RebKell's Junkie Boards

[ShockPR]

Afternoon everybody:

A week or two ago somebody posted Mark Cuban's weblog article about back-to-backs and I referenced a study I had done once. Finally tracked it down through Clay Kallum at Full Court. I'll probably try and update it at some point during the offseason since we now have another several years of data we can plug in, but here it is with apologies for the lack of coding for the charts. I hope it still makes sense.

-------------------------------------

The WNBA released its 2002 schedule a few weeks ago, which means that for the week just after, head coaches and general managers were on the phone to the league complaining about the specific injustices foisted upon their respective teams. Weather it's too many back-to-backs, not enough home weekend games or that long road-trip to the East Coast to start the season, there's usually something to complain about with regards to every team's schedule if you look at it in the right light.

But is all this hand-wringing over scheduling merely a case of people needing something about which to complain, or is there really a disadvantage inherent in some team's schedules? Today, I'll take a look at back-to-backs and their impact on a team's winning percentages.

Conventional wisdom suggests that back-to-backs are bad for teams because of the fatigue factor inherent in playing that second game -- the back end of the back-to-back. Therefore, it makes intuitive sense that teams with the most back-to-backs would suffer with regards to their won-loss record as compared to teams with fewer back-to-backs.

First assumption first, though: Do teams lose more back ends of back-to-backs then we would expect?

In the WNBA's five-year history, there have been 348 regular-season back-to-backs. Teams are 180-168 in front ends of a back-to-back for a 51.7 winning percentage, but in the back ends of back-to-backs they're 158-190 for a 45.4 winning percentage. Winning any game, league-wide, is a 50-percent proposition, so we conclude that winning the back end of a back-to-back is tougher than winning the average game.

The problem is, though, that the back end is not an average game, and I don't mean because of the fatigue factor. Stating the obvious here for a moment, not only is the average winning percentage in the league 50 percent, but 50 percent of a team's games are home games and 50 percent are road games. However -- and here's the kicker -- a disproportionate number of back end games are played on the road. In fact, 242 of 348 back ends have been played on the road -- 69.5 percent.

Here's where the math gets a little heavy. Teams win road games 39.6 percent of the time and home games 60.4 percent of the time over the league's five-year history. If we apply those winning percentages to the home/road distribution of back end games, we get the following.

106 home games * .604 = 64 wins
242 road games * .396 = 96 wins
348 total back ends = 160 wins


We would expect teams that play this distribution of home and road games, regardless of their context, to win 160 of them while losing 188. Take a quick peek back at the actual record in these back end situations: 158 and 190. In five years, teams have won just two fewer games than we would expect them to if they were playing random games throughout the schedule.

I don't know if anybody else is as shocked by that as I am. Not that it will alter anybody's game strategy or that any team that knows this can have some advantage over another team, but I always bought into the idea that fatigue played a roll in the difficulty of winning the back end of a back-to-back. Looking at the data though, playing the back end is no different then playing any other game. It's playing those games on the road that makes it seem like a negative proposition.

Studying the figures with regards to the total number of back-to-backs in a season and how they relate to a team's winning percentage paints an interesting picture as well.

Winning Pct. vs. No. of Back-To-Backs
Back-To-Backs W-L WPct
1 14-14 .500
2 0-0 -----
3 18-10 .643
4 163-207 .441
5 237-199 .544
6 326-296 .524
7 101-123 .451
8 93-99 .484
9 14-18 .438


It seems to suggest that although it's more difficult to win games with more than seven or more back-to-backs on the schedule, it's also more difficult to win games with four or fewer back-to-backs. The sample sizes in those outlying areas are pretty small, making it difficult to draw any sort of conclusions from them. Only time, additional games and additional studies will tell if what we are seeing here is significant.

The more interesting question, if these winning percentages hold up over time and are deemed significant, is why would a team that has four back-to-backs or less suffer in the same fashion as a team with seven or more? The only thing that occurs to me is that there may be other elements of those team's schedules that are impacting their winning percentages.

I took a look at one other possibility with this issue -- that we might see the effect of back-to-backs manifesting itself not in specific back-to-back games or throughout an entire season, but over the second half of the season. The thinking being that it takes time for a team to feel the wear and tear of multiple back-to-back situations. If this hypothesis were true, we would expect teams with more back-to-backs to see a decrease in their winning percentages from the first half of the year to the second, but we don't.

2nd Half Winning Pct. vs. No. of Back-To-Backs
Back-To-Backs No. of Teams 1st Half WPct 2nd Half WPct
1-4 14 .459 .458
5 14 .541 .539
6 20 .522 .529
7 7 .464 .438
8-9 7 .438 .518


Actually, the seven teams that have had eight and nine back-to-backs have improved in the second half of the season. Basically, though, this chart suggests to us that there is no effect of back-to-backs on a team's winning percentages in the second half of the season.

As of now, I'd have to conclude that there is not enough concrete evidence to suggest that back-to-backs have a negative impact on a team's winning percentage.

[dtsnms]

Admiral quoted Cuban, I mentioned that you had a study that disagreed. Thanks for sharing it!

