How Much Should You Actually Be Running for a 50K?

The jump you are bracing for is smaller than you think

You have run a marathon. You have signed up for a 50K, which is 5 kilometres further, and somewhere in the back of your mind is the assumption that the training must be substantially bigger. More miles. Longer weeks. A different life.

The finisher data says otherwise. Runners who complete 50K races typically train in the region of 35 to 50 miles a week, roughly 56 to 80 kilometres, or about six to nine hours of running. That is the same range a reasonably serious marathon build already reaches. If you trained properly for your marathon, you are already training in the band that 50K finishers occupy.

What changes is not the size of the week. It is how you are allowed to grow the long run, and that turns out to be the one part of the conventional advice with real evidence behind it, pointing in a direction most runners have never been told.

Where your marathon build already sits

The numbers are worth stating precisely, because vagueness is what makes this question hard to answer.

In the one study that tested 50K runners directly, the 50-kilometre group averaged 66 km a week in the month before the race, about 41 miles, across 7.9 hours of running. Widen the lens to ultrarunners generally and the ULTRA Study's 1,212 active runners reported a median of 3,347 km over the prior year, which averages to roughly 64 km or 40 miles a week. Both land in the same place: a typical range of about 35 to 50 miles, with individuals scattered well above and below it.

Three honest caveats, and they matter more than the numbers.

First, there is no single study that establishes a 50K training band. The figures above are assembled from a small 50K cohort and a large survey of ultrarunners at all distances. Anyone quoting you a precise floor and ceiling for this race distance is interpolating, and so is this article.

Second, these describe what finishers did. They are not a prescription and they are not a threshold. Nobody has run a controlled trial of training volume at any ultra distance, and nobody is likely to. Every figure here comes from asking people who finished hard races what they had been doing, which cannot tell you what would have happened had they trained differently.

Third, the 50K evidence is one cohort of 21 finishers. It is the weakest kind of finding that is still worth acting on. Treat the band as a sanity check on a plan you have built from your own history, not as a target to generate one from.

There is one thing volume does buy at this distance specifically. In that study, which compared runners at 50, 80 and 160 kilometres on the same course on the same day, weekly training distance correlated with performance only at 50 km, and fairly strongly: r=-.66 for the month before the race. At 80 and 160 km no training variable predicted finish time at all. So more miles will probably make you faster over a 50K, in a way they will not over a hundred miler. The floor for finishing is low. The ceiling for going quickly is not.

The rule that actually binds

Here is the finding that should change how you build the next four months.

In 5,205 recreational runners tracked prospectively over 18 months, using Garmin data from 588,071 sessions, researchers looked for the training pattern that predicted injury. Week-over-week change in mileage did not predict it. The acute-to-chronic workload ratio, which a decade of running apps have been built around, did not predict it either, at least not in the direction anyone expected: higher ACWR was associated with a lower injury rate, with the largest spikes carrying a hazard rate ratio of 0.75.

What did predict injury was the size of a single session measured against the runner's own longest run in the previous 30 days. Relative to running the same or less, a session between 10% and 30% longer than that 30-day longest carried a hazard rate ratio of 1.64. Between 30% and 100% longer, 1.52. More than double, 2.28.

Read those three numbers carefully, because the shape is not the smooth ramp you might expect. The rate jumps as soon as you pass 10% and then sits roughly flat until you double, at which point it climbs sharply. There is no gentle slope to ride up. The penalty arrives early and then waits.

That gives you a usable rule, and it is the best evidenced thing in this article:

No run should exceed 110% of your longest run in the last 30 days.

It bites hardest in the situation nobody plans for. You are twelve weeks out. The plan says 22 miles this Saturday. But you had a chest infection three weeks ago and your longest run since is 14 miles. The plan does not know that. It has kept advancing while you have not, and Saturday's run is 57% above your actual 30-day longest, which puts it in a band the study associates with a roughly 50% higher injury rate.

The rule is not that you can never run further. It is that the step up has to be taken from where you actually are, and that if you have lost a block to illness, travel or work, you rebuild from the run you last completed rather than from the row in the spreadsheet.

Two limits on that rule are worth knowing. The cohort was general recreational runners, median 9.5 years of experience, and sessions over 100 km were excluded from the analysis, so applying it to ultra training is an extrapolation rather than a direct finding. And the authors are explicit that the rule governs one session at a time: running 11 km, then 12.1, then 13.3 in a single week each clears the 10% bar and may still be too much. The threshold is a floor for caution, not a licence to ratchet.

There is a second reason to care about the long run at this distance. In 63 male runners at a 24-hour race, the longest single training session before the event correlated with distance covered at r=.56, ahead of weekly volume at r=.31, and it was one of only two variables retained in the final regression model. Both were significant, so this is a difference of degree rather than a clean win for one over the other. But time on feet in one go appears to matter at least as much as the same distance spread across a week.

The ten percent rule is not what you were told

Nearly every training plan you will read is built on the idea that weekly mileage should rise by no more than 10%. It is stated as though it were a finding. It is not.

The study people reach for when they cite it tracked 874 novice runners with GPS. Across the three progression groups, overall injury rates did not differ at all. Within one specific subset of injuries, the distance-related ones such as patellofemoral pain, iliotibial band syndrome and medial tibial stress syndrome, those progressing by more than 30% carried a hazard ratio of 1.59 against the under-10% group, and even that result did not reach statistical significance: the confidence interval ran from 0.96 to 2.66, with a P value of .07.

So the detectable danger signal, such as it is, sits somewhere near 30% rather than 10%, applies to a subset of injury types, and is soft even there.

