Marathon race time predictor
This marathon predictor takes a recent race and estimates your marathon finish time, then adjusts it for race day conditions and the miles you run each week.Enter a recent race to estimate your marathon time, adjusted for race day and weekly miles.
- From a 5K, 10K or half
- Riegel and Cameron
- Heat, hills and wind
- Weekly miles counted
Your recent race result
Enter a race you ran in the last 8 to 12 weeks
--
Pick a distance shorter than a marathon and enter your time. Your prediction appears as you type.
Predict every distance from 1 mile to 50K
--
Enter a race time and press the button. You get a prediction for every distance, with three formulas side by side.
Check your marathon readiness
--
Enter a race result to see your predicted marathon times.
More weekly miles, faster marathon. Make the miles fun.
Your prediction gets better as your miles go up. Motera turns those miles into streets you own, XP and a spot on the local board.

You claimed a new territory!How to predict your marathon time
Take a recent race, multiply by a factor for the distance and expect to land on the slower side of the number.
3:53:10
That is what a typical 1:45 half runner ran in the marathon, in the reader survey Slate ran with Andrew Vickers. Riegel called it 3:38:55, about 14 minutes too fast.
Bars start at a shared offset so the gaps are easy to see.
From a half
Multiply by 2.085 for Riegel, or 2.19 for what runners tend to run. Double it and add 10 to 20 minutes to do it in your head.
From a 10K
Multiply by 4.60 for Riegel, or 4.67 for the 1.07 exponent.
From a 5K
Multiply by 9.59 for Riegel, or 9.80 for the 1.07 exponent. Add 5 to 10 minutes as a buffer.
The most accurate race predictor knows your mileage
Andrew Vickers and Emily Vertosick surveyed 2,303 recreational runners and tested the formulas on 156 of them who had not been used to build their models. Riegel at an exponent of 1.07, the same as the Riegel 1.07 column in All distances mode, held up to the half. At the marathon it ran at least 10 minutes too fast for half of the runners.
- Riegel, exponent 1.07
- One race plus weekly mileage
- Two races plus weekly mileage
Predicted minus real finish time, minutes and seconds. The centile ranks runners from the most overpredicted to the least. Table 4, Vickers and Vertosick 2016.
Mean squared error, lower wins
Nearly a quarter of runners finished more than 20 minutes slower than Riegel said. Even with weekly mileage added, about a quarter were still 10 or more minutes too fast, so no formula gives you a promise.
After controlling for other factors, only a prior race time and weekly mileage improved the prediction. The weekly miles inputs above exist for that reason. Their published model needs a formula from the paper's supplement, so this tool does not run it. Its mileage adjustment is a simpler planning scale of its own.
Sources: the study and Slate's write up.
Riegel, Cameron or the wall buffer?
Each formula suits a different runner. Pick the one that sounds like you, and read the others as your range.
Riegel
The standard prediction, and optimistic for marathons. Trust it if you run 40 or more miles a week and have raced a marathon before.
Best for experienced, high mileage runners on flat courses
Cameron
A fatigue curve in place of a fixed exponent. A useful second opinion when you are stepping up to the distance, especially from a 5K or 10K.
Best for a first or second marathon
Wall buffer, Riegel +6%
Readiness mode adds 6 percent to Riegel as a planning buffer for the late miles. Use it if you worry about the wall, or if your longest run is under 18 miles.
Best for under-trained runners, first marathons and hilly courses
Three adjustments, three names. Riegel 1.07 is the steeper exponent in All distances mode. The wall buffer adds 6 percent to Riegel in Readiness mode. The conditions and mileage adjustment is the percentage Marathon mode adds for heat, humidity, hills, wind and weekly miles.

The Riegel formula, worked through
T2 = T1 x (D2 / D1)^1.06
Peter Riegel published it in 1981 in American Scientist, in a paper called Athletic Records and Human Endurance. An exponent of 1.0 would be linear, with no slowdown at all. Take a 10K in 45:00, which is 2,700 seconds. The ratio is 42.195 / 10 = 4.2195, and 4.2195^1.06 = 4.60. So 2,700 x 4.60 gives a 3:27:01 marathon.
f(x) = 13.49681 - 0.000030363x + 835.7114 / x^0.7905
That is the Cameron fatigue function, built from world best times across distances, with x in meters. The prediction is T2 = T1 x (D2 / D1) x f(D1) / f(D2), so a 45:00 10K gives 3:30:50.
Tanda race predictor and watch estimates
Two ways to predict a marathon without a recent race: a published formula that reads your training log, and the predictor on your watch.
