The 50-yard freestyle at Men’s NCAA Championships is the fastest, arguably most intense race in all of swimming. It’s always one that gets a lot of attention because of the pure speed, power, and athleticism required to excel in such a short, explosive race. In 2023 though, the event was the center of attention for another reason.
The officials held the athletes at the start of this race for an unusually long time, enough so that it interfered significantly with their reaction time off the blocks. This was a bit controversial at the time, and there was a ton of discussion around the start after the race.
I took a quick look at the race from a statistical perspective – and published an article on swimswam.com – to compare the starts (measured by reaction time) of this heat relative to previous performances and to measure how important/predictive the start is for the final results.
Data
The dataset we’ll use for this analysis was scraped from swimmeetresults.tech and includes all 188 of the 50-freestyle results from the 2021, 2022, and 2023 Men’s NCAA Championships. This contains both prelims and finals results, but not relay takeoffs (so, the individual event only). The data includes the meet, swimmer name, swimmer team, round (final or prelim), reaction time, time to 25 yards, and final time.
1. A Comparison of the 2023 Final vs. the Rest
Here we highlight the general distribution of reaction times for competitors in the 50 freestyle. The distribution collects primarily around about 0.63 seconds and most of the data falls within about a tenth of a second between about 0.58 and 0.67. There are certainly slower outliers than the 2023 Final, but with more than half of the athletes getting off in 0.7 or slower, it certainly seems to be unusual.
Additionally, we highlight Jordan Crooks here because he was the event winner and the athlete that reportedly caused the officials to delay the start.
Making a more direct comparison of the reaction times from the 2023 Final against the rest of the results, it definitely appears to be a significantly slower race than normal for NCAA level competition.
If we consider the null hypothesis: The reaction times from the 2023 final don’t differ from normal, we can quantify this a bit more precisely. I ran a t-test to determine the significance. The p value 0.028 suggests that we can reject the null hypothesis and there is a high probability the delay in start impacted the results. (There is only a 2% chance we could have observed these reaction times if the starts weren’t impacted by the delay).
So there is strong evidence that the unusual delay of the start did impact how fast the competitors were able to get off the block. However, each swimmer had to deal with the same delay, so that may not necessarily affect the competitive nature or fairness of the race. So, we may still wonder whether the final results were comprised at all.
2. How Important is the Start to the Final Results?
Though it will be difficult to determine just how much the final results were impacted by the starting delay, we can at least examine how the reaction times compare to the final results generally.
The overall results are displayed in the scatter plot to the right. If we model the final swim time using linear regression based on the reaction time, we observe a pretty weak correlation. Specifically, considering all swims, we have an adjusted R^2 value of 0.026.
However, considering that the 2023 final was likely impacted by the delay, it may not be fair to include this. While removing that specific race from the data and remodeling does improve the result, the r^2 of 0.075 isn’t much stronger.
So, the variation in outcomes for the swim time aren’t explained well by the variation in reaction times.
In fairness, this is what you would expect. There is so much that goes into swimming a fast 50 freestyle after the initial ~0.65s it takes to leave the block.
