Statistically Speaking: Back-To-Backs

Five years ago I did a study on back-to-backs in the WNBA to see, among other things, what impact they had on a team�s record and if it truly was more difficult to win the back end of a back-to-back than any other game.

The general conclusions were surprising:

• Winning the back-end of a back-to-back does not appear to be more difficult than winning any other randomly selected game

• A preponderance of back-to-backs has no perceivable negative impact on a team�s overall performance.

At the time, I was pretty surprised at those conclusions. You can read that original study here, but I�ll give you the Cliff Notes version.

The record of WNBA teams from 1997-2001 in the back end of a back-to-back was indeed poor at 158-190 for a 45.4 winning percentage. However, the back end game of a back-to-back is played on the road an inordinate amount of time � close to 70 percent of the time, actually. Since teams only win on the road 39.6 percent of the time, it is only natural that teams have below .500 records in those back end games. The �projected record� of teams in those back end games, based on the distribution of home and road games, was 160-188 � just two games better than the actual record.

So why am I writing about this again? Well, I meant to do it a year ago because in 2005 Dallas Mavericks owner Mark Cuban asked the Elias Sports Bureau and his stats gurus from Indiana University to do some analysis on back-to-backs in the NBA. You can read the blog he posted about their work here. His conclusions were basically a complete 180 from mine, although the studies looked at the issue in different ways.

Mark Cuban smart guy, wealthy guy

Now Cuban is a smart guy and a wealthy guy, so it is certainly possible that his study was more accurate than mine. Or it may just be that the NBA and WNBA are vastly different due to the length of their respective seasons, or the lengths of the games. So I thought I would re-do part of my study to see if anything had changed over the past five years in the WNBA, and also to add another element or two to it.

Since that first study, there have been 312 new sets of back-to-backs. Given the home/road distribution of those games (98/214), we would expect teams to have won 144 of the back ends (with 168 losses) for a winning percentage of 45.7 percent.

Here's the math on that . . .

• Home games, 98 * .604 (avg. home win pct.) = 59 wins

• Road games, 214 * .396 (avg. road win pct.) = 85 wins

• Total 144 wins

Uh oh . . . We�ve got a problem . . . In reality, teams have won 163 of those 312 back end games � 52.2 percent.

This makes no sense. If back-end games are so difficult, how is it that teams have winning records in those back-end games over the past five years? I had a couple of theories.

My first thought was that perhaps the teams that were engaging in most of these back-to-backs were the better teams in the league. Afterall, teams don�t play the same number of back-to-backs in a season � last year Detroit had eight sets while Houston had just one. If the better teams are playing the back-to-backs, then they could conceivably have better records.

However that wasn�t the case. The record of teams from 1997-2001 engaged in back-to-backs was 50.0 percent. From 2002-06, it was 49.9 percent. Essentially, no change.

It could also be that the opposition hasn�t been as good in those back-end games. But that also wasn�t the case. In the first five years of the league, back-end opponents had a winning percentage of 49.0. Since then, it has been 48.8. Once again, essentially, no change.

So then I thought maybe teams were winning more games on the road than they used to? I had rather lazily used the old road winning percentages from 1997-2001 to do my calculations.

Year	Road WPct
1997	.375
1998	.400
1999	.396
2000	.402
2001	.398
2002	.398
2003	.378
2004	.394
2005	.353
2006	.399

Nope. Nothing there either. It has actually been much more difficult than usual to win on the road in two of the past four seasons (2003, 2005).

I�m open to suggestions.

Maybe I just entered some information wrong. I�ll go back and triple-check when I get some time. Perhaps I'll shoot a copy of this to Cuban to see if he's interested in weighing in.

I also revisited how the number of back-to-backs a team has correlates to its final winning percentage, and found no correlation whatsoever. For instance, as mentioned a few graphs ago the Detroit Shock had the most back-to-backs in the league last year with eight, and their season turned out okay.

Here�s another chart for you. It shows the winning percentage of that back end game over the years.

Year	Back End WPct
1997	.333
1998	.417
1999	.538
2000	.454
2001	.462
2002	.526
2003	.416
2004	.508
2005	.500
2006	.586

As you can see, in four of the past five years, the winning percentage has been at or above .500. Before that, it was below .500 four out of five years. This past season, the back end winning percentage was close to 60 percent.

So who are the best teams at winning the back-end of back-to-backs? Here�s the complete list.

Team	W	L	Pct
Los Angeles	31	12	0.721
Houston	30	12	0.714
Connecticut	14	6	0.700
Detroit	27	22	0.551
New York	27	25	0.519
Cleveland	20	21	0.488
Utah	15	16	0.484
Sacramento	23	25	0.479
Orlando	10	11	0.476
Phoenix	21	29	0.420
Charlotte	23	32	0.418
Washington	19	27	0.413
Seattle	14	20	0.412
Portland	6	9	0.400
Chicago	2	3	0.400
Indiana	13	23	0.361
Miami	6	11	0.353
Minnesota	11	24	0.314
San Antonio	4	11	0.267