Tanda pace calculator
Pace (sec/km) = 17.1 + 140.0 x exp(-0.0053 x km per week) + 0.55 x training pace (sec/km)
Giovanni Tanda built it in 2011 on 22 runners aged 28 to 54 and 46 marathons, with a standard error of about 4 minutes. At 80 km a week and 5:30 /km it gives 4:50 /km, a 3:24:06 finish. A 2022 follow up on mostly Strava data found an error of 5.4 minutes for 2:47 to 3:36 finishers and 9.5 minutes for faster runners. First timers can add 2 to 3 percent.
Is the COROS race predictor accurate?
We found no independent study of its marathon numbers. COROS says the predictor uses recent training data to estimate 5K, 10K, half and marathon times, shifts as your fitness changes and can be calibrated with a Running Fitness Test.
The closest published test covers another brand. A 2025 study of a HUAWEI watch, with 154 amateurs and 288 tests, measured these errors, and it did not validate the marathon at all.
So check any watch against your own last race. If it says 3:30 and your half points to 3:50, believe the race.
Predict a marathon from your half
The half is the best race to predict from, because it is closest to the marathon. Find your time, then read across.

Quick math: double your half and add 10 to 20 minutes. A 1:45 half points to 3:40 to 3:50.
From a 25:00 5K, the half comes out near 1:55 on Riegel. Before you trust a marathon number from there, get 4 long runs of 10 or more miles into your last 8 weeks.
Need a half marathon pace from a half marathon time? Divide by 13.11 miles or 21.0975 km. A 2:00 half is 9:09 /mi or 5:41 /km.
For the half itself, see the half marathon calculator and half marathon splits.
Marathon prediction chart
Scan these for a first idea of your time, then enter your own race in the calculator.
From any race to a marathon
No conditions or mileage adjustment, which is what Marathon mode shows at 40 to 49 miles a week on a good day. Below 40 miles a week the tool adds 1 to 8 percent. Cameron is the published form, the same in every mode.
From a 5K to every distance
Riegel at 1.06, and it assumes comparable training for each distance. It works downward too, as a half or 5K race predictor. Running at altitude or with walk breaks? See the altitude pace calculator and the run walk calculator.
Performance level benchmarks
The All distances mode compares your time with these. Slower than Beginner reads Novice. Each cell is the slowest time for that level.
Why a marathon is harder to predict
Formulas extend your speed across the distance. The marathon adds problems a short race never tests.

Glycogen runs out
Your muscles and liver hold a limited store of glycogen, often not enough for a full marathon at race pace. When it runs low, usually in the last third of the race, the body leans on fat, which releases energy more slowly. That is the wall.
Muscle damage adds up
A marathon is about 35,000 impacts if your stride is 1.2 meters, each at 2 to 3 times your body weight. A runner can hold a 1:30 half and still fade well past what the half predicted.
Mental fatigue
A 5K lasts 15 to 30 minutes. A marathon lasts 3 to 5 hours, and the mental load grows after the second hour. Runners who have done a few tend to beat first timers with the same 10K.
More time in the weather
You sit in the weather for 20 to 35 minutes in a 10K and 3 to 5 hours in a marathon. In the Ely heat data, top finishers lost 4.5 percent on warm days, about 11 minutes over four hours, and a shorter race barely registers it.
Two runners with the same 5K can finish a marathon far apart, because a 5K says little about fueling or long run training. For where your 5K sits, see the 5K run time calculator, and for the late miles read how to avoid the bonk.
What heat, hills and wind do to your time
A prediction assumes a cool, flat day. Here is how much the real day can take back.
Top male finishers, slower than the course record
WBGT is wet bulb globe temperature, which counts humidity and sun as well as air temperature. Ely and colleagues looked at seven marathons in 2007, and slower runners lost more than these top finishers. On a four hour marathon, 4.5 percent is about 11 minutes. The heat toggle in Marathon mode adds 3 to 8 percent, which is 16 to 43 seconds per mile at a 9:00 pace. For your own pace by dew point, use the heat adjusted pace calculator.
These are the ranges in Marathon mode, set as planning guides rather than measured effects. On a 3:45 marathon, the hilly course toggle adds about 2 to 7 minutes. For pace on a grade, use the elevation adjusted pace calculator.
Turn your predicted time into a plan
How the tool works, why a prediction misses and what to do once you have a number.
Enter a recent race
Use your best result from the last 8 to 12 weeks. A 10K or half is ideal. A parkrun or a solo track effort works too.
Compare the formulas
Every mode uses Riegel and Cameron. Marathon mode averages them, All distances adds Riegel 1.07 and Readiness adds a 6 percent wall buffer for a likely range.