[ShockPR]

That's right. I forgot that you and I had discussed it.

[BCBG25]

[dtsnms]

[Admiral_Needa]

[Admiral_Needa]

[BCBG25]

I guess to analyze back-to-backs you'd have to start with the players. I think players react differently. Not sure how that would add up, but I would say that it has to be factored in and not taken as a whole. Average age would be a good way to start, maybe. Or that could explain why some teams have no problems with back-to-backs. Anyways, if it's a myth, I know we won't hear much about it, but I will pay close attention to how the Shock perform on the 2nd night of their countless back-to-backs next season.

_________________
Kings of the World!

[ShockPR]

What I would have liked to do, and still may at some point, is look at the quality of opponents in those back end games and the margin of victory. It may be that even though teams are winning as many back ends of back-to-backs as they should based on the ratio of home games to road games, that the opponents in those back ends are worse than the average team or that teams aren't winning by as many points as we might otherwise expect.

[Admiral_Needa]

[ShockPR]

I guess to analyze back-to-backs you'd have to start with the players.

This is something that we generate on our XL sheets that we keep for every team in the league. I've never really used it for anything, but it would be interesting to note which players are most immune from the back-to-back impact and which players suffer big time.

Looking at last year's primary Shock players (and remembering that we're talking about a sample size of four games when it comes to back ends of back-to-backs . . .

Braxton
Back Ends - 10.3 ppg, 4.3 rpg, 48.5 FG%, 90.0 FT%
Other Games - 6.4 ppf, 2.9 rpg, 45.8 FG%, 48.0 FT%

Cash
Back Ends - 6.7 ppg, 5.3 rpg, 3.3 apg, 47.4 FG%, 33.3 3G%, 25.0 FT%
Other Games - 5.5 ppg, 4.0 rpg, 1.8 apg, 36.4 FG%, 14.3 3G%, 71.4 FT%

Ford
Back Ends - 9.3 ppg, 11.5 rpg, 38.2 FG%, 47.8 FT%
Other Games - 9.5 ppg, 9.5 rpg, 43.7 FG%, 48.8 FT%

Nolan
Back Ends - 14.3 ppg, 3.8 rpg, 1.8 apg, 45.5 FG%, 57.1 3G%, 72.2 FT%
Other Games - 16.1 ppg, 4.9 rpg, 3.9 apg, 39.2 FG%, 28.9 3G%, 81.0 FT%

Riley
Back Ends - 7.0 ppg, 5.0 rpg, 33.3 FG%, 100.0 FT%
Other Games - 7.7 ppg, 4.7 rpg, 38.0 FG%, 76.9 FT%

Pierson (Shock only)Back Ends - 6.7 ppg, 1.3 rpg, 66.7 FG%, 0.0 FT%
Other Games - 7.9 ppg, 2.9 rpg, 40.5 FG%, 70.5 FT%

Powell
Back Ends - 3.0 ppg, 3.3 rpg, 0.5 apg, 38.5 FG%, 66.7 FT%
Other Games - 6.0 ppg, 2.7 rpg, 3.0 apg, 44.2 FG%, 59.7 F%

Smith (Shock Only)
Back Ends - 8.5 ppg, 1.5 rpg, 3.5 apg, 33.3 FG%, 28.6 3G%, 100.0 FT%
Other Games - 9.7 ppg, 2.3 rpg, 1.7 apg, 38.4 FG%, 33.3 3G%, 75.8 FT%

Team Offense
Back Ends - 66.8 ppg, 39.8 rpg, 10.4 apg, 40.9 FG%, 34.8 3G%, 64.9 FT%
Other Games - 66.0 ppg, 35.2 rpg, 14.5 apg, 40.3 FG%, 31.8 3G%, 65.9 FT%

Team Defense
Back Ends - 73.0 ppg, 30.5 rpg, 16.0 apg, 40.1 FG%, 37.1 3G%, 72.7 FT%
Other Games - 66.5 ppg, 29.9 rpg, 13.6 apg, 40.3 FG%, 33.4 3G%, 75.3 FT%

Detroit's back end games last season were at New York (a win), at Sacramento (a loss), vs. Phoenix (a win) and vs. New York (a win). The Sacramento loss (which was by 40 points . . . ugh) skews things quite a bit.

[ShockPR]

Just look at the odds line from Las Vegas.

The only problem (yet a big one) with looking at Vegas lines, is that a Vegas line is not designed to be a predictor of the actual outcome. It is designed to get 50 percent of the betting population on either side of the bet so as to gaurantee a profit.

[BCBG25]

Thanks for the info, ShockPR. Very interesting indeed. You see that on a young roster like the Shock's the impact isn't too hard.
I wanna see that analysis done on the Comets' roster *

* it feels good to make fun of the Comets because they are OLD. Lord knows what us Lib fans had to put up with until after Kym, Sue and Spoon were ALL gone.

_________________
Kings of the World!

[Admiral_Needa]

[Admiral_Needa]

[womens_hoops]

[ShockPR]

Houston Comets players in back-to-backs last season . . .

