The 700-millisecond question: what 110,000 starts reveal about reaction time

Picture a 50m freestyle final. Two swimmers hit the wall 0.3 seconds apart — about the time it takes to blink. On poolside it feels obvious that the start must have mattered. But how often is that actually true? And is a quick reaction something you can coach, or is it just something some kids are born with?

Reaction-time talk in swimming tends to run on anecdotes: one athlete, one meet, one slow start that cost a medal. We wanted to see what it looks like at population scale, across Irish competitive swimming, so we worked through every usable reaction time in the MyAthlete database and kept only the comparisons we trusted.

What we measured, and what we didn't

The raw file holds 111,330 reaction-time rows from 66 competitions between February 2018 and July 2026. We kept starts between 0.10 and 2.00 seconds where a race time was also recorded, which leaves the 110,883 figures quoted through this piece. The 0.10 floor is there to screen out pad errors and false-start artefacts. It removes 350 rows, about 0.3%, and the suspicious 0.10–0.20 band holds only 33 rows, so nothing here hangs on where exactly we drew the line.

Three things to know before any of the numbers below. First, roughly 40% of swims in the database have no reaction time at all — timing pads tend to appear at bigger, better-equipped meets, so this is a slightly elite-skewed picture. Second, about a quarter of the starts are backstroke, which begins in the water rather than from the block. We separate those wherever the claim is about blocks, and the dive-only checks confirm the main patterns hold. Third, FINA points (see the guide above) let us compare calibre across strokes and distances. Where we say "paired" or "same-swimmer," we mean we tracked the same athlete in two settings and averaged the difference, rather than comparing different groups of people.

Typical reaction: mean 0.691, median 0.690, middle half between 0.64 and 0.74. The distribution is tight — almost 80% of starts land between 0.60 and 0.80 — with thin tails (10th percentile 0.59, 90th 0.80, 95th 0.84).

Dive and back starts live in different places

The table above mixes both apparatuses, so read it as the whole population before we split it. Backstroke starts in the water, off different apparatus, so it gets its own distribution and its own standard throughout this piece — an elite sharp start, with a separate standard for backstroke. The shapes tell you why. Both are tight, both lean right, but backstroke clusters about 60 milliseconds quicker with a longer tail: dive median 0.69, back median 0.66.

Elite-level starts only (FINA 600+, 0.02s bins, 0.45–0.85s window). The backstroke sample is smaller, so read the shapes against their own zones rather than bar-for-bar across charts.

That average — and those two shapes — are the starting point, not the story. The story is when a few hundredths actually change anything — and for that, distance matters more than anything else.

0.7 seconds means something different every time the distance doubles

This was the clearest pattern in the data, and the easiest to picture. A 0.7 reaction is a fixed cost. The race around it keeps getting longer:

If you coach sprinters, that reframes the "work on your starts" conversation. It is not that distance swimmers should react slowly. It is that a sprinter pays for a sleepy start thirty times over compared with a miler.

So does the quickest reaction win?

Rarely on its own — and that is exactly what makes the exceptions worth attention. Across heats of four or more, the winner owned the best reaction in only about one heat in six (15.5 to 18.3% depending on distance). That might sound low, but consider chance: in a ten-lane heat you would expect the winner to hold the best start roughly one time in ten. The observed rate sits well above that. Reacting well tilts things toward winning without settling them — which is why the next two tests ask how often that tilt is the margin.

The sharper question is what happens when the race is already tight. For each heat we asked a simple what-if: if the runner-up had matched the winner's reaction, with everything else unchanged, would that have been enough to flip the result? It is a mechanical counterfactual, not proof of causation — the rest of the race wouldn't really stay identical — but it puts a scale on how often the blocks contain the margin.

Put faces on those percentages. In 50m freestyle the average winning margin is around half a second, and the average reaction gap between first and second is 0.076. Most finals are not decided there. But when two swimmers are already stroke for stroke, 76 milliseconds is no longer a footnote. Looked at from the other end, the best reactor in a heat wins 14.7% of the time and makes the top three 41.0%, against 9.4% and 30.3% for swimmers reacting in the bottom half. A small, persistent edge — and it points one way. Faster wins more, exactly when everything else is equal.

What if everyone started like the best starter in the heat?

The flip test above only asks about first versus second. We then asked the wider question: if every swimmer had produced the best start in their own heat, with everything else unchanged, how many places would move? It is the same mechanical counterfactual, just run for the whole heat rather than the top two — not proof of causation, but a scale for how often the blocks contain the margin.

Across 63,218 swims in trusted, fully-timed heats of four or more, 3.7% gain at least one heat place — about 1 in 27 — and 0.3% gain two or more, with a verified maximum of +5. By construction, about as many places are lost as gained: every place a slow starter takes is one a sharp starter gives up. At Summer Nationals 2019, for instance, a 50m freestyler moved 11th → 6th of 19 on a 0.19 concession (0.75 against the heat's 0.56).

