NCAA gymnastics teams travel a median of 6,700 miles each season. This is the equivalent of a round trip between Anchorage and New York City. We found that different conferences had travel schedules that looked very different from one another, and even within a conference, some teams rack up thousands of more miles on the road than others do. With gymnasts spending hours on the way to meets, we wanted to know if these tough travel schedules affect scores. Does traveling farther result in lower scores?
Methodology
To find out, we pulled scores for meets in 2022, 2023, and 2024 and classified each meet as home, away, or neutral-site. We excluded neutral site meets for this analysis. For each away meet, we calculated the as-the-crow-flies distance in miles between the host campus and the away teams’ campuses. Home meets were assigned a distance value of zero miles. These distances are approximate because some teams may have multi-meet weekends that involve traveling from one away meet to another away meet nearby, rather than starting all travel from their home campus.
For each event in each meet, we calculated the mean score deviation. This measures how much a team’s mean, or average, score on an event differs from the team’s mean score for the entire season. For example, Arkansas averaged a 9.90 on beam at Missouri in 2024. Its season average on beam was 9.79. Subtracting the season average from the meet average, we get a mean score deviation of +0.11.
We also calculated a hit rate for each event and the deviation from the season average hit rate for that team. Hit rates were defined as scores greater than or equal to 9.75. Using the same example above, Arkansas’ hit rate on beam that day was 1.0, meaning that all gymnasts scored a 9.75 or higher. Arkansas’ average beam hit rate in 2024 was 0.79, making its hit rate difference on beam for this meet +0.21.
Overall Results
To begin this analysis, we looked broadly for a relationship between away scores and distance traveled. In this chart, we looked at mean score deviation by distance. Mean score deviations are plotted along the vertical axis, and distance traveled is plotted along the horizontal axis. The blue dashed line is a trend line, which shows a very slight negative relationship between mean score deviation and distance traveled. This means that as distance increases, scores tend to be a bit lower than the season average.
Next, we plotted hit rate differences by distance traveled. The trend line is trending downwards again, meaning that as distance increases, hit rates tend to be slightly lower than the season average.
Weekly Impact
From the trend lines above, we’ve learned that scores and hit rates trend slightly lower as distance grows. Since we know that scores tend to build as the season goes on, we wanted to test this relationship for each week of the season.
To do this, we calculated correlation coefficients which tell you the strength and the direction of a linear relationship. A correlation coefficient closer to one or negative one indicates a strong linear relationship, while a correlation coefficient closer to zero indicates a weak linear relationship. In this example, a positive correlation coefficient means that as distance increases, scores increase too. A negative correlation coefficient means that as distance increases, scores decrease.
We found that at the beginning of the season, there was a positive relationship between distance traveled and mean score deviations, meaning teams scored higher the farther they traveled. This changed after week four. Later in the season, as distance traveled increased, scores tended to be lower than the season average.
We wanted to know if this pattern differed by event. The pattern was repeated across all four events. In the beginning of the season, most events had a positive relationship between distance traveled and mean score deviation. As the season drew on, meets with longer travel tended to have below-average scores.
Hit rate difference and distance traveled had a negative correlation throughout the season, with exceptions in week one, week four, and week eight. Across the board, these correlations were weak; all were within +/- .2.
Looking at each event separately, we noticed two distinct patterns. On uneven bars and floor, the relationship between hit rate difference and distance traveled did not change much from the beginning of the season to the end. On the other hand, the relationship between hit rate differences and distance traveled trended downwards throughout the season on vault and beam. As the season went on, longer distances were associated with lower hit rate differences.
Grouping By Distance
For the next portion of this analysis, we divided home and away meets into five categories: home meets (no travel), 1 to 99 miles of travel, 100 to 499 miles, 500 to 999 miles, and 1,000 or more miles.
Within each category, we calculated the percent of meets by the number of gymnasts who hit their routines. Hits were defined as scoring a 9.75 or higher. Across all distances, it was most common for 3, 4, or 5 athletes to hit their routes. During home meets, it was slightly more common for 5 or 6 athletes to hit than for meets in any other distance group.
Next, we examined hit rate differentials by distance category. We calculated the percentage of event hit rates that were lower and higher than the team’s season average hit rate. Events with hit rates within 10% of the season average were considered similar to the average and are excluded from the chart below. During home meets, about 20% of event hit rates were lower than season averages, and about 30% of hit rates were higher than season averages. This relationship was inversed for meets that were 1,000 miles away or farther: about 35% of event hit rates were lower than season averages, and about 20% of meets had higher event hit rates, implying the large distance traveled does have an impact on the team hit rates.
Lastly, we looked at mean score deviations. In the chart below, we considered deviations that were within one-tenth of the season average to be similar enough to the season average to exclude. At home meets, a higher percentage of team event scores were higher than their season average than were lower. For all away meets, team event scores tended to be lower than season averages more often than higher. This remained true across all distance categories. However, for meets 500 miles away or farther, the gap widened between meets with scores lower than the mean and meets with higher scores, once again showing the impact of the distance traveled.
All in all, there does seem to be an association between longer distance traveled and lower scores. However, this relationship is not statistically strong. We noticed that as distance traveled increased, scores and hit rates decreased, especially later in the season. Hit rates on vault and beam at late-season meets had the most pronounced negative correlation. Ultimately, it seems that while traveling longer distances has a slightly negative effect on performance, many other factors determine a team’s scores.
READ THIS NEXT: A Day in the Life of a College Gymnast: Home Meets Versus Away Meets
Article by Jill Walsh and Dara Tan
