Your health comes first. This educational article provides general information. Discuss your individual circumstances, pregnancy, illness or symptoms with your clinician; the article cannot replace an individual assessment.

After a half-marathon close to an hour and a half, a calculator can easily display a marathon time that is both exciting and suspiciously attractive. The calculation may be flawless while the conclusion is too bold. Predicting from the shorter distance means estimating how your sustainable speed changes over several more hours of exertion. Your half-marathon time alone provides too little information for that.

The useful question is therefore not which calculator to trust unconditionally. Instead, examine the assumptions of the chosen formula, how closely you resemble the runners studied, and what reveals the specifics of your marathon preparation. The calculations below are illustrative examples, not individual target-time prescriptions.

1. What does a distance-ratio formula do?

One common form of the relationship known as Riegel’s formula is: predicted time = known time × (new distance / known distance)k. The exponent expresses how time increases with distance. If k were exactly one, twice the distance would take exactly twice the time. With k greater than one, we allow for a slower average speed over the longer distance.

The distance ratio from half-marathon to marathon is two. For a 1:50 half-marathon, or 110 minutes, using k = 1.07 gives 110 × 21.07, approximately 231 minutes. That is a marathon of about 3:51. The formula has not asked how much running fits into your week, whether you have had a long interruption, or how you managed the final part of your latest long run.

The seconds on the display indicate mathematical precision, not certainty of prediction. If the program says 3:50:56, that does not mean it knows your future performance to the nearest second. For practical decisions, a rounded range accompanied by a list of missing conditions is often more useful.

2. The assumption hidden in the exponent

For the same 110-minute result, k = 1.05 gives roughly 228 minutes, while k = 1.10 gives 236 minutes. That is a difference of more than eight minutes without your half-marathon time changing by a single second. The difference comes from the model used. A parameter tucked away among decimal places can therefore noticeably alter a race target.

The exponent does not simply describe how determined you are. It compresses the relationship between your shorter- and longer-distance performances into one number. Training history, adaptation to the duration and race conditions may all contribute. Two runners can finish the same half-marathon in the same time yet slow by different amounts over the longer distance.

Adjusting the exponent until your desired target appears is not a good solution. That would fit the result to your wish. If you have several races from a similar period, you can examine the shape of your own performance curve. This also needs caution, because a hot 10K and a cool half-marathon are not a clean comparison.

3. What did research find in recreational runners?

In the original study by Vickers and Vertosick, distance-ratio prediction worked better for shorter races but was consistently too optimistic for the marathon. Among the marathon predictions examined, it gave many runners substantially faster times than they actually achieved. Models also using weekly running volume provided better estimates in this database.

This does not mean the formula is wrong for everyone, or that an accurately predicted time has no value. The lesson concerns how reliability is distributed: a rule that sounds good on average can miss substantially for a particular runner. The study’s voluntary questionnaire sample is not a perfect representation of every marathon runner either, so it cannot yield a universal correction.

A common misunderstanding would be simply to add ten minutes to every prediction based on this research. That would place another poorly generalised rule beside the poorly generalised formula. Instead, the result supports treating the number derived from a shorter race as a conditional starting point and separately examining your background in marathon-duration exertion.

Two recreational runners on an easy, conversational long run along a riverside promenade.Enlarge image
AI-generated editorial illustration; not a documentary photograph.

4. Why is it more than a longer half-marathon?

Doubling the distance also pushes your finishing time much further into the day. The same morning may become a different part of the day, access to food and fluids may change, and you must take far more steps at your initially chosen rhythm. These factors cannot all be folded back into a half-marathon average pace.

A half-marathon can show the speed you can sustain that day for well over an hour and a half or for two hours. A marathon decision also raises the question of how stable your functioning remains later. With an identical early pace, the final third can differ. The relationship between the two races is therefore not solely about speed.

The longest training run is not the only evidence of specific preparation. Consider whether longer runs were repeatable, how they fitted into the rest of your week, and how you felt when continuing your normal responsibilities afterwards. None of this guarantees a finishing time, but it provides more relevant information than one excellent shorter race.