Set the real day
Add heat, humidity, hills, wind and your weekly miles, or see every distance from 1 mile through 50K with pace per km and mile.
Why a prediction goes wrong
How to use your number
The weekly miles behind your marathon can claim real ground.
In Motera, a run that loops back to where it started turns the area inside into your territory. Draw a loop on the map with your finger or mouse and see how much ground one training run can take.
- Zones0
- Area0 mยฒ
- Loop0.00 km
- XP+0
Demo numbers. In the app, your real GPS run draws the loop.
ยฉ OpenStreetMap, OpenFreeMap
Are marathon race predictors accurate?
Are Marathon Race Predictors accurate
MidLife Runner walks through what goes into predictor algorithms and how to use them in training, with CR Plots, Metathon and Runalyze as examples. Worth a watch before you pick a goal time.
About this marathon predictor

The Motera marathon race time predictor is a free marathon time estimator. Enter a recent race and it returns an estimated finish time and the pace to hold, so you get an expected marathon time without running one.
It has three modes. Marathon mode is a marathon race predictor with four race day conditions and a weekly volume slider. All distances mode is a race predictor for every distance from 1 mile to 50K, with a performance level. Readiness mode uses your weekly mileage to give a likely range and works in miles or km.
It suits runners picking a goal for a first or next marathon, and anyone looking for a 5K to marathon predictor or a half marathon to marathon predictor. There is no best marathon predictor for everyone: this one gives you several formulas and shows how far off they have been in published research.
The average US marathon in 2024 was 4:24 for men and 4:51 for women, according to RunRepeat, whose full data set covers 10,482,628 results from 6,352 events between 1984 and 2025. For an age adjusted view of your time, use the age grade calculator. Compare a predicted marathon with the Boston Marathon cutoff times to see whether it would have been fast enough.
Marathon predictor questions
How accurate is a marathon time predictor?
Less accurate than most runners assume. In the Vickers and Vertosick 2016 test, Riegel held up to the half marathon but predicted the marathon too fast for most runners. The accuracy section on this page charts the full results. Treat any single number as a target with a window of 10 to 20 minutes around it.
What is the most accurate race predictor?
In the largest published test, a model using one recent race plus weekly mileage beat Riegel, and two races plus weekly mileage did better still. Mean squared error was 380.7 for Riegel, 227.6 for one race plus mileage and 208.3 for two races plus mileage. So the most accurate race predictor is the one that knows how much you train, which is why the weekly miles input on this page matters.
What is the Riegel formula?
The Riegel formula, also typed as the reigel formula, is T2 = T1 x (D2/D1)^1.06. Peter Riegel published it in 1981 in American Scientist. T1 is your race time, D1 the race distance and D2 the target distance. An exponent of 1.0 would mean no slowdown as the distance grows, and 1.06 slows your pace about 4 percent each time the distance doubles. A 10K in 45:00 gives a 3:27:01 marathon.
What is the Cameron formula?
Dave Cameron built his model from the top times in the world at a range of distances. The curve is f(x) = 13.49681 - 0.000030363x + 835.7114 / x^0.7905, with x in meters, and the prediction is T2 = T1 x (D2/D1) x f(D1)/f(D2). From a 1:45 half that gives 3:43:08, a few minutes slower than Riegel. Every mode and table on this page uses this same form.
What is the Tanda race predictor?
Giovanni Tanda published a marathon formula in 2011 that uses training rather than a race: marathon pace in seconds per km = 17.1 + 140.0 x exp(-0.0053 x km per week) + 0.55 x mean training pace in seconds per km, both taken from the eight weeks before the race. It was built on 22 runners and 46 marathons, with a standard error of about 4 minutes. A 2022 follow-up on mostly Strava data found an error of 5.4 minutes for 2:47 to 3:36 finishers and 9.5 minutes below 2:47. The calculator on this page runs it.
Is the COROS race predictor accurate?
There is no independent study of it that we could find. COROS says the predictor uses your recent training data to estimate 5K, 10K, half and marathon times, shifts as your fitness changes and can be calibrated with a Running Fitness Test. The closest published test covers a different brand: a HUAWEI watch was off by -1.47, 0.10 and 2.30 percent at 5K, 10K and half across 154 amateurs, and the marathon was not validated. Check any watch against your own last race before you trust its marathon number.
How do I predict my marathon time from a half marathon?
Double the half and add 10 to 20 minutes, so a 1:45 half points to 3:40 to 3:50. Riegel multiplies the half by 2.085, which gives 3:38:55, Cameron gives 3:43:08, and the runners behind the Vickers data ran closer to 2.19 times their half, which gives 3:49:57. Choose the slower end if your weekly mileage is under 40 miles. The half is the most reliable input because it is the closest race to the marathon.
