Hiking: What Does It Train, and How Should You Walk?

What a hike trains is decided by gradient, not distance. The climb is where your lungs complain and the descent is where your muscles take damage — two halves of one trail that are close to two different exercises, which is why each has its own way of being walked and recovered from.

Updated August 16, 2026

AI-researched

The answer, with evidence

At the base: several unbroken hours of walking is a documented dose in its own right — this layer does not depend on gradient at all.

15 cohorts, 47,471 adults, 3,013 deaths over a median 7.1 years: against the lowest quartile of daily steps (median 3,553), the other three carried all-cause mortality hazard ratios of 0.60, 0.55 and 0.47. The curve flattened around 6,000–8,000 steps a day for adults over 60 and 8,000–10,000 for those under 60. [1] How fast you walk depends on which measurement you trust: across 11 British cohorts of 50,225 walkers, among the 49,731 who had no event in the first two years of follow-up, self-rated average and brisk pace carried 20% and 24% lower all-cause mortality than a slow pace [2], while the device-measured meta-analysis above found cadence largely stopped predicting once total step count was accounted for. [1] Different instruments, so this page keeps both.

The real variable: not walking versus running, but gradient — the same trail up and down is close to two different exercises.

Measured on 10 runners across grades from −45% to +45%: walking cost 1.64 J per kilogram per metre on the level, rose to 17.33 J at +45% — more than tenfold — and fell to 0.81 J at −10%, cheaper than flat ground. Above +15% and below −15%, mechanical efficiency approached that of purely concentric and purely eccentric contraction respectively. [3] Note who was measured: ten trained runners on a treadmill. The magnitudes are the spine of this page, but carrying a walking-economy curve from trained runners to a general hiker is an extrapolation, not a result of the study.

The other half: the descent is not the way home, it is a training stimulus of its own — and also the day’s real bill.

127 sedentary adults were allocated to two groups; the 102 in the active one took a cable car up and walked only down, 540 m of descent, three to five times a week for two months. Insulin resistance, fasting glucose, glucose tolerance, triglyceride tolerance, body mass index and C-reactive protein all improved significantly, while a matched sedentary control group did not change. [4] The cost sits on the same side: after one ascent and descent of Mount Snowdon, 37 active adults were still losing maximal voluntary contraction at 48 hours. [5] The half that did not feel hard is the half still measurable two days later.

Trekking poles: they work, and not only by saving effort — what they reduce is muscle damage.

In the same study the pole group reported lower exertion on the ascent, lost less maximal voluntary contraction immediately and at 24 and 48 hours, reported less soreness at 24 and 48 hours, and showed lower creatine kinase at 24 hours; vertical jump did not differ. [5] One mountain, one day, 37 people — hold the effect size loosely, but four measures point the same way.

Pacing: judge it by whether you can still speak in full sentences — and here that test genuinely applies.

The talk test is validated against the ventilatory threshold: comfortable speech is likely below it and not likely above it. [7] Two other guides on this site cite the same review for its caveat that the test is impractical during intervals. A long sub-threshold day on foot is precisely the setting it was built for.

Recovery: steep descent is a better thing to scale it to than distance — and expect the second time to be far easier than the first.

Eccentric-biased lower-limb exercise costs 24–48 h of maximal voluntary contraction, on a recovery time-course that varies with how severe the damage was — a systematic review and meta-analysis of 141 studies. [8] Join that to descending being eccentric [3] and to a mountain day leaving force below baseline for two days [5], and the rule above follows. One step in it is not measured: no trial has compared a climb-heavy day against a descent-heavy day for recovery, and a search returns none, so that last inference comes from mechanism rather than from a result. As for the second time being easier: the repeated bout effect is well established, while its mechanisms — neural, mechanical, extracellular matrix remodelling, biochemical signalling — are not. [6] That review quantifies no duration of protection, so this page does not offer one either.

Up and down the same trail are not the same exercise

Comparing hiking with running to see which "burns more" asks the wrong question. On the level, running costs about 3.4 J per kilogram per metre and walking 1.64 — running is roughly double. But hiking rarely happens on the level, and gradient moves that number by an order of magnitude more than the choice of gait does.

