If the road from Bengaluru to Kolkata sloped downhill all the way, you could coast the full 2,000 km or so with the engine off at a steady 60 km/h only if the road started about 35 km above Kolkata, high in the stratosphere and well above the 19 km Armstrong limit where an unprotected body's fluids boil. The slope itself is modest: about 4.7% near the ground, easing to 1.2% where the air is thin. The real road drops only about 2,936 ft (895 m) over about 2,000 km, a 0.045% grade, which would let a coasting bike roll roughly half a kilometre to a kilometre before stopping.
- 0.045% is the real average grade from Bengaluru to Kolkata: far too gentle to roll on.
- ~1.2% is the slope a bike needs just to beat its tyres' rolling resistance.
- 4.7% holds a steady 60 km/h engine-off in sea-level air, because drag is about 3× the tyre loss.
- ~35 km up is where the coasting road would have to start, above the 19 km Armstrong limit.
- 2.2 MJ from the real 895 m drop pays for only about 1% of the trip at the wheel.
Bengaluru sits at about 900 m on the Deccan Plateau and Kolkata at 1.5 to 9 m in the Ganges delta, so the trip is, technically, downhill already. It just isn't downhill enough.
This what-if works out how much slope a bike needs to hold speed without an engine, what that does to the starting height, and why the thin air of the upper atmosphere quietly halves the answer.
How far would you coast on the real Bengaluru–Kolkata road?
About half a kilometre to just over a kilometre. The average slope of the real route is 2,936 ft (895 m) over 2,000 km, or 0.045%. A rolling motorcycle needs about 1.2% just to beat its tyres, so the real slope is 27 times too gentle to matter.
Calculated in this post from the height drop (Bengaluru at about 900 m, Wikipedia; Kolkata at 1.5 to 9 m, Wikipedia, taken as about 5 m) and the roughly 2,000 km route length of Sapta's November 2025 ride. The 1.2% is the tyre rolling-resistance coefficient of 0.012, the middle of the range in Wiratkasem and Pattana, 2021.
About 26.5 km west of Leh, on the Srinagar–Leh road between Nimmoo and Leh, there is a stretch known as Magnetic Hill. Leave a car free to roll there and it appears to creep uphill on its own, against gravity. Some visitors still credit a magnetic pull from the hill.
There is nothing magnetic about it. The road actually runs gently downhill; the layout of the surrounding slopes hides the true horizon, so the eye reads a mild descent as a climb. It is a neat lesson for this what-if: a gentle grade is almost invisible to a rider, and gravity only moves you once the slope beats your tyres.
Wikipedia: Magnetic Hill (India), citing Bangalore Mirror, "Nothing magnetic about Magnetic Hill" (2015)The working
Assumptions, with these example inputs: bike, rider and luggage weigh \( m = 250 \) kg (an example, not a measurement); rolling resistance coefficient \( C_{rr} = 0.012 \), the middle of the 0.0089–0.0137 range measured on five commercial motorcycle tyres; drag area \( C_dA = 0.5 \) m², the upright-rider value measured in a full-scale wind tunnel; sea-level air density \( \rho_0 = 1.225 \) kg/m³ (standard atmosphere); \( g = 9.81 \) m/s².
- Ignore drag first. With the engine off at 60 km/h (16.7 m/s), the net slowing force per kilogram is \( g(C_{rr} - s) \), where \( s \) is the slope. The stopping distance is
\[ x = \frac{v^2}{2g\,(C_{rr} - s)} = \frac{16.7^2}{2 \times 9.81 \times (0.012 - 0.00045)} \approx 1{,}230\ \text{m} \]
- Add drag. Air resistance at 60 km/h is about 85 N, more than twice the 29 N of rolling resistance, so it shortens the roll. A numerical integration with both forces gives about 570 m.
Calculated in this post, with these example inputs (250 kg, drag area 0.5 m², Crr 0.012), from the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere).
- Energy check. Dropping 250 kg through 2,936 ft (895 m) releases \( mgh = 250 \times 9.81 \times 895 = 2.2 \) MJ. Riding 2,000 km at 60 km/h in sea-level air costs about 229 MJ at the wheel. The whole real descent pays for about 1% of the trip.
What forces does a coasting motorcycle fight?
Two forces push back, and one pulls you along:
- Rolling resistance from the tyres, which stays roughly constant with speed;
- Aerodynamic drag, which grows with the square of speed;
- Gravity along the slope, \( mg\sin\theta \), the only force on your side once the engine is off.
The first two together are the road load.
