Ask hikers about Mount Washington and they’ll tell you about the boulder field on the way up — Tuckerman Ravine’s headwall, a scramble over loose rock that turns a hike into a workout. Fewer mention the way down. White Mountain guidebooks routinely warn that the descent takes as long as the climb, sometimes longer, on a pitch steep enough to chew through knees and confidence both. Naismith’s rule, in its plain form, treats that descent as basically free.
That’s not a knock on Naismith. His 1892 formula was built for Scottish hill walking, not scree fields, and it holds up well for what it actually covers — distance and ascent. We’ve written about the base rule and its fitness/terrain corrections before. This one is about the correction most calculators skip. Including, for now, ours.
The correction Langmuir added
Eric Langmuir’s mountaineering handbook extended Naismith’s rule downhill, and the fix isn’t what most people guess. On a gentle decline — Wikipedia’s summary of his work puts the threshold at 5–12 degrees — you genuinely move faster than flat ground, so Langmuir subtracts 10 minutes per 300 m descended. Past 12 degrees, steep enough that Tuckerman’s headwall clears it easily, the correction flips: add 10 minutes per 300 m instead.
Steep downhill isn’t free time. It’s slower than flat ground, not faster.
Our hiking time calculator doesn’t apply either half of that correction yet. The source code says why, in a comment right above the formula: descent is “intentionally NOT modeled in v1,” because Langmuir’s fix needs a slope angle most hikers can’t read off a trailhead sign. That’s an honest limitation. Still a limitation.
What that costs on a real descent
Take the return trip down Tuckerman Ravine Trail: about 4.2 miles (6.8 km), roughly 4,250 ft (1,300 m) of descent, rocky terrain most of the way.
| Leg | Our calculator (no descent correction) | Langmuir-corrected |
|---|---|---|
| Distance + rocky terrain (×1.35) | 2.64 h | 2.64 h |
| Steep-descent add (12°+, 10 min/300 m) | — | +0.72 h |
| Total | 2.64 h | 3.36 h |
That’s a 43-minute gap on one leg of the hike — call it 27% too fast, on exactly the kind of trail where being 27% too fast matters. You’re timing whether you clear the boulder field before dark, not whether you make it home for dinner.
When to trust the number, when to pad it
Compare that to a mellow counterexample: an easy lakeside loop, 6 km (3.7 mi) of good trail dropping 200 m at roughly 8 degrees — squarely inside Langmuir’s gentle-decline band. Our calculator has that leg at 1 h 43 min. Langmuir’s correction subtracts about 7 minutes for the gentle grade, landing closer to 1 h 36 min. That’s a rounding error, not a planning risk.
So: on a maintained, moderate-grade trail, don’t bother correcting for this. The gap runs small, and our omission even trends slightly conservative there, since gentle downhill should technically move a hair faster than we credit it. Don’t trust the raw output on anything with a rock-scramble reputation, though — Tuckerman, Half Dome’s cables, any descent locals describe using the word “knees.” Add 20–30% to the return leg by hand until a real slope-angle input ships.
The bigger caveat
Even Langmuir’s fix isn’t settled science. Wikipedia’s own summary notes the accuracy of descent corrections is disputed, particularly at gentle gradients — different hikers, different footwear, different rock all move the number around. No formula, ours included, replaces watching the clock against the map once you’re actually on the way down.
Run your ascent numbers through the hiking time calculator, then trace the real grade of your descent in the route planner before you set a turnaround time. Plan the way down. It’s the half of the hike Naismith’s rule was never built for.