Gradient

Walking cost (J/kg/m)

What it means

+45% (very steep up)

17.33

Over ten times level ground; near-pure concentric work

Level

1.64

The minimum, at roughly 1 m/s

−10% (gentle down)

0.81

The cheapest point anywhere on the curve

−45% (very steep down)

3.46

Back up again, and now almost entirely eccentric

The counter-intuitive row is −10%: a gentle descent is the cheapest place on the entire curve, cheaper than flat ground. The same study offers a practical corollary — the most efficient gradient for a mountain path is somewhere around 20–30%, which is why a well-built trail switchbacks instead of running straight up the fall line. [3] Worth stating plainly: these numbers come from 10 runners on a treadmill, not from field measurement, and trained runners are not the reader this page addresses. Trust the magnitudes, not the decimals.

One conclusion in that table is easy to read backwards. The −10% row is the cheapest terrain on the curve, but cheap describes oxygen, not training: the descent-only study whose numbers are in the conclusions above is precisely the one that pulls those apart, since its participants rode a cable car up and their metabolic markers improved anyway. [4] Three labels it needs: groups were allocated rather than randomised, the absolute changes are all small, and the control group was sedentary — so the comparison is descending versus doing nothing, not descending versus climbing. What it supports is one sentence, that the descent is training too, and no promise at all about weight.

The descent is the half that sends the bill

Put those two facts together and you get the sentence this page exists for: the descent costs the least oxygen and does the most muscle damage. Climbing, your quadriceps shorten under load and your cardiorespiratory system is what complains. Descending, the same muscles lengthen while resisting gravity — eccentric contraction, the kind that produces delayed-onset soreness, and the kind that barely raises your breathing at the time.

The Snowdon study put a duration on the cost: 37 active adults carrying roughly 5.6 kg day packs made one ascent and descent, and maximal voluntary contraction had still not returned to baseline immediately afterwards, at 24 hours, or at 48 hours. [5] An ordinary day hike leaves your legs unable to produce their usual force for two days afterwards — and none of that is perceptible while you are walking down. (What the abstract reports is that the pole group lost less force at each of those points; it does not print each group’s drop, so what travels here is the two-day duration, not a figure.) Whether the soreness itself tells you anything about training effect is the soreness guide’s question; here it is only used to plan the calendar.

Do not estimate recovery from how hard it felt

Cardiorespiratory fatigue clears in hours; the loss of force from eccentric work is measured in days. A route that felt easy because it was mostly downhill can need more recovery than a climb of the same length that left you breathing hard. Judge it by how much descent there was and how steep it was, not by how out of breath you got.

Whether you are walking it right is judged by whether you can still talk

The intensity rule is simple: you should be able to produce full sentences comfortably, all day. The talk test is validated against the ventilatory threshold — comfortable speech below it, not above. [7] Worth noting that two other guides here cite this same review for its limitation, that it "may not be practical for high-intensity interval training". A hike is the inverse case: hours long, held below the threshold, which is exactly the job the instrument was designed for.

As for how often to stop — no trial has compared break schedules, so any precise ratio is convention rather than a finding. What follows is reasoned from that convention plus the evidence above, and labelled as such:

A short stop every 45–60 minutes and a longer one before a sustained climb. That is convention, not a trial result — the real test is whether your speech is still easy.

Break a long steep descent into segments with flatter stretches between. Eccentric damage accumulates with time under load; charging straight down is not the same as taking it in three parts.

Use poles if you have them, especially on the way down. What they reduce is not only perceived effort but measured markers of muscle damage. [5]

A heavier pack costs more energy per metre — that is arithmetic from the per-kilogram cost, not a finding about damage. [3] And for the record: the day packs in the pole study weighed 5.6 ± 1.5 kg, which is simply what those participants carried. No arm of the study varied the load, so that number is not a tested ceiling.

"This feels easy today" is not a reason to add distance. Gentle descent is the cheapest terrain on the whole curve, and the bill is posted the following morning.

How many days to leave after a long one

Look at how much steep descent there was, not only at the distance — and steep is the operative word: a −10% grade is the cheapest terrain on the whole curve, and the eccentric share only becomes pronounced below about −15%. [3] After a day that was mostly flat or gently graded, train normally tomorrow. After a long steep descent, plan on the rule that no heavy lower-body lifting goes into the next two days — that is the window in which maximal voluntary contraction has not yet returned to baseline. [5] How to spread load across a week belongs to the recovery guide; the hiking-specific addition is only this: descent is the variable nobody counts. It is not the only thing driving recovery from a long day — time on feet, load and sleep all matter — but those you were going to weigh anyway, and this one you were not.