At 60 km/h on our assumptions the tyres take \( 0.012 \times 250 \times 9.81 = 29.4 \) N and the air takes \( \tfrac12 \times 1.225 \times 0.5 \times 16.7^2 = 85 \) N. Above about 35 km/h the air is the bigger cost, which is why fuel economy collapses at highway speed.
Rider posture moves the drag area by a third: the Warwick wind tunnel measured 0.50 m² upright and 0.35 m² tucked on the same bike.
The authors measured drag area (CdA) for four motorcycles, two petrol and two electric, in a full-scale wind tunnel at 17–35 m/s, with a rider in two-piece leathers in three postures. Bare bikes came in at 0.26–0.31 m²; adding an upright rider pushed that to 0.50–0.59 m², and a full tuck brought it down to 0.33–0.43 m².
A 30% cut in CdA from upright to tucked cut energy use over the WMTC test cycle by up to 19%. The limit: no rolling road or spinning wheels, and the engine was off.
Read the paper (free)How steep would the road need to be to coast at highway speed?
For a steady 60 km/h in sea-level air the road needs a 4.7% slope; at 80 km/h it needs 7.4%. At a steady speed gravity's pull along the road exactly balances the road load. For small angles \( \sin\theta \approx \tan\theta = s \) and \( \cos\theta \approx 1 \), so:
Calculated in this post, with these example inputs, from the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere).
- Balance the forces.
\[ m g\,s = C_{rr}\,m g + \tfrac{1}{2}\rho\,C_dA\,v^2 \quad\Rightarrow\quad s = C_{rr} + \frac{\rho\,C_dA\,v^2}{2\,m g} \]
- Plug in 60 km/h.
\[ s = 0.012 + \frac{1.225 \times 0.5 \times 16.7^2}{2 \times 250 \times 9.81} = 0.012 + 0.0347 = 0.0467 \]A 4.7% grade, about 1 in 21.
Calculated in this post, with these example inputs, from the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere).
- Starting height, if the air stayed as thick as at sea level.
\[ h = s \times L = 0.0467 \times 2{,}000\ \text{km} \approx 93\ \text{km} \]That is near the 100 km line where space conventionally begins.
Calculated in this post from the 4.67% slope above and the roughly 2,000 km route length of Sapta's November 2025 ride.
What Indian road standards allow
Indian highway engineers would not build that road. IRC practice for plain and rolling terrain sets a ruling gradient of 3.3%, a limiting gradient of 5% and an exceptional 6.7% for short stretches. Put the numbers the other way round and ask what speed each standard gradient gives a coasting bike:
| Gradient | IRC class (plain or rolling) | Coasting speed | Start height for 2,000 km (sea-level air) |
|---|---|---|---|
| 0.045% | The real Bengaluru–Kolkata average | None; you stop within about 1 km | 0.9 km |
| 1.2% | Below ruling | Just above walking pace | 24 km |
| 3.3% | Ruling gradient | 47 km/h | 66 km |
| 5.0% | Limiting gradient | 63 km/h | 100 km |
| 6.7% | Exceptional gradient | 76 km/h | 134 km |
Gravity only moves you once the slope beats your tyres; below about 1.2%, a downhill road is just a flat road that looks hopeful.
Check a slope yourself
Use this to check what you just read: set a grade and your own load, drag and tyre numbers, and it returns the engine-off speed where gravity and road load balance.
Engine-off coasting speed on a slope
Coasting speeds are from \( v = \sqrt{2mg(s - C_{rr})/(\rho C_dA)} \) with the assumptions above; gradients from the IRC table reproduced in NPTEL course notes.

Why does thin air cut the starting height in half?
Because drag depends on air density, and air density halves roughly every 6 km of height.
Near the top of our imaginary ramp the air is so thin that drag nearly vanishes and the road only needs to beat the tyres, about 1.2%. That changes the shape of the road: steep at the Kolkata end, where the air is thick, and almost flat up high.
The atmosphere's density falls off exponentially with a scale height of about \( H = 8.5 \) km (NASA's Earth fact sheet). The slope needed at height \( h \) is then
This has no neat closed form along a fixed length, so I integrated it numerically from Kolkata (5 m) back to Bengaluru in 100 m steps. At 60 km/h the road has to start 35.2 km up. Two quick checks on that answer: it can never be below \( C_{rr} L = 0.012 \times 2{,}000 = 24.0 \) km, the tyre-only floor, and it must be well below the 93 km of the constant-density case. It sits between them.
Calculated in this post by numerical integration, with these example inputs, from the 8.5 km scale height in the NASA Earth fact sheet and the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere).