Although you guys only had two back-to-backs, one of which was the first two games of the season. The first was a home win over Minnnesota on May 22 and the second was a road win over New York on July 27.

Janeth Arcain
Back Ends - 11.0 ppg, 2.5 rpg, 1.0 apg, 46.7 FG%, 100.0 3G%, 87.5 FT%
Other Games - 10.0 ppg, 2.8 rpg, 1.6 apg, 41.9 FG%, 13.3 3G%, 88.4 FT%

Dominique Canty
Back Ends - 9.5 ppg, 4.5 rpg, 3.5 apg, 38.9 FG%, 62.5 FT%
Other gzames - 8.1 ppg, 3.2 rpg, 3.0 apg, 40.0 FG%, 73.5 FT%

Sancho Lyttle
Back Ends - 5.0 mpg, 1.0 ppg, 3.0 rpg, 33.3 FG%
Other Games - 14.5 mpg, 4.5 ppg, 3.8 rpg, 59.2 FG%

Michelle Snow
Back Ends - 15.0 ppg, 8.5 rpg, 54.5 FG%, 75.0 FT%
Other Games - 11.8 ppg, 6.7 rpg, 55.1 FG%, 70.5 FT%

Sheryl Swoopes
Back Ends - 24.0 ppg, 3.0 rpg, 3.0 apg, 45.9 FG%, 66.7 3G%, 90.9 F%
Other Games - 18.3 ppg, 3.6 rpg, 4.4 apg, 44.5 FG%, 33.3 3G%, 84.6 F%

Comets Offense
Back Ends - 75.0 ppg, 31.5 rpg, 12.5 apg, 45.9 FG%, 77.8 3G%, 72.1 FT%
Other Games - 67.5 ppg, 28.0 rpg, 13.8 apg, 44.7 FG%, 28.1 3G%, 76.7 FT%

[Admiral_Needa]

Interesting stats, however, I just can't accept that back-to-backs don't have a negative or substantially negative impact on teams or players.

It reminds me of the stat that is often put up about Pittsburgh Steeler QB Ben Roethlisberger. Specifically, that Roethlisberger has a better completion percentage or accuracy when hit, compared to when protected (throwing without any defender hitting him).

However, if you ask any QB or former QB working as a football analyst whether they can be more successful as a passer when hit or not hit, no doubt nearly all of them would say when not hit.

Similarly, the NBA analysts always say that NBA players prefer to play every other day or every 2-3 days, as opposed to back-to-backs. Considering the W has even more travel difficulties than the NBA, I think that the W players prefer the same game spacing for the same reasons.

_________________
2002 WNBA Virtual GM Overall Winner
2006 WNBA Triple Threat Overall Winner
2007 NBA ESPN Fast Break Overall Winner

[ShockPR]

Looking at the broader issue you raised of what athletes tell us about these kind of things, I would suggest that as much as those who play the games can tell us about "how" they play the game and "how they feel" when they play it, that they are far from the best people to actually analyze performance in these kind of situations. Noted baseball author and current Boston Red Sox front office big-wig Bill James called it not being able to see the forest for the trees.

I agree with you that it makes intuitive sense that the back end of a back to back should be more difficult, but how many athletes are saying it is tougher because they've honestly looked at it and can articulate what is different about the way they play in the back end, and how many are just parotting conventional wisdom that has been passed down from generatiom to generation.

I have heard basketball color commentators say on a number of occasions that three pointers get rebounded by the offensive team more often than two-pointers do. Studies I have read indicate that this is not true, but it doesn't stop them from saying it.

Getting back to the back-to-backs, it may be that we aren't giving these world-class athletes the credit they deserve for being in incredible shape and having the ability to bounce back given the rigors of playing, practicing and treveling

Perhaps when teams are winning the first game of a back-to-back by a decent margin, coaches tend to rest players more than they otherwise would to save them for that second game. Perhaps players in that first game don't run full throttle Tamika Catchings style throughtout the game taking more breaks than they otherwise would.

The thing to do is ask the question, 'if it is true that the back-end of a back-to-back is tougher on players, and that difficulty impacts the performance of players playing in those games, how would we know?'

Statistically, we would expect teams to lose more back-ends than they would if the game was a random single game on the schedule. We might expect offensive shooting percentages to decrease and defensive shooting percentages to increase. We might expect the same to be true of rebounding percentage. Perhaps teams take more threes in the back end of a back to back rather than aggressively attack the rim. Perhaps they foul more.

These are all things we can test for, if we have the inclination. If I get some time over the weekend, I'll try and put together the league averages for the past several years just to see what they look like.

Of course, you could go even deeper than this if you wanted. Do we treat a game where both teams are playing the back end of a back to back differently from a game where only one team is. Are we adjusting for the quality of the opposition in those back-end games. I doubt I'll have the time to do that, but we'll see what teams and players over time have been better than others in the these situations.

Your earlier point is true though, just because we don't find any evidence of an impact, doesn't mean their isn't an impact. It may just mean we haven't found a way to measure it yet.

[threadkiller1201]

here's a link to Kevin Pelton's analysis (from Storm web page last year):

http://www.wnba.com/storm/news/back2back050620.html