How to Predict Your Marathon Time and Why Formulas Run Fast

2026-10-24

The simplest way to predict marathon time is to take a recent race and scale it with Peter Riegel's formula: predicted time = known time × (42.195 ÷ known distance)^1.06, with distances in kilometres. A 10K in 45:00 predicts a 3:27:01 marathon, and a half marathon in 1:45:00 predicts 3:38:55. Treat those numbers as a best case. For most recreational runners the real finish is slower, and the research says by how much.

Most guides stop at the formula or a single multiplier. Below is what the formula assumes, where it was never meant to work, and how to turn a prediction into a goal you can actually hold for 42 kilometres.

Start with pace: time divided by distance

Before predicting anything, it helps to know your pace. Pace is time divided by distance: 50:00 over 10 km is 5:00 per kilometre. Per-mile pace uses the international mile, exactly 1.609344 km, so the same run is 8:03 per mile. Speed is the reverse, distance over time in hours, here 12.0 km/h.

Pace is written in minutes and seconds, so round to the nearest second before you split it up. 10 km in 49:56 is 299.6 seconds per kilometre, which shows as 5:00. A careless split gives the impossible 4:60.

It also helps to know what a goal time means per mile. These are the average paces over the official 42.195 km:

• 3:00:00 marathon: 4:16 per km, 6:52 per mile.
• 3:30:00: 4:59 per km, 8:01 per mile.
• 4:00:00: 5:41 per km, 9:09 per mile.
• 4:30:00: 6:24 per km, 10:18 per mile.
• 5:00:00: 7:07 per km, 11:27 per mile.

Use the official distance, not what your watch says. GPS usually measures a race slightly long, so pace based on the watch's distance looks faster than the pace the clock records. If you train on a treadmill set in miles per hour, 6.0 mph is a 10:00 mile, 6:13 per kilometre and 9.7 km/h.

How to predict marathon time from a shorter race

Riegel published the formula in Runner's World in 1977 and expanded it in American Scientist in 1981. The exponent of 1.06 is the key part. Because it's above 1, your predicted pace slows as the distance grows: doubling the distance multiplies the time by about 2.08, not 2.

Here's what the formula gives for a few published examples, all with the default exponent:

• 5K in 20:00 predicts a 10K in 41:42.
• 10K in 50:00 predicts a half marathon in 1:50:19 and a marathon of about 3:50.
• 10K in 45:00 predicts a 3:27:01 marathon.
• Half marathon in 1:45:00 predicts a 3:38:55 marathon.
• 5K in 25:00 predicts a 3:59:47 marathon.

The input matters as much as the formula. Use a race you ran all-out, recently, on a reasonably flat course. A training run is not a race effort, and the formula only scales whatever effort you give it.

Why the marathon prediction is usually too fast

The formula assumes you're as well prepared for the marathon as you were for the race you entered. Riegel built it from top performances, and for most recreational runners that assumption doesn't hold.

The clearest test is a 2016 study by Andrew Vickers and Emily Vertosick in BMC Sports Science, Medicine and Rehabilitation. They compared Riegel's predictions with the real results of 2,303 recreational runners. Up to the half marathon, the formula held up well. For the marathon it didn't: the median runner finished 10:09 slower than predicted, and for half of the runners the prediction was at least 10 minutes too fast.

What made the difference was training volume. In the same data, running 50 rather than 30 miles a week was associated with a marathon about 25:32 faster for men. A 10K result can't see that. When the authors added weekly mileage to their model, the share of runners predicted 10 or more minutes too fast fell to about a quarter.

Other analyses point the same way. Runner's World South Africa cites a statistical model that multiplies a half marathon time by 2.19 to get the marathon, instead of about 2.08 from the classic formula, and its columnist still calls that slightly fast for many runners. He also names the real cost of an optimistic number: it tempts you into a start you can't sustain, and the price comes in the last few kilometres.

The 230-minute limit most calculators ignore

Riegel fitted the formula to efforts lasting roughly 3.5 to 230 minutes. Outside that window it's extrapolating, not describing data.

That matters for a lot of marathoners. Take the 5K in 25:00 from the list above. Its predicted 3:59:47 is beyond 230 minutes, so the formula is being stretched twice at once: from a short race to a long one, and past the range it was built on. Our calculator marks that prediction as less reliable. The 10K in 45:00, with its 3:27:01, stays inside the window and isn't marked.

The closer the two distances, the better the estimate. A 10K predicting a half marathon is a modest stretch, and the 2016 study found the formula well calibrated there. A 5K predicting a marathon stretches 5 km into 42.195 km, and every bit of error in the input grows with it.

How to add a realistic margin

You can't fix the formula's blind spots, but you can stop treating its answer as a promise. A few practical options:

• Predict from a half marathon if you've run one. It's the closest distance to the marathon and the one the research trusts most.
• Raise the exponent. 1.06 is the most quoted value. Vickers and Vertosick used 1.07 as an average, Riegel himself reported 1.08 for elite runners, and RunningAhead calls 1.06 to 1.10 typical. One data scientist's fit to amateur marathon results came out near 1.16.
• If your weekly mileage is low or your long runs are short, use the upper end of that range, or read the default prediction as the fastest you could hope for.
• If you've raced two distances recently, the exponent that connects them is your personal one. It captures whether you're more of a speed or an endurance runner, which no fixed number can.

Changing the exponent is a quick sensitivity check. At 1.00 every prediction keeps your current pace, and the gap between that and 1.06 or 1.10 shows how much fatigue the formula is adding.

What to calculate next

The running pace calculator below does all of this from one race. Enter the distance and your finish time, and you get your average pace per kilometre and per mile, your speed, and predicted times for 5K, 10K, half marathon and marathon, each with its own pace. Predictions outside the 3.5 to 230 minute window are marked. You can change the exponent in the extra settings to see how far the marathon estimate shifts.

What it doesn't account for: hills, heat, altitude, age grading or heart-rate zones. Use the result as a starting point for a goal, then check it against your long runs before race day.