See the data as a table
| Air as thick as sea level all the way (km) | Real atmosphere (thins with height) (km) | |
|---|---|---|
| 40 km/h | 54.8 | 30.7 |
| 60 km/h | 93.4 | 35.2 |
| 80 km/h | 147.3 | 39.0 |
| 100 km/h | 216.7 | 42.2 |
| 120 km/h | 301.5 | 45.0 |
Source: Worked on this page from Kirca and McGordon (CdA 0.5 m²), Wiratkasem and Pattana (Crr 0.012) and NASA (8.5 km scale height) · 250 kg bike, rider and luggage. Ignores temperature changes with height.
Could a rider survive at the top?
No, not without a pressure suit. Above about 4,500 m most people need supplemental oxygen, and at the Armstrong limit of about 18–19 km the pressure falls below 6.3 kPa, low enough that saliva and the fluid in the lungs boil at body temperature.
Consciousness goes within seconds and death follows in a minute or two unless pressure returns. On the 60 km/h road, the first 1,190 km or so from Bengaluru, more than half the trip, runs above 19 km. The profile only drops below that line about 810 km from Kolkata.
Calculated in this post from the integrated road profile (8.5 km scale height, NASA) and the 19 km Armstrong limit (Wikipedia), over the roughly 2,000 km route of Sapta's ride.
The bike would have its own troubles, though none would matter on a coast. A naturally aspirated engine breathes air in proportion to density; at 35 km the air density is estimated at about 1.6% of sea level by the exponential model used above, so there is nothing to burn even if you wanted to restart.
Tyres would see a big pressure difference across the carcass, and the rider's cooling depends on air too. For what thin air does to riders and machines at merely Himalayan heights, see RoadHook Rides' guide to altitude sickness on a motorcycle tour.

How sensitive is the answer to the assumptions?
Not very: across the measured range of tyres, postures and loads the start height stays between 30 and 38 km. That is the useful thing about this problem. The tyre term sets a hard floor and the air term adds a little on top.
Calculated in this post from the tyre range in Wiratkasem and Pattana (2021) and the drag-area range in Kirca and McGordon (2023); see the table below.
| Change (60 km/h, real atmosphere) | Start height | Why it moves |
|---|---|---|
| Base case: Crr 0.012, CdA 0.50 m², 250 kg | 35.2 km | - |
| Best measured tyre, Crr 0.0089 | 30.4 km | Lower tyre floor (17.8 km) |
| Worst measured tyre, Crr 0.0137 | 37.9 km | Higher tyre floor (27.4 km) |
| Full tuck, CdA 0.35 m² | 33.1 km | Less drag near the ground |
| Upright naked bike, CdA 0.59 m² | 36.2 km | More drag near the ground |
| Lighter, 200 kg | 36.6 km | Drag is the same but gravity pulls less |
| Heavier, 300 kg | 34.1 km | More weight to push through the same air |
Five motorcycle tyres sold in Thailand were tested to ISO 18164:2005 and had rolling resistance coefficients of 8.9, 10.3, 11.4, 12.4 and 13.7 N per kN of load. Each was then fitted to a motorcycle on a chassis dynamometer at a constant 50 km/h; fuel use rose steadily from the lowest- to the highest-resistance tyre. A 10% drop in the coefficient cut fuel use by about 1%. The limit: small commuter tyres at one speed, not a loaded touring bike on a highway.
Read the paper (free)How much fuel would the downhill road save?
All of it, roughly 13 to 34 litres for a typical touring bike, depending on how you count. The energy you need at the wheel is the road load times distance. In the thin-air version the work done by gravity equals the drop: \( mgh = 250 \times 9.81 \times 35{,}200 \approx 86 \) MJ. In the sea-level-air version it is about 229 MJ.
Petrol holds about 33.6 MJ per litre. If a small petrol engine delivers about 20% of that to the rear wheel, which is my rough estimate for part-throttle cruising, the fuel equivalent is:
The 20% efficiency is the shakiest number in this post, so treat the litres as an order of magnitude. To plan the real version with fuel stops, use RoadHook Rides' fuel range planning guide.
What happens if the slope is too steep?
Then the brakes take the difference as heat, continuously. On a 7% grade at 60 km/h, gravity pulls with \( 250 \times 9.81 \times 0.07 = 172 \) N against 114 N of road load. The extra 57 N, at 16.7 m/s, is about 0.95 kW of heat that the brakes must shed for as long as the descent lasts.
Calculated in this post from a 7% grade (a scenario, not a real road) and the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere).
That is why trucks lose brakes on long ghat descents and why RoadHook Rides' piece on ghat road design and hairpin bends matters more than any coasting fantasy. It is also why real descents such as the 2,700 m drop from Zuluk to the plains are ridden in gear, the way every ghat road is meant to be ridden.

