Earlier this fall, Lee Feinswog at VolleyballMag suggested that this was the season of reverse sweeps in college volleyball. It does seem that every day there’s another reverse sweep or two. Last night (Wednesday, November 28th, 2024) there were three reverse sweeps in forty-four D1 women’s volleyball matches: Stanford came back from losing the first two sets to Georgia Tech, Purdue reverse swept Oregon, and Northern Colorado avoided being upset in the first round of the Big Sky tournament by reverse sweeping Montana State. Wisconsin avoided being reverse swept by Ohio State by pulling out the fifth set after dropping sets three and four. In the first three rounds of the Division 3 women’s volleyball tournament (56 games) last week, however, there was just one reverse sweep.
In my high school season this year, my team was both on the receiving and giving end of reverse sweeps, which led me to wonder how common reverse sweeps actually are. I looked around to see if anyone had figured out how common they were, but the internet failed me. Then I realized I have a huge dataset with the exact answers I’m looking for, so here we are! Last weekend, I scraped data from the NCAA Stats website for more than 10,000 D1 and D3 volleyball games from the 2024-25 season.
As can be seen in the figure above, just over half of matches (52.6%) this season were won in three sets. Nearly thirty percent (29.2%) are won in four sets, and just under one-fifth of matches (18.2%) took the full five sets.
By definition, three set matches all follow a single set pattern: all of the sets are won by the team that wins the match (WWW). There are three patterns for four set matches (LWWW, WLWW, WWLW) and six patterns for five set matches (LLWWW, WWLLW, WLWLW, LWLWW, WLLWW, LWWLW).
Intuitively one might expect that certain patterns are more common than others. For example, in the four set matches, one might imagine that the LWWW might be most common—this is the pattern in which the winning team comes out cold in the first set, but then ‘wakes up’ and wins the next three. Alternatively, perhaps some teams that go up 2-0 might let their foot up off the gas and drop the third set before buckling down and coming back to win in 4 sets. However, the data doesn’t suggest that any particular version of four set matches are much more common than the others: in my data LWWW, WLWW, and WWLW are 33%, 34.2% and 32.8% of four-set matches.
Similarly, all of the patterns of five-set matches are almost equally likely to happen. Reverse sweeps happen in 16.7% of five-set matches. ‘Almost’ reverse sweeps–games in which a team wins sets one and two, drops sets three and four, but wins set five–happen as often (16.6%). The full set of distributions of volleyball set outcomes can be seen in the figure below.
Overall, there does not seem to be strong evidence of momentum affecting the overall patterns of volleyball match outcomes, at least not momentum based on set outcomes, but it is still really interesting to see how often reverse sweeps actually happen!
Data Notes:
- The data reported above are from 10,680 D1 and D3 box scores from the Fall 2024 season played from the beginning of the season through November 22nd.
- To explicitly answer the question in the title: both reverse sweeps (LLWWW) and almost reverse sweeps (WWLLW) are both 3.0% of the 2024-25 D1 and D3 regular season matches.
- While this article has focused on the patterns of sets won and lost, one can also look at the same data in different ways. For example, in this data, 81.2% of teams that win the first set win the match, and teams that go up 2-0 won 95.5% of the time.
- Another bit of evidence that momentum may not play as big a role in the aggregate as one might expect: when teams split the first two sets, there is almost no difference in the probability of winning the match based on the order in which sets were won. Teams that won the first set went on to win the match 50.4% of the time, teams that won the second set went on to win the match 49.6% of the time.
- Whether a team won or lost a set may be less important than how (and by how much) they won by. My data includes more than just the set win-loss records–it has the scores and play by play describing how points were won, and I hope to have time to analyze and write more based on this data in the coming months.
- Photo Credit: American University Athletics