What this race time predictor does
Give it one race you have actually run — a 10K, a half, a parkrun, a marathon — and it returns the finish time you are currently capable of at a different distance, plus the pace you would have to hold to get it. It is a race time predictor, not a training plan: it reads the fitness you have already demonstrated and moves it along the distance curve.
Four things come back, and all four are on screen at once: the predicted finish time, your average pace in both kilometres and miles, your average speed, and a split table for the target race so you know what each kilometre or mile has to look like on the day. One click copies the whole thing as text.
It is the same job a marathon time predictor or a half marathon predictor does. Those are not separate tools with separate maths — they are the same calculation pointed at a different target distance, which is why this page handles 100 metres to 100 miles from one form.
How the number is calculated
The default engine is Peter Riegel's endurance equation, published in 1981 and still the standard baseline in running:
T₂ = T₁ × (D₂ ÷ D₁)C
- T₁ — your recent race finish time, in seconds (the time you typed in).
- D₁ — your recent race distance, in metres (your entry × 1000 for km, × 1609.344 for miles).
- D₂ — your target race distance, in metres.
- T₂ — the predicted finish time, in seconds.
- C — the slowdown exponent. Default 1.06, editable under Method.
If you give a second race
Two results at two different distances pin down your own exponent instead of the population average:
C = ln(T₂ ÷ T₁) ÷ ln(D₂ ÷ D₁)
where the two races are (D₁, T₁) and (D₂, T₂), both converted to metres and seconds. The prediction is then anchored on the longer of the two races, because extrapolating from the closer distance is the smaller extrapolation. If the solved exponent falls outside 0.90–1.40 — which is what you get from a mistyped time, or two races run at very different efforts — the page says so and falls back to 1.06 rather than reporting a number it does not trust.
How the splits are derived
Splits are arithmetic on the predicted time, not a second model. With Even pacing, every segment gets the same pace: segment time = total time × (segment distance ÷ target distance). With Negative split, the first half of the distance is run 2 % slower than average and the second half 2 % faster, which leaves the total unchanged because the two halves are the same length:
T = (D/2)·p·(1+0.02) + (D/2)·p·(1−0.02) = D·p = T₂
The bar beside each row is scaled to that segment's pace, so the shape of the race is visible at a glance. Segments longer than a whole kilometre or mile (the tail of a 21.0975 km target, say) are shown with their true partial length. Above 40 segments the table groups rows so it stays readable.
Unit conversions used
These are definitions, not estimates: 1 km = 1000 m exactly; 1 mile = 1609.344 m exactly (the international mile); pace per km = seconds ÷ (metres ÷ 1000); pace per mile = seconds ÷ (metres ÷ 1609.344); speed in km/h = (metres ÷ 1000) ÷ (seconds ÷ 3600).
Supported range and limits
- Distance: 100 m to 100 miles (0.0622 – 160.9344 km), on both the recent race and the target. That is the widest range any of the tools this page was measured against accepts — the same span as Run Hive, which goes further than SportTracks (stops at the marathon) and MyProCoach (stops at the half).
- Units: kilometres or miles, switchable independently on each field.
- Time: hours, minutes and seconds, up to 47:59:59.
- Cost: none. No account, no email, no paywall, no daily limit, no request to a server.
- Privacy: every number is computed in your browser. Nothing is sent anywhere and nothing is stored — reload the page and it is gone.
- Works offline once the page has loaded, and on a phone.
What this page deliberately does not do
Each of these is a real feature on at least one of the tools measured for this page. They are missing on purpose, and here is the reason:
- Age and sex adjustments. A marathon race time predictor with an age-grading column needs a published age-factor table, and this page has no source for one it can stand behind. Rather than ship a made-up coefficient, it leaves them out.
- Weather correction (heat, humidity, wind). The tools that offer it do not publish the model. Adding an unsourced adjustment would make the number look more precise than it is.
- Bulk mode, accounts, saved history. This is a one-race, one-answer page. Multi-row input, sign-in and stored results all need a server, and none of them are needed to answer the question.
- Elevation or trail adjustment. Not modelled. If your race is hilly, treat the prediction as a flat-course number.
Questions
How accurate is a race time predictor?
Close when the two distances are near each other and the input was a genuine race effort; progressively worse as the gap widens. A 10K predicted from a 5K is reliable. A marathon predicted from a 1K is not, because endurance, fuelling and pacing become the deciding factors and no distance formula contains them. Treat the number as the time your current fitness is worth, then discount it for a hilly or hot course.
Is this the same as a marathon time predictor?
Yes — same equation, different target distance. Set the target to 42.195 km or 26.2 miles and the page is a marathon time predictor; set it to 21.0975 km and it is a half marathon predictor. There is no separate model per distance, and no need to visit a different page.
What time should I enter?
A recent race you ran all-out, ideally within the last two to three months. A training run, a tempo effort or a parkrun you jogged will make the prediction too conservative. If the recent race was run in very different conditions from the target, expect the output to drift.
Why is my predicted marathon time so much faster than I think I can run?
Because the formula only knows pace and distance. It assumes you have done the endurance work for the longer event. Most runners who see an unrealistic marathon prediction from a fast 5K have not yet run the long runs the marathon needs — the formula is describing potential, not certainty.
Can I predict a race time for a custom distance?
Yes. The distance fields are free numbers, not a fixed list, so any distance inside the supported range works — a 7-mile race, a 30K, a 100-mile ultra.
Can I use it for a parkrun or a treadmill run?
A parkrun is fine: it is 5 km, it is measured, and it is usually run hard. A treadmill run is a weaker input, because the belt removes wind and pace drift and the distance is only as good as the machine's calibration. It still works; just expect a slightly optimistic result.
Does it work on a phone?
Yes. The form is a single column on a narrow screen, the number pad comes up for the distance and time fields, the split table scrolls inside its own box rather than pushing the page sideways, and the copy button puts the whole result on your clipboard.