Energy conservation settles it. Coasting about 2,000 km at 60 km/h needs about 86 MJ of gravitational energy even in the most favourable thin-air layout, which means starting about 35 km up, above the Armstrong limit and about four times the height of Everest. The real road drops about 2,936 ft (895 m) and supplies 2.2 MJ, around 1% of what the trip needs at the wheel.
Calculated in this post from the 895 m drop (Bengaluru about 900 m, Wikipedia; Kolkata 1.5 to 9 m, Wikipedia), the roughly 2,000 km route of Sapta's ride, the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere), and Everest at 8,848.86 m (Wikipedia).
No tyre, posture or load in the measured range brings the start below 30 km. The nearest version that works is an electric motorcycle that recovers energy on long descents, or a gentle 1.2–1.5% grade a few kilometres long, where a bike will hold a slow, engine-off roll.
Calculated in this post from the tyre and drag-area ranges in Wiratkasem and Pattana (2021) and Kirca and McGordon (2023); see the sensitivity table above.
What's new
- At least 11 killed when a truck descending NH-135 near Mirzapur suffers suspected brake failure ThePrint, Apr 2026
- Three killed as a loaded truck loses its brakes descending Bherughat on the Agra–Mumbai highway Free Press Journal, Oct 2026
What if next?
- What if you stacked every speed breaker in India? Every hump on the route above would cost you speed you could never get back.
- What if the Earth stopped spinning while you were on the highway? On NH16 you would suddenly be doing about 1,630 km/h.
- What if gravity doubled halfway through your ride? Twice the pull downhill, twice the load on the tyres, and a surprise about top speed.
Questions people ask
Is Bengaluru higher than Kolkata?
Yes. Bengaluru sits at about 900 m on the Deccan Plateau (Wikipedia) and Kolkata at 1.5 to 9 m on the Ganges delta (Wikipedia), so the road loses about 2,936 ft (895 m) over about 2,000 km. That is an average grade of about 0.045%, far too gentle to coast on.
Calculated in this post from the two elevations above (Bengaluru about 900 m, Kolkata taken as about 5 m) and the roughly 2,000 km route length of Sapta's November 2025 ride: 895 / 2,000,000 = 0.045%.
Does coasting in neutral save fuel on a motorcycle?
Usually not on modern fuel-injected bikes, which cut fuel almost completely on a closed throttle in gear, while idling in neutral still burns some. In neutral you also lose engine braking, so it is a poor trade on any real descent.
What slope does a motorcycle need to roll on its own?
About 1 to 1.5%, set by the tyres' rolling resistance coefficient of roughly 0.009 to 0.014. To hold 60 km/h with an upright rider it needs about 4.7%, because air drag adds roughly three times as much resistance as the tyres at that speed.
Calculated in this post, with these example inputs, from the sourced drag area (Kirca and McGordon, 2023), tyre coefficient (Wiratkasem and Pattana, 2021) and air density (standard atmosphere).
What is the steepest gradient allowed on Indian highways?
For plain and rolling terrain, IRC practice uses a ruling gradient of 3.3%, a limiting gradient of 5% and an exceptional 6.7% for short stretches. Hilly and steep terrain allow 5–8%.
Sources
- Kirca MC, McGordon A: Effect of Rider Position on the Energy Consumption of an Electric Motorcycle, University of Warwick (WRAP) (2023; source for CdA 0.35–0.59 m²)
- Wiratkasem K, Pattana S: The effect of motorcycle tyre rolling resistance coefficient on the saving of fuel consumption, Energy Reports 7 (2021; source for Crr 0.0089–0.0137)
- NASA NSSDC: Earth fact sheet (scale height 8.5 km, surface air density)
- NPTEL, Transportation Engineering, Lecture 17: Vertical alignment I (IRC gradients 3.3 / 5.0 / 6.7%)
- Wikipedia: Magnetic Hill (India) (26.5 km west of Leh on the Srinagar–Leh road; an optical illusion on a road that actually runs downhill; source for the story)
- Wikipedia: Armstrong limit (18–19 km, 6.26 kPa; oxygen above 4,500 m)
- Wikipedia: Bengaluru (elevation about 900 m)
- Wikipedia: Geography of Kolkata (elevation 1.5–9 m)
- Wikipedia: Gasoline (33.6 MJ per litre, lower heating value)
- Wikipedia: Rolling resistance (how Crr changes with load and pressure)
- ThePrint: At least 11 killed in multi-vehicle crash in Mirzapur after suspected truck brake failure (Apr 2026)
- Free Press Journal: 3 killed after wheat-laden truck's brakes fail at Bherughat (Oct 2026)