5. The same 1:50, two different backgrounds

Imagine two runners who both run 1:50 in the same flat half-marathon. One has trained regularly with balanced demands for months and completed several longer runs under control. The other performs well over shorter distances but has recently had few opportunities for long training runs. The calculator gives both the same number.

This does not mean the first runner will definitely run a faster marathon. It does show why the prediction’s foundation is incomplete. More questions remain for the second runner about how performance holds up over a longer duration. That uncertainty deserves a place in the race target; there is no need to pretend everything is known.

Weekly mileage alone is not a complete explanation either. Forty kilometres could mean five easier outings, or two short runs and a disproportionately demanding weekend session. The number has a structure and a history behind it. Better research outcomes in a higher-volume group do not justify a sudden increase or prescribe the same weekly target for everyone.

6. Three checks before using the half-marathon result

First, examine when and under what conditions the result was achieved. When did you run it, what was the course like, and was it an all-out race? A previous best does not necessarily represent your current condition. Conversely, there is no reason to compare pace on a hilly course with a fast, flat race solely on the basis of kilometres.

Second, check the basis of timing. Official chip time, the watch’s elapsed time and automatically paused moving time may differ. If stops are excluded from the half-marathon input, the marathon estimate may rest on a condition that does not exist in an actual race. A GPS-measured 21.1 kilometres is not automatically an official half-marathon result either.

Third, examine what has happened since the result. Have you prepared consistently, trained for another distance, or had a substantial interruption? You do not need to invent penalty minutes for every small change. Recognising that the input’s validity has changed is enough: the original number can no longer be used with the same degree of confidence.

7. What can critical speed or a laboratory add?

A different type of estimate requires a different type of data. The research examining critical speed from training files used fast efforts of several durations. This can provide a more detailed performance profile than one race. Even here, however, the results involved prediction error; a large database does not eliminate individual uncertainty.

Laboratory testing can show which speeds are associated with particular physiological changes. A study following sixteen recreational runners measured threshold speeds and running economy at several points during a marathon training cycle. Its small sample and particular preparation system mean that it, too, cannot provide a time-prediction recipe transferable to every runner.

More measurements are useful when they genuinely clarify different questions. Agreement between three apps based on the same short race is not three independent pieces of evidence. A race result, a consistent training history and an interpretable measurement can, however, illuminate different sides of the situation. Even together, they cannot turn missing marathon experience into an event that has already happened.

A recreational runner on a controlled, easy long run along a quiet lakeside footpath.Enlarge image
AI-generated editorial illustration; not a documentary photograph.

8. How can a number become a cautious target-time range?

Start by recording the calculator’s result and the input used. In a separate column, list supporting signs and open questions. A recent, comparable race and consistent preparation may be supporting signs. Experience of the longer duration or the expected race environment may remain open questions.

Do not create the range by averaging the three most favourable predictions. Instead, describe the conditions under which the more optimistic option would become realistic, and what circumstances would justify a more cautious choice. Until there is an individually validated model, the range’s width is a constructed decision-making tool. It cannot automatically be called a ninety-five per cent statistical interval.

Record the decision and your wish in separate sentences. “I would like to finish within four hours” expresses a goal. “My current information leaves uncertainty about how stable my performance will remain over that duration” assesses the situation. Both can be true simultaneously. Wanting a particular boundary does not require you to force every second of it from the opening part of the race.

9. A short worksheet for your next calculation

Before next using a calculator, make a five-line note: known race and date; timing basis; course and weather; characteristics of your preparation since then; and available experience of the longer duration. Then add the formula’s result. This keeps the prediction connected to the conditions that produced it.

Add a specific review date, such as your next discussion with your coach. You do not need to recalculate the future after every run. A review chosen in advance can prevent an unusually good or bad day from determining the entire marathon plan. Relevant new information may deserve more weight than the app’s daily change of mood.

Finally, check what decision the number would change. If it affects neither when you enter, the structure of preparation nor your starting plan, there is no need to argue over seconds. If an uncertain prediction would make you cling to an excessively narrow target, return to its assumptions. A calculator helps when it organises questions, rather than being used as though the finish gantry has already decided the result.

Sources

Helpful? Pass it on.