How the race time predictor works
The predictor uses the formula Pete Riegel published in 1981: T2 = T1 x (D2/D1)^1.06. Take your finish time from a recent race (T1), multiply it by the ratio of the new distance to the old distance (D2/D1), and raise that ratio to an exponent of 1.06. The exponent is the whole trick. If pace never decayed with distance, the exponent would be exactly 1 and a 45:00 10K runner would run a 3:09 marathon. Real endurance decays, so the exponent sits above 1, and every doubling of distance costs you a little over 4 percent in pace. The calculator applies the formula to every standard distance at once and shows the predicted time and pace per mile for each.
When to use it
Use it when you are picking a goal for a new distance, sanity-checking a goal you already have, or deciding which pace group to start with. It answers the question that follows every good race: what does this result mean for my next one? Run a strong 10K in October and the predictor tells you what an equivalent half marathon looks like in December, which is a far better starting point than a round number you picked because it sounded nice.
The one assumption you cannot skip: predictions assume you are trained for the target distance. The formula converts fitness across distances; it does not create endurance you have not built. A 5K result predicts your marathon only if you have done the long runs and weekly mileage a marathon requires.
Worked example: from a 45:00 10K
Say you just ran a 45:00 10K. With the standard 1.06 exponent, the predictor gives you about 1:39:18 for the half marathon and 3:27:02 for the marathon. The half marathon number is usually trustworthy: the distance jump is small and most 10K runners have the endurance to cover it. The marathon number deserves suspicion. Switch the exponent to 1.10 and the marathon prediction slows to roughly 3:41, which is a much more realistic target for a 45:00 10K runner training on modest mileage. That 14-minute spread is not a flaw in the tool; it is the honest range of outcomes, and where you land in it depends on your training.
Choosing the exponent (an honest note)
Riegel derived 1.06 from well-trained club and elite runners, and most online predictors still use it without comment. For runners on modest mileage, 1.06 flatters the marathon badly. The endurance to hold predicted pace past mile 20 comes from volume, and if the volume is not there, the last 10K collects the debt. In coaching practice, 1.08 fits most recreational runners stepping up in distance, and 1.10 is the right call for a first marathon or anyone training under roughly 40 miles per week. Predictions get less reliable the further the target is from the input race, so a half marathon predicts a marathon better than a 5K does. When in doubt, pick the slower number: a conservative goal you negative-split beats an optimistic one you walk home from.
Related tools and guides
The training pace calculator answers the same question with the Daniels/Gilbert physiological model instead of a power law: it turns your race into a fitness score, reads out equivalent times from 1500m to the marathon, and hands you training paces from the same number. The two models usually agree in the middle distances; prefer the Riegel predictor here when you want an adjustable exponent to keep a marathon goal honest on modest mileage, and the score-based tool when you are anchoring workouts to a result you trust.
Once you have a predicted time, the pace calculator converts it into pace per mile and kilometer, and the splits calculator turns it into a printable pace band with even or negative splits. And because a prediction is a promise you still have to keep, read the marathon pacing mistake that costs you the last 10K before you bank the marathon number.
Frequently asked questions
How accurate is the race time predictor?
For distances close to your input race, quite accurate: a 10K predicted from a 5K is usually within a minute or two for a trained runner. Accuracy drops as the distance gap grows. Predicting a marathon from a 5K stacks two assumptions on top of each other: that the formula holds over a fourfold jump in distance, and that you are as well trained for the marathon as you were for the 5K. Treat long-range predictions as a starting point, not a promise.
Why is my marathon prediction optimistic?
The 1.06 exponent was derived from well-trained runners with high mileage. If you run 25 to 40 miles per week, your endurance does not scale the way theirs did, and the marathon is exactly where that shows up. Switch to 1.08 or 1.10 for a more honest marathon estimate, and remember that the wall at mile 20 does not appear in any formula.
Which exponent should I use?
Use 1.06 if you train 50 or more miles per week and have raced the target distance before. Use 1.08 if you are on moderate mileage or stepping up to a distance for the first time. Use 1.10 for a first marathon on modest training, or any time you would rather finish strong than blow up chasing a number.
How recent should the input race be?
Within the last six to eight weeks is ideal, and within the current training block is best. Fitness changes: a personal best from last year reflects the runner you were, not the runner you are. Use your most recent all-out effort, and prefer an actual race over a hard workout, since races reliably pull out a harder effort than training does.
Does it work for ultramarathons?
Mostly no beyond 50K. The calculator will happily print predictions for 50 miles, 100K, and 100 miles, but at those distances hiking, fueling, terrain, elevation, and time on feet dominate the outcome far more than road speed does. A flat 50K prediction from a marathon is a reasonable ballpark; a 100-mile prediction from a 10K is a fiction.
Can I use a training run instead of a race?
You can, but expect the predictions to be conservative. The formula assumes an all-out effort, and almost nobody runs a time trial alone as hard as they race with competition, a crowd, and a clock at the finish. If a solo time trial is all you have, treat the predictions as a floor rather than a target.