Does the 1.06 race-time prediction rule hold? Tested on today's world records

Overview

Testing Riegel’s 1.06 endurance formula on the current world records shows that this classic race-time prediction rule runs a little too slow for the fastest runners, and even more so for women. The best-fit exponent in the formula is lower than 1.06, particularly over the longer distances amateurs care about most. Here’s how the numbers shake out, what the exponent really does to your own predictions, and what it can—and cannot—tell you.

Riegel’s 1.06: What it predicts, what the records do

Peter Riegel’s formula, written as T₂ = T₁ × (D₂/D₁)^1.06, is used in most race-time calculators today[2]. It assumes that the time needed for a new race distance can be predicted from a previous race by multiplying by the ratio of distances raised to the 1.06 power.

When the site's owner tested this rule on the current world records for men and women—from 1500 m to the marathon—they found that the best-fit exponent was not 1.06 but lower: 1.0546 for men (using records from 5000 m to marathon) and 1.0451 for women[1]. In other words, the actual record-holders slow down less with distance than the rule expects.

Here’s what the data shows for the records as of 30 October 2026[1]:

When predicting the marathon from the half marathon, 1.06 gets the men’s record within 2.2 seconds, but from the 10,000 m or 5000 m, the prediction is 1–3% too slow for both men and women[1]. For women, the over-prediction is even larger.

How much 1.06 over-predicts: Actual race-by-race differences

When you use 1.06 to predict a marathon from world-record times at shorter distances, the error is clear in seconds and percentage. Here are the results from the owner’s test[1]:

Men:

• Half marathon to marathon: predicts 1:59:32.16; actual record 1:59:30 (2.2 sec too slow)
• 10,000 m to marathon: predicts 2:00:26.91; 56.9 sec (0.79%) too slow
• 5000 m to marathon: predicts 2:00:44.73; 74.7 sec (1.04%) too slow

Women:

• Half marathon to marathon: predicts 2:11:04.36; actual record 2:09:56 (68.4 sec, 0.88% too slow)
• 10,000 m to marathon: predicts 2:12:57.39; 181.4 sec (2.33%) too slow
• 5000 m to marathon: predicts 2:13:57.92; 241.9 sec (3.10%) too slow

In both cases, the 1.06 exponent is most accurate when used from the half marathon to the marathon, but increasingly overestimates the time needed when starting from shorter distances, and more so for women[1].

The exponent is not constant: How the fit changes with distance

A key finding is that the best-fit exponent changes with the pair of distances you use. For men, the exponent from 1500 m to 5000 m is 1.0792, but from 10,000 m to marathon it is 1.0545. For women, the same distances give 1.0787 and 1.0440, respectively[1]. This means Riegel’s formula is more accurate for predicting times over longer distances (5000 m to marathon) than when comparing shorter to longer races.

Here’s a summary table from the test[1]:

The lesson: the “1.06” rule is a simplification. For world-record performances, the slowdown with distance is less than 1.06, especially for women, and the fit varies by which distances you use[1].

A worked example: What the exponent does for a recreational runner

To illustrate how the exponent shifts marathon predictions, suppose a runner clocks 50:00 for 10K. Using the same formula as in our own 2026 test of Riegel’s rule[1], here’s what the predicted marathon looks like as you adjust the exponent:

Each 0.01 step up in the exponent adds about seven minutes to the predicted marathon in this example[1]. If you’ve tried a marathon based on a 10K time, you may have noticed that 1.06 can set a pace that feels ambitious once you’re past 30 kilometers—especially if your training leans less on endurance than the elites that set the records.

No source gives the best-fit exponent for recreational runners; this worked example is for illustration only, not a data-backed answer for non-elites. The only measured exponents in our data are from world records, where the best fit is lower than 1.06[1].

What these results can—and cannot—tell you

This is a direct test on world records as of 30 October 2026: Riegel’s 1.06 formula slightly overestimates marathon times when you predict from shorter records, and the gap is wider for women. The best-fit exponent is about 1.055 for men and 1.045 for women from 5000 m up to the marathon, based on our calculation[1]. For example, using the half marathon record as a base, Riegel’s 1.06 predicts the men’s marathon to within 2.2 seconds of the actual world record—nearly a perfect match. But from the 5000 m or 10,000 m, the same formula runs 0.8% to 1% slow for men, and 2–3% slow for women[1].

But this only tells you how the current world records line up. Each of these records is held by a different athlete, set in a different year, on different tracks or courses, under different conditions—not the same person moving up in distance, and not a model for what would happen if you, or any single runner, doubled their race. For anyone outside this elite tier, how much marathon pace falls depends mostly on your training volume for the marathon itself. The test does not—and cannot—control for that[1].

For women’s marathon and half marathon records, both were set in mixed races with male pacemakers[1, 3]. No source gives what the exponent would be in women-only races, and our dataset does not separate out this effect; what the records show is that the implied exponent is lower for women, but the role of pacing setups is not measured here[1].

Questions readers ask

Does this mean I should use a lower exponent for my own predictions?

Not necessarily. These exponents fit the world records, not recreational runners. Most non-elite runners slow down more than these record-holders do as the distance increases, so the best exponent for you may be higher—sometimes 1.07 to 1.10 or more, especially if your long-run training is limited[1].

Why does the exponent change with distance?

The exponent reflects how much extra time is needed as you run farther. Elite runners are relatively better at longer distances, so the slowdown shrinks with distance. Over shorter intervals (like 1500 m to 5000 m), the drop-off is steeper, but from 5000 m up, the difference is less pronounced[1].

Can I trust any race-time calculator?

Race calculators are best when you input a recent, all-out race at a similar distance to your goal. If the gap is large (like 5K to marathon), the predicted time gets less reliable, and your actual marathon fitness and training have the final say[1].

Sources

1. The site owner's own test data (2026-10-30): Riegel's 1.06 tested on world records: our calculation, 2026-10-30
2. Peter Riegel - Wikipedia
3. List of world records in athletics - Wikipedia