The Science Behind It

How the Calculator Works

The calculator has two modes: calculate from ride data, and predict climb time.

Shared inputs (both modes)

• Climb distance (km)
• Elevation gain (m)

Average gradient is derived automatically: gradient = (elevation / (distance × 1000)) × 100

Mode 1: Calculate from ride data

Additional input: Climb time (HH:MM:SS or MM:SS format)

VAM formula:

VAM (m/hr) = elevation_gain_meters / time_hours

Equivalently: VAM = (elevation_gain_meters × 60) / time_minutes

This is a pure rate calculation with no assumptions.

W/kg estimation:

The calculator uses a simplified climbing power model - gravity plus rolling resistance, with aerodynamic drag excluded:

W/kg = g × sin(θ) × v + C_rr × g × cos(θ) × v

Where:

• θ = arctan(gradient / 100)
• v = average climbing speed in m/s (distance / time)
• g = 9.8067 m/s²
• C_rr = 0.005 (fixed rolling resistance coefficient)

Aerodynamic drag is excluded because it is negligible on steep gradients (above ~8%, aero contributes under 5% of total resistance). At 5% gradient, aero drag can account for 10-15% of resistance - the calculator surfaces an info warning when gradient falls below 5% to flag this limitation.

Outputs:

• VAM (m/hr) and category rating
• Gradient (%)
• Average speed (km/h)
• Estimated W/kg

Mode 2: Predict climb time

Additional inputs:

• Sustainable W/kg (1-8)
• Rider + Bike weight (kg, 50-150)

The same simplified model is inverted to solve for speed:

speed (m/s) = W/kg / (g × