An Analysis of Reaction Times

This article was originally published in the 2019 Spring Issue of SwimSwam Magazine. With so much discussion on reaction times throughout the broadcast of International Swimming League (ISL) matches this season, we thought it’d be valuable to offer our readers a statistical look at the true impact reaction times have on a race.

Every swimming race starts the same way: with the sound of a horn.

Swimmers hear that and begin the race. However, they don’t all leave at the same time. Some react quicker than others. Some get off the blocks faster. We have a stat for this: reaction time.

At high-level meets, there is a pad on the block that measures how long it takes for a swimmer’s feet to leave it.

That reaction-time number isn’t just a measure of how quickly swimmers register the horn and begin moving, but of their start mechanics. A swimmer who does an old-school two-foot start will take longer to get off the blocks than one who does a modern track start.

So let’s dig into some data and see what we can learn about starts. I grabbed times data from every World Championships meet, long course and short course, since 2011.

In events off the blocks (no backstroke starts), swimmers had a median reaction time of 0.7 seconds, with a standard deviation of 0.063 seconds (19,215 data points).

Men were a tiny bit faster than women off the blocks. Men had a median reaction time of 0.69 seconds and standard deviation of 0.059 seconds (10,573 data points), and women had a median reaction time of 0.72 seconds with a standard deviation of 0.062 seconds (8,459 data points). Those small differences are statistically significant, p < 0.0001.

Sprinters were quicker off the blocks than distance swimmers. The median reaction time for women in 50s and 100s was 0.7 seconds, and reaction times got progressively slower as the distance went up. Each higher increment of distance saw swimmer getting off the blocks about 0.02 seconds slower: 200s swimmers took a median of 0.72 seconds to get off the blocks, 400s swimmers took 0.73 seconds, 800s swimmers took 0.76 seconds, and 1500s swimmers took 0.78 seconds.

The men saw a similar pattern: 50s swimmers had a median reaction time of 0.65 seconds, 100s swimmers took 0.68 seconds, 200s swimmers took 0.69 seconds, 400s swimmers took 0.71 seconds, 800s swimmers took 0.75 seconds, and 1500s swimmers took 0.74 seconds.

There are a lot of swimmers at Worlds that FINA lets swim despite their lack of a qualifying time — they’re called “universality selections.” They’re from smaller countries, and the goal is to encourage growth of the sport in those places. It’s a nice program, but the presence of those swimmers skews the data a bit, with their slower starts pulling down the averages. Removing all prelims swims from the data removes their reaction times, lowering the overall medians by a hundredth or two of a second at every distance. Table 1 has all the details.

Table 1: Reaction times by distance at World Championships since 2011

50 100 200 400 800 1500
Women, all swims Median 0.7 0.71 0.73 0.74 0.76 0.78
  Average 0.707 0.714 0.728 0.739 0.759 0.784
  Standard deviation 0.067 0.058 0.053 0.059 0.058 0.060
  Sample size 2,470 2,743 2,085 1,589 328 129
Women, no prelims Median 0.68 0.7 0.72 0.72 0.75 0.77
  Average 0.679 0.704 0.721 0.726 0.745 0.771
  Standard deviation 0.043 0.047 0.048 0.046 0.054 0.052
  Sample size 603 748 451 129 115 33
Men, all swims Median 0.67 0.68 0.69 0.71 0.75 0.74
Average 0.683 0.685 0.697 0.711 0.75 0.74
Standard deviation 0.066 0.054 0.052 0.057 0.048 0.0512
Sample size 3,157 3,453 2,508 1,840 189 351
Men, no prelims Median 0.65 0.67 0.69 0.7 0.75 0.74
Average 0.654 0.671 0.693 0.702 0.758 0.741
Standard deviation 0.04 0.041 0.048 0.059 0.042 0.053
Sample size 617 739 449 129 33 141

Swimmers in shorter events were better off the blocks than swimmers in longer events. This makes quite a bit of sense: The shorter the race, the more the tiny improvements to start mechanics matter to a swimmer’s final time. Also, sprint-oriented athletes tend to be able to generate more power than distance athletes, allowing them to get off the block quicker.

Elite swimmers are pretty consistent with their starts. The average change in a swimmer’s reaction time from one round to the next (prelims versus semifinals, semifinals versus finals, same swimmer, same event) was only 0.02 seconds (standard deviation 0.02).

A good or bad reaction time has only a tiny correlation with how a swimmer’s time changes from one round to the next. Changes in reaction times accounted for less than 2 percent in the variability of a swimmer’s first 50 from one round to next. The same was true in individual 50s (the 50 free, the 50 breast, and the 50 fly). The change in a swimmer’s reaction time in 50s accounted for about 2 percent of the swimmer’s time change from prelims to semis or semis to finals.

The observation that a swimmer typically gets off the block faster than their competition is useful, but the observation that they got off the block 0.05 seconds faster in finals than they did in prelims is borderline useless.

How long it takes a swimmer to leave the block after the starting horn isn’t the only reaction time reported on most meet results — there are also relay reaction times.

Men had an average relay reaction time at Worlds of 0.38 seconds, while women had an average reaction time of 0.41 seconds.

Swimmers tightened up their exchanges in later rounds of relays. In finals, men had an average reaction time of 0.36 seconds, and women had an average reaction time of 0.40 seconds. I’ll leave it to someone else to psychoanalyze why men appear to be more aggressive off the blocks on relay starts, even by a small amount, at the highest level of the sport.

Swimmers were also more aggressive on relay starts in the shorter relays. Men had an average reaction time of 0.33 seconds in the finals of 200 relays (contested only at Short Course Worlds), 0.36 seconds in 400 relays, and 0.38 seconds in 800 relays. Similarly, women had an average reaction time of 0.37 seconds in the finals of 200 relays, 0.39 seconds in 400 relays, and 0.44 seconds in 800 relays.

In the case of relays, we can go deeper. Because the relays also exist as individual events, there’s something to compare with. To do that, I took swimmers in the individual 100 free and compared their times with non-lead-off relay splits at the same meet. I used the best individual time and relay split that a swimmer put up at a meet.

Swimmers who swam the individual 100 had faster reaction times on 400 relays than swimmers who did not. Men had an average 400 relay non-lead-off reaction time of 0.24 seconds, and women averaged 0.30 seconds. Each group was about a tenth of a second better than the overall average.

Men’s times on relay splits were on average 0.61 seconds faster than their individual swims. Their reaction times were on average 0.44 seconds faster, leaving 0.17 seconds of improvement down to improved momentum from a relay start and whatever energy swimmers got from the relay atmosphere.

Women saw a smaller relay bump: They improved their times by an average of 0.44 seconds, but of that, 0.41 was from improved reaction times, meaning they improved by only 0.03 seconds from the extra start momentum and relay intangibles.

This is an odd result. Why would women benefit much less from a relay start than men? To see if it’s repeatable or the result of some other unknown confounding factor, I gathered more data, from the NCAA Championships over the past four years.

Men in the individual 100 free had an average reaction time of 0.67 seconds, and women had an average reaction time of 0.69 seconds — totals similar to what we saw in the World Championships data.

In the NCAA data, swimmers were more conservative in their reaction times than at Worlds. The men had an average relay-start reaction time of 0.51 seconds, and men who also did the individual 100 had an average reaction time of 0.40 seconds, about a tenth of a second worse than Worlds swimmers.

Women saw the same trend. They had an average relay-start reaction time of 0.50 seconds, and swimmers who also did the individual 100 had an average reaction time of 0.39 seconds.

Despite the worse relay reaction times, NCAA women saw the same improvement in their relay splits versus individual events as Worlds women. NCAA women’s 100 free relay splits were an average of 0.46 seconds faster than their individual events, and 0.30 seconds of that was from improved reaction time.

Men’s relay times were 0.44 seconds faster than their individual swims, and 0.28 of that came from reaction time.

Unlike in the Worlds data, NCAA men and women got very similar benefits from relay starts. I’m not sure what the cause of the differences in the data is. I assume the reaction-time gap between the relay starts and flat starts couldn’t be caused by an equipment difference, because the reaction times in the individual events were almost identical.

The bottom line seems to be that the extra momentum from the arm swing and step on a relay start is worth about 0.15 to 0.17 seconds, but the biggest reason by far that relay splits are better than individual swims is the improved reaction time off the blocks.

Leave a Reply

Notify of

oldest most voted
Inline Feedbacks
View all comments
2 years ago

One other thing to note other than whether it is a old/modern start is which leg the swimmer rests his/her weight on. For example, someone like Manadou or Peaty rests their weight on their front leg to react quicker, while Dressel likes to sit back on his back leg and spring forward, and the former method would result in a slightly faster reaction time. Could it be that male swimmers do the former more often?

Thomas Selig
2 years ago

Thanks for an interesting article.

I’m a bit surprised relay reaction times are so pedestrian on average. I get that in some circumstances, particularly heat swims, it may not be worth the risk of a tighter takeover. But I think 0.3 secs is usually considered fairly “safe”, and the averages are quite a bit higher than that. I wonder if there’s a discrepancy between heats and finals?

I also sometimes wonder if teams practice relay takeovers enough. A team averaging 0.1 secs on takeovers – something that feels achievable with enough practice – gains nearly a second to the average takeover time, which is really quite a lot. In athletics (track and field), relay hand-overs are practised thoroughly. Obviously there… Read more »

2 years ago

Thank you Andrew Mering for a really well done research report. This had to have taken some time, and pointing out variants like the the universality selection cohort was well done. I would like to read some other nerdy numbers research reports like this but I would have no clue about how deep the data would be on things like stroke count or breath counts, etc.?

2 years ago

Why do I feel like this article is directed at rowdy

Reply to  Yabo
2 years ago

What are you talking about, it never once mentions the sides swimmers breath to

2 years ago

Quick, hide this from Rowdy, he’ll overdose on reaction times

2 years ago

I, for one, would love it if we could something like a 10m or 15m split. That’s a much better proxy for start performance than reaction time.

Would be especially cool to look at for some of the more dramatic differences we see in this sport… like Dressel vs Peaty in the 50m breast. Peaty got off the blocks faster (0.56 vs 0.62), but Dressel completely destroyed him to to the 15m mark. But the only numbers we have right now are reaction time and final time (25.98 vs 26.15). This race especially would be wild since even if we had 25m splits, they turned about even too.

CJ Cunningham
Reply to  Barry
2 years ago

The old Race Analysis software had 15m velocity split.

Reply to  Barry
2 years ago

But rowdy will sit there and talk about how peaty had a better start while he chases Dressel from behind

Reply to  Barry
2 years ago

I like the idea of the 15m split time, but that measurement would actually be combinations of reaction time, entry velocity, and underwater streamlining. The 25m split time would be super exciting and I would like to see that!

Honest Observer
2 years ago

Well done. A very polite, scholarly nudge to our most prominent announcer to focus on something else for the first lap of these short course races. (BTW, I think Rowdy takes more grief than he deserves — all of us would make occasional mistakes if it were our job to fill up dead air — but he does pay way too much attention to what is only an extremely minor component differentiating the swimmers.)

Reply to  Honest Observer
2 years ago

Rowdy has improved slightly the last few weeks. But he deserves criticism . Doesn’t do his homework. You should see Kurt Hanson’s instagram ( he’s announcing for the teams over there). You can see how much data analysis and homework he does.

Reply to  Joel
2 years ago

I wonder why the felt the need to inflict Rowdy on us when they had Hanson available? I guess they were trying to pander to the American market?

2 years ago

The nerd in me enjoys this and also knows how much time this data took to collect!