One more thing, useful for anyone who hikes seasonally: the first one is always the worst one. After a single damaging bout, the same stimulus produces markedly less damage next time — the repeated bout effect. Its existence is well established and its mechanism is not: the review lists neural adaptation, altered muscle mechanical properties, extracellular matrix remodelling and biochemical signalling, and proposes they work in concert. [6] It puts no number on how long the protection lasts, so neither does this page. "The second time is much easier" is settled; "and it still counts N weeks later" is not.

Two hiking days back to back: make the second one climb-heavy and descent-light, and save the long descent for the last day.

Hiking in the same week as lower-body lifting: put heavy squats and deadlifts before the hike, or two days clear of it.

On the first long hike of a season, deliberately go short — buy the repeated bout effect with a moderate day first.

If stairs are still noticeably hard 48 hours later, that is not a discipline problem but unrecovered force production. Let it push the next one back.

Frequently asked questions

Does hiking count as real cardio?

Yes, and the evidence for it is unusually direct. A meta-analysis of 15 cohorts and 47,471 adults found that the more steps people took, the lower their all-cause mortality: against the least active quartile, hazard ratios were 0.60, 0.55 and 0.47. The benefit levelled off around 6,000–8,000 steps a day for people over 60 and 8,000–10,000 for those under 60. A hike is several hours of exactly that, usually on a gradient, so it lands well inside the range where the curve is still falling.

Is hiking as good a workout as running?

For a given distance on flat ground, no — running costs roughly twice as much energy per kilometre. But hiking is rarely flat, and gradient changes the arithmetic sharply: walking uphill at a 45% grade costs over ten times what walking on the level does. A steep hour of hiking is not a lesser version of a run; it is a different stimulus, with far more eccentric loading on the way back down and much less repetitive impact on the way up.

Why is walking uphill so much harder than walking on flat ground?

Because you are lifting your whole body against gravity, and the cost scales steeply with gradient. Measured on a treadmill across grades from −45% to +45%, the energy cost of walking was about 1.64 J per kilogram per metre on the level and 17.33 at +45% — more than a tenfold increase. The same study found the most efficient gradient for a mountain path is around 20–30%, which is why well-built trails switchback rather than going straight up.

Why do my legs hurt more the day after a hike than during it?

Because the descent, not the climb, does most of the damage, and it does it quietly. Walking downhill at a shallow grade costs less energy than walking on flat ground, so it never feels like the hard part — but your quadriceps are working eccentrically, lengthening under load, which is the contraction type that produces delayed soreness. In a study of 37 people hiking up and down Mount Snowdon, force production was still measurably reduced 48 hours later. What that soreness does and does not mean is covered in the muscle-soreness guide.

Do trekking poles actually do anything?

They do more than they look like they should. In a controlled field study of 37 hikers carrying day packs up and down Mount Snowdon, the group using poles reported lower perceived exertion on the ascent, lost less maximal voluntary contraction immediately after and at 24 and 48 hours, reported less soreness at 24 and 48 hours, and had lower creatine kinase at 24 hours. It is one mountain and 37 people, so treat the size of the effect loosely — but the direction is consistent across four separate measures.

How often should I stop and rest on a long hike?

There is no trial that compared break schedules, so anyone quoting a precise ratio is quoting convention. What is evidence-based is the pacing test: you should be able to speak in full sentences comfortably for most of the day, because the talk test tracks the ventilatory threshold well in exactly this intensity range. A practical default is a short stop every 45–60 minutes and a longer one before a sustained climb, adjusted by whether your speech is still easy.

How many days should I rest after a long hike?

Judge it by the descent rather than the distance. If it was mostly flat or gently graded, the next day can be normal. After a long steep descent, expect reduced force production for one to two days and plan around it — do not schedule heavy lower-body lifting into that window. The good news is that the protection builds: after one damaging bout, the same effort produces markedly less damage next time, an effect that is well documented even though its mechanisms are still not fully understood.

Do I need to hit 10,000 steps a day?

No. The 10,000 figure is a marketing number, not a finding. In the 15-cohort meta-analysis, the risk curve flattened around 6,000–8,000 steps a day for adults over 60 and 8,000–10,000 for younger adults, and the steepest part of the benefit came well below any of those. The same analysis found that once total step count was accounted for, walking faster added much less than most people assume.