Timing resolution in racing

Why do swimming records stop at hundredths when the clock can read thousandths?

The electronics in a modern pool could report your finish to a millionth of a second if asked. Yet swimming's official times stop at two decimal places, and when two swimmers tie at that resolution the sport awards two gold medals rather than looking at the next digit. Deliberately discarding information that you already possess feels like an odd thing for a results-obsessed sport to do.

So the question is not whether the clock can read further. It plainly can. The question is what a further digit would actually be telling us — and whether that thing is the swimmer.

Begin with the event that forced the issue. Munich, 1972, the men's 400 metre individual medley: Gunnar Larsson and Tim McKee were identical at hundredths, and the placing was awarded on thousandths — 4:31.981 against 4:31.983. Gold and silver separated by two thousandths of a second. The result stood, and swimming subsequently abandoned the third digit altogether. Why would you retreat after the extra digit had just done its job?

Here is the move that answers it. A time is not a measurement of the swimmer alone. It is a measurement of the swimmer plus the course they swam. So ask: how precisely is the course itself defined?

Not very. Pool construction rules permit a lane to be slightly longer than nominal — the tolerance on a 50 metre length runs to a few centimetres, with lanes allowed to be over but not under. A legal pool can therefore have lanes whose lengths differ by more than the width of your hand.

Now convert that into time, because that is the currency we care about. An elite freestyler moves at roughly two metres a second, so a centimetre of extra distance costs about five thousandths of a second. A three-centimetre discrepancy between two lanes is worth something like fifteen thousandths.

Look at what that does to Larsson and McKee. Their two-thousandth gap is several times smaller than the amount that lane geometry alone could have handed one of them. The third digit was computed correctly, arrived reliably, and was still not a fact about who swam faster. It was a fact about the building.

That is the general principle, and it is worth stating plainly: a measurement carries no more information about a quantity than the least reproducible link in its chain. Refining the clock does not refine the answer once the clock has stopped being the limiting element. Past that point extra digits are perfectly precise measurements of noise.

Then push the argument one step further, because it does not stop conveniently at hundredths. A hundredth of a second is about two centimetres of swimming — still within the range that lane variation could account for. So why keep the second digit at all? Honestly, because it is a convention rather than a derivation: hundredths are the coarsest resolution that separates most races while staying close to the limit of what the course can support. And the sport hedges the remaining doubt by refusing to break ties: when two swimmers share a time, both are recorded as sharing it, which is a governance decision admitting that the instrument has run out of meaning.

Compare a sport where the course is not the weak link. Track distances are laid out to far tighter tolerances relative to the times involved, and athletics does use thousandths from the photo finish to order athletes who share a hundredth — while still rounding published times and records to hundredths. The difference is not that one sport has better clocks. It is that in one of them the extra digit still refers to the athlete.

There is a second limiting link worth naming, since it is not only the pool. The swimmer stops the clock by pressing a touchpad that requires a real, if light, force, and the amount of that press — finger angle, reach, how flat the hand lands — varies from touch to touch by more than a thousandth of a second's worth of movement. Even a perfect pool would not rescue the third digit.

The natural misreading is that this is about equipment being imperfect — that better touchpads and better pools would eventually earn the third digit back. Some of it would improve, but the frame is wrong. The point is that precision is a property of a chain, and the chain is only as informative as its loosest link. Buying a finer clock while the pool tolerance is unchanged does not buy any additional truth; it buys the appearance of truth, which is worse, because a published thousandth invites everyone to treat it as real.

A second refinement: this is not the same as measurement error in the everyday sense. The lane difference is not a random wobble that averages out over repeated swims — it is a systematic advantage attached to a lane. Extra digits do not merely add noise; they can encode a bias faithfully and present it as an athletic result.

• How do sports with even looser courses — open-water swimming, road marathons — handle records?
• What would it take, structurally, for swimming to justify a third digit again?
• Why does athletics separate placings by thousandths while still rounding the published time?

Every term the collection defines is gathered in the glossary.