard deviation of records throughout the league, which measures the spread amongst team's records. Since I used wins as opposed to winning percentage, we have to account for the fact that the WNBA's schedule has expanded from 28 games to 34. If you do the math (and I admit I had to look on the Internet for a refresher), you can estimate what the standard deviation of team wins would be if all teams were exactly equal and wins were totally random based on schedule length. What I've done is compared that to the actual standard deviation of wins each season:

Year Proj. Actual Ratio
1997 2.65 3.70 139.6
1998 2.74 7.13 260.2
1999 2.83 4.81 170.0
2000 2.83 6.29 222.3
2001 2.83 5.55 196.1
2002 2.83 4.83 163.6
2003 2.92 4.62 165.1
2004 2.92 3.67 125.7
2005 2.92 5.83 199.7

Many things here should not be surprises. For example, of course there was a lot of parity in 1997, when talent was (in theory) distributed pretty evenly. The notable exception, of course, was the Houston Comets, but that wasn't really true until Sheryl Swoopes returned to full health in 1998, which not only featured the least parity of any season in WNBA history but also the best (27-3, .900, Comets) and worst (3-27, .100, Washington Mystics) winning percentages in league history.

The talk about parity really began in 2003, following the contraction of the league by two teams and a Dispersal Draft. From this perspective, however, 2003 really featured no more parity (and in fact slightly less) than the previous season. It was in 2004 that parity really became reality; only one team (Los Angeles) won more than 20 games, only one won fewer than 10 (San Antonio) and the entire East was separated by only three games.

Last year, there were still good playoff races in fairly tight middles of each conference, but one team separated themselves in each conference at the top (Connecticut, Sacramento) and at the bottom (Charlotte, San Antonio).

Another way to look at this information is across leagues. In the NBA, the ratio of actual standard deviation to projected deviation in 2004-05 was enormous - 280.1, in keeping with the format used on the chart above. That's larger than in any WNBA season to date.

While this is the definition of parity I'd naturally start with, it's not the only one. After all, we may be less concerned with how competitive teams are within a given season and more concerned with season-to-season competitiveness. That is, does my team still have a chance next year despite being terrible last year? That's something important to consider when we think about parity's impact on attendance.

So here's another list, this time showing the league-wide correlation between a team's winning percentage the first year and the second year. Correlation shows how closely two things are related, with values closer to 1 (and -1, though in this case it's safe to rule out a negative relationship, which would mean teams that were worse in year one did better in year two) indicating a stronger relationship (that is, good teams in the first year are still good in the second year and vice versa) and values closer to zero indicating more parity (teams move up and down in the standings).

Year Correlation
1997-98 .888
1998-99 .290
1999-00 .655
2000-01 .666
2001-02 .464
2002-03 .125
2003-04 .418
2004-05 .385

(Chart note: 2002-03 means the correlation between teams' performance in 2002 and 2003.)

And now we see why the parity articles popped up in 2003. The amount of turnover amongst successful teams - headlined by Detroit's shift from worst to first in the entire league - was relatively unprecedented. (There has never been an NBA season with less year-to-year consistency.)

The two least consistent years should be no surprise. Between 1998 and 1999, the ABL folded, providing a huge infusion of talent to the WNBA. (Also, as mentioned above, 1998 was the least-competitive season ever, so it would have been hard for Houston and Washington to stay as good and as bad as they were.) The combination of a Dispersal Draft before 2003 and a fairly strong rookie class meant the league turned upside down.

What surprises me is why there wasn't more turnover in 2004, when there was a one-team Dispersal Draft as well as arguably the strongest crop of rookies since the ABL-laden class of '98. What's also interesting is how consistent the league was from year one to year two. The rookie class was presumably lacking in depth because the WNBA was still competing with the ABL for talent (the top two picks, Margo Dydek and Ticha Penicheiro, were quite good, but the fourth and sixth picks (Korie Hlede and Cindy Blodgett) were notably non-factors).

To return to our question, if you balance the two measures, it's clear that there is indeed quite a bit of parity in the WNBA. With the expansion Chicago Sky joining the league in 2006, parity will probably take a short-term hit, but with a hard salary cap and unrestricted free agency forcing good teams (like the Storm) to make tough decisions about their players, look for parity to continue to be a reality in the WNBA.