How do I predict my marathon time from a 10K?
Multiply the 10K by about 4.60 for Riegel or 4.67 for the 1.07 exponent. A 50:00 10K gives 3:50:01 or 3:53:21. These numbers assume you have done the long runs. At 50 or more miles a week, Marathon mode takes 2 percent off, about 5 minutes on a 3:50. Without weekly long runs, expect to finish slower than the number.
Can I predict a marathon from a 5K?
Yes, with caution. Riegel multiplies a 5K by 9.59, so a 20:00 5K gives 3:11:49, and Cameron gives 3:15:11. The bigger the gap between the race and the marathon, the less the prediction can be trusted, because a 5K tests speed and a marathon tests endurance and fueling. Add a 5 to 10 minute buffer to a prediction from a 5K, or use a 10K or half if you have one.
Why do the formulas give different predictions?
Each one models the slowdown with distance in its own way. Riegel uses a fixed exponent of 1.06, Riegel 1.07 in All distances mode uses a steeper one and Cameron uses a curve. Readiness mode adds a wall buffer of 6 percent on top of Riegel, and Marathon mode applies its conditions and mileage adjustment to the Riegel and Cameron average. The further your race is from a marathon, the more the formulas spread.
How does weekly mileage change a marathon prediction?
It is one of only two things that improved prediction in the Vickers study after controlling for other factors, the other being a prior race time. This tool turns that into a planning adjustment of its own, not a research figure. Marathon mode and Readiness mode use the same scale: add 8 percent under 25 miles a week, 5 percent at 25 to 29, 3 percent at 30 to 34, 1 percent at 35 to 39, nothing at 40 to 49 and take off 2 percent at 50 or more.
How much does heat slow a marathon?
A study of seven marathons found top male finishers were 1.7 percent slower than the course record at a wet bulb globe temperature of 5.1 to 10 C, and 4.5 percent slower at 20.1 to 25 C. Slower runners lost more. On a four hour marathon, 4.5 percent is about 11 minutes. The heat toggle in Marathon mode adds 3 to 8 percent.
Which race should I enter, and how recent should it be?
Use a race or an all out time trial from the last 8 to 12 weeks, and no older than 3 months. A 10K or half works best. If you have improved since, your fitness may be better than the race says. If you took time off, the prediction will be optimistic.
Should I use my training pace or race pace?
Race pace. Training paces usually sit 30 to 60 seconds per mile slower, which makes the prediction too slow. If you have not raced lately, run a solo time trial on a track or a parkrun at full effort. The one exception is the Tanda calculator, which is built to take your mean training pace.
What is the average marathon time?
RunRepeat puts the 2024 US average at 4:24 for men and 4:51 for women. Its full data set, across all years from 1984 to 2025, holds 10,482,628 US marathon results from 6,352 events. Women ran about 10 percent slower than men at the marathon in the Vickers data. For a first marathon, finishing with a pace you held to the end is a better goal than any average.
What does the performance level mean?
It compares your race time with fixed benchmark times for your gender at the 5K, 10K, half and marathon. A male 5K in 20:00 or faster reads Advanced, and in 25:00 or faster reads Intermediate. The level appears when you enter one of those four distances. The full benchmark table is on this page.
Do these formulas work for ultramarathons?
Less well. Elevation, terrain, sleep and nutrition start to matter more than the fatigue curve. Treat the 50K row as a rough estimate, and for anything longer use a model made for ultras that counts elevation and aid station stops.
What is a marathon prediction chart?
A lookup table of predicted marathon times for common 5K, 10K and half times. The chart on this page covers 5K times from 18:00 to 35:00, with the matching 10K, half and marathon on the same row. Use it to scan, then enter your own result in the calculator.
Why does a predictor show a slower marathon than I expected?
The formulas build in the natural slowdown over distance. If your training has been short and fast, that slowdown may be right. If you have done 18 to 22 mile long runs, you may beat the number by a few minutes.
How do hills change a marathon prediction?
The hills toggle in Marathon mode adds 1 to 3 percent, about 2 to 7 minutes on a 3:45 marathon. That range is a planning guide, not a measured course effect, so look at the course profile and pick the higher end when the climbs come late. For a pace on a given grade, use the elevation adjusted pace calculator.
How can I improve my predicted time?
Train for the distance. Weekly mileage and a recent race were the strongest predictors in the Vickers data, and their model put the marathon at 3:52:57 for runners who did intervals most weeks against 4:01:28 for those who did not, about 3.5 percent faster. Mix easy runs, tempo runs, intervals and long runs, and stay consistent.