That gradient is the share-of-race table arriving by independent maths: roughly 1 in 13 sprinters moves on the start, fading to 1 in 100 past 400m. Heats decide lanes; overall standings decide finals — so we re-ran the same counterfactual across the whole event. There, 24.7% of 24,032 top-30 finishes move at least one overall place, 8.5% move two or more, and 258 swims move into the top 8 with 155 into the top 3. The largest verified overall swings are +11 twice — Summer Nationals 2024, 50m freestyle 21st → 10th (27.36 off 0.68 against 0.49) and 50m backstroke 29th → 18th (31.49 off 0.84 against 0.58).

Swimmers stay anonymous here; meets and margins are real. McCullagh 2018, 50m fly: 4th → 1st overall (24.96 off 0.80 against the heat's 0.65). Summer Nationals 2018, 50m breaststroke: 7th → 3rd on just 0.09 conceded (29.88 off 0.71 against 0.62). Irish Open 2018, 50m fly 6th → 3rd; McCullagh 2024, 100m breaststroke 6th → 3rd; McCullagh 2026, 50m freestyle 4th → 1st of heat on 0.09. Every case comes from a trusted, date-keyed heat inside the 0.40–0.85s clean bounds (full method at the foot of the piece).

It improves a lot as swimmers grow up. It also sticks.

Older swimmers react faster in the cross-section, but that on its own proves little — older groups are also stronger, fitter, and more selected. The convincing version follows the same kids over time. With at least three swims on each side of each age boundary:

Three hundredths in the early teens, another two hundredths through the late teens — real, coachable margins, visible in dive starts alone and in 50m freestyle on its own. This is not just the paddock getting faster because the slow kids quit, though some of that selection is always present. Individuals are genuinely getting off the blocks quicker.

And yet early quickness tends to stay with you. Among 220 swimmers with enough races both at 16-and-under and at 19-plus, early and late averages correlate at 0.66 — high for something measured in hundredths. Four out of five kids from the fastest early quartile were still in the faster half as seniors, while fewer than one in five from the slowest quartile had climbed into the top half.

Both things are true at once, which is why starting sharp is worth coaching early: the age curves leave no doubt that reaction time trains, yet early quickness behaves a bit like a trait — some swimmers start out sharp and stay sharp. For context on what "sharp" means at the top end: among swims rated 600-plus FINA, men average 0.655 and women 0.675, a 20-millisecond gap that persists even after holding calibre constant.

Where the raw averages mislead you

None of the above weakens the case for faster starts — but three of the best stories in this dataset share one cautionary shape: the headline average points one way, and the careful comparison points the other. All three are the same mistake — reading a population average as if it compared like with like.

Big meets look faster — until you compare like with like

On raw averages, nationals (0.683) look clearly sharper than regionals (0.693) and club or schools galas (0.732). Easy headline: pressure sharpens. Except nationals also gather better swimmers swimming a different mix of events. Hold calibre fixed at 550–650 FINA and the order flips — club 0.664, regional 0.668, national 0.674. The headline was mostly composition.

The test that cuts through it follows the same swimmer between meet types. There, nationals really are quicker: 12.6ms quicker than the same swimmer at club meets (740 swimmers), 4.9ms quicker than at regionals (1,024 swimmers). Finals tell the same quiet story — 2.3ms quicker than prelims at nationals in the paired comparison, with 54.7% of swimmers sharper in the final. Pressure, on this evidence, focuses rather than freezes. (The global "finals are 9ms slower" figure is a data quirk: club timed-finals get entered as "Final" and drag the average. Within nationals, finals are 6ms faster.)

Middle lanes look faster — until you check who swims in them

Centre-lane swimmers average 0.676 against 0.694 three or more lanes out. But the middle lanes are seeded with better swimmers — average 525 FINA in the centre, 461 at the edges. Restrict to swimmers rated 550–650 and the gap shrinks to 0.6708 versus 0.6737, about 3 milliseconds. The lane wasn't doing much. The seeding was.

Strokes look different — until you compare within the swimmer

Raw stroke averages invite theories about stroke-specific aggression. Track the same swimmer across long-course 50s, 100s and 200s and almost nothing remains: fly 2.2ms slower than their own freestyle, breast 1.5ms slower, IM 5.0ms slower, with fewer than half quicker outside free in any comparison. Reaction time seems to belong to the swimmer more than the stroke.

One related note for the statistically inclined: the overall correlation between reaction and FINA points is only −0.07, which looks like no relationship — yet elite 700-plus swims average 0.657 against 0.778 for novices under 150. Both are correct. The relationship is curved with a long novice tail, so a single straight-line correlation undersells it. Tier means describe it better; paired tests describe it best.