That does not make a 25% weekly jump a good idea. Growing slowly is still the right default, because adaptation takes time and because a build you can repeat beats one you can survive. But the reason is adaptation, not a proven injury threshold, and knowing which is which changes what you do when the arithmetic gets tight. It means a runner whose base cannot reach their target inside the weeks available should change the race, the timeline or the goal, rather than agonising over whether 12% is safer than 10%.

The acute-to-chronic workload ratio deserves its own answer, and it is messier than a clean dismissal. In the 5,205-runner cohort it ran backwards, with higher ratios predicting fewer injuries. In a Dutch cohort of 435 runners it ran backwards too. But in 119 runners training for a 56-km ultra, a ratio above 1.5 was associated with higher injury incidence. The honest summary is that the ACWR has not been shown to do the job it is sold as doing in runners, and that the one ultra-specific dataset pointing the conventional way is small. If your watch reports one, do not treat it as a safety check.

The risk runs the other way

The thing most first-time ultra runners worry about is doing too much. The evidence points the other way.

In the same 50, 80 and 160 kilometre cohort discussed above, non-finishers averaged 51 km a week in the final month before their race against 77 for finishers, and 36 against 58 over the preceding year. Both gaps were statistically significant. But the sample was 51 runners with only nine non-finishers, six of them in the 160 km race, so read this as a distance-scaling effect at the longest distance rather than a 50K finding. The direction is consistent with everywhere else it has been looked for.

In 106 runners tracked around the Comrades ultramarathon and modelled with generalised additive models, lower training distance across the twelve weeks before the race was associated with higher injury risk, not lower, with those injuries occurring mainly during or after the race itself. That study was retrospective and small, so treat it as directional.

The honest reading is not that more is safer. Every cohort showing declining risk at higher volume is a cohort of people who tolerated that volume, and no cliff has been observed within the range studied is a very different statement from high mileage is safe. But for a marathon runner stepping up to 50K, the familiar worry has the sign backwards. Arriving undertrained is the better-documented failure.

One caution is worth stating separately. If you have a history of stress fracture, the general injury literature supports being more conservative with volume growth and favouring frequency over single large sessions. That is a reasonable inference from bone-loading principles rather than a finding from any of the ultra studies here, so weigh it with a clinician who knows your history rather than treating it as settled.

What to do with this on Monday

Five things, and none of them is a plan template.

1. Look up your longest run in the last 30 days. That number, not your goal race, sets what you are allowed to do this weekend. Cap the next long run at 110% of it.
2. Check your marathon peak against the band. If you peaked in the mid thirties to fifty miles a week, you are already where 50K finishers sit. Your build is about holding that consistently, not growing it.
3. Build a repeatable average, not a single hero week. The volume that shows up in the finisher data is a sustained habit across a block, not one enormous week off a low base. A peak you reached once and could not repeat matches neither the performance data nor the injury data.
4. Run at least three days a week. Weekly running frequency did influence injury risk in the Comrades cohort, though the relationship was not a simple straight line, so treat this as sensible practice rather than a proven dose. Spreading volume out also makes the 110% rule far easier to obey.
5. If your race climbs, stop counting miles. On steep courses, weekly mileage stops describing what you are actually doing. Switch to hours plus vertical gain. This is a practical convention rather than a research finding, but the bands above stop applying in any useful way once a large share of your time is spent hiking.

None of that is difficult. The hard part of a first 50K was never the arithmetic, it was turning up often enough for long enough, which is a problem of systems rather than of numbers and is covered here.

The mileage you need is probably the mileage you already ran. What you need to change is the long run, and the rule for that one is better evidenced than almost anything else you will be told.

Research Cited

1. Schuster Brandt Frandsen, J. et al. (2025). How much running is too much? Identifying high-risk running sessions in a 5200-person cohort study. British Journal of Sports Medicine, 59(17), e109380.
2. Nielsen, R.O. et al. (2014). Excessive progression in weekly running distance and risk of running-related injuries: an association which varies according to type of injury. Journal of Orthopaedic & Sports Physical Therapy, 44(10), 739-747.
3. Coates, A.M., Berard, J.A., King, T.J. & Burr, J.F. (2021). Physiological determinants of ultramarathon trail-running performance. International Journal of Sports Physiology and Performance, 16(10), 1454-1461.
4. Knechtle, B., Knechtle, P., Rosemann, T. & Lepers, R. (2011). Personal best marathon time and longest training run, not anthropometry, predict performance in recreational 24-hour ultrarunners. Journal of Strength and Conditioning Research, 25(8), 2212-2218.
5. Hoffman, M.D. & Fogard, K. (2011). Factors related to successful completion of a 161-km ultramarathon. International Journal of Sports Physiology and Performance, 6(1), 25-37.
6. Hoffman, M.D. & Krishnan, E. (2013). Exercise habits of ultramarathon runners: baseline findings from the ULTRA Study. Journal of Strength and Conditioning Research.
7. Burgess, T., Durand, P. & Buchholtz, K. (2025). Assessing the relationship between training load and injury in ultramarathon runners: a novel approach using Generalised Additive Models. South African Journal of Sports Medicine, 37(1).
8. Craddock, N.L., Buchholtz, K. & Burgess, T.L. (2020). Does a greater training load increase the risk of injury and illness in ultramarathon runners? South African Journal of Sports Medicine, 32(1), 1-6.
9. Nakaoka, G. et al. (2021). The association between the acute:chronic workload ratio and running-related injuries in Dutch runners. Sports Medicine, 51, 2437-2447.