The PB that got away — and the one the start saved

Population patterns are one thing; best times are personal. So we turned the same counterfactual on the swimmer's own history. For each athlete-event we took the all-time best, then looked only at strictly slower swims after it — 5,968 athlete-events living after their peak — and asked: with a fast start, holding the rest of the swim fixed, would this have been a best? The bar is a credible one: an elite sharp start or the swimmer's own best start, whichever is fairer, so good starters are judged against their own sharpness and slower starters against a bounded elite standard.

On that bar, 621 swims by 462 athletes come back — 499 dive, 122 back — including 24 backstroke cases a flat elite standard would have missed. But the split underneath is the honest part. Decompose each miss into start versus water, and only about 1 in 9 is pure start-cost:

Starts rarely alone cost best times, but they often decide them — which is exactly why faster is worth having. The ledger also has a credit side. Run the counterfactual in reverse — with a typical start instead of the sharp one produced — and 107 best times across 103 athletes disappear, and 1,979 heat places (2.0%) go the other way — 11.7% at 50m, 6.5% at 100m, 2.8% at 200m. Costs ask "with a fast start would it have flipped?"; gains ask "with a typical start would you have lost it?"

When the start costs a qualifying time

The highest-stakes version of the same test is the selection clock. For each target meet we took the swimmer's best in the qualifying window, measured it against the time for their age band on the same course, and kept near-misses within half a second. A miss counts as rescued when an elite sharp start covers the deficit, rest of swim fixed.

Every count here is a lower bound — qualifying slots come from entry-file imports only, so several meets carry partial coverage, and the regional figure further requires the swimmer's club to have been present at the target meet. The typical shape is a close shave: a 50m breaststroke missed by 0.02 (0.62, needed 0.60), a 200m breaststroke by 0.03 (0.63), a 100m freestyle by 0.10 (0.70). And most misses are not rescuable — the median needed start across near-misses is 0.44s, which is not a coaching target but a timing artefact. Only the close ones are coachable, and those are the ones to rehearse.

Are Irish swimmers getting quicker off the blocks?

One trend we are deliberately not banking. The unadjusted data says yes. Nationals-only averages fall from 0.711 in 2018 to 0.660 in 2026, about 51 milliseconds, and the series stays intact when we hold calibre fixed at 550–650 FINA (0.703 down to 0.657). Then you look at the steps: 21 milliseconds vanish between 2024 and 2025 alone, with a partial rebound the year after. Whole federations don't gain 21 milliseconds of fast-twitch fibre over one winter.

Conclusions: faster starts win more

Reaction time is a small share of any race — 2.19% of a 50m, 0.065% of a 1500m — and that share explains when it matters. Across 63,218 swims in trusted heats, about 1 in 27 moves a place on the start alone, rising to roughly 1 in 13 at 50m and fading to 1 in 100 past 400m. Re-ranked overall, a quarter of top-30 finishes move, with 258 into the top 8 and 155 into the top 3. Close races are where a fixed cost becomes the margin: 8–10% of finishes decided by under half a second contain the winning margin on the blocks.

The same pattern reaches personal and selection outcomes. On a per-start elite bar, 621 post-peak swims come back as bests that the start decided, while 107 bests on record only exist because of a sharp start — and only about 1 in 9 of the misses is pure start-cost, the rest also involving the water. At the selection clock, 160 qualifying swims across national and regional meets turn on a credible fast start, all at the close margin; most larger misses need starts no swimmer could credibly produce.

Reaction time also carries both development and predisposition. The same swimmers improve by roughly 30 milliseconds through the mid-teens with three-quarters improving at each step, yet early and late averages correlate at 0.66 and four in five early sharp starters remain in the faster half as seniors. Context sharpens rather than freezes — the same swimmer is 12.6ms quicker at nationals than at club meets and 2.3ms quicker in finals — while lane position and stroke, once calibre and the swimmer are held fixed, shrink to a few milliseconds. Backstroke remains its own apparatus throughout, with its own sharper distribution and standard.

Reaction time isn't the race. But it is milliseconds on offer before anyone has taken a stroke — measurable, persistent, and decisive exactly when everything else is equal. That is why it matters, and roughly how much.

See what your start is doing for you

If you found this interesting, you can see the same analysis for your own swimming in MyAthlete. The Reaction Lab looks at your electronic starts and shows whether your start is helping you or costing you — split into dive and backstroke, since the two starts work differently.

• Your grade. Your recent average, rated from Elite to Costly against real top-level starts, with a simple chart showing where you sit.
• Places. Any heat places your start has lost or saved, with the exact swim and margin shown.
• Qualifying times. Any QTs your start decided — missed or made — with what you swam, what you needed, and the start that would have done it.
• Personal bests. Any bests your start decided, including whether the miss was on the blocks or in the water.

Everything links back to the swim it came from, so you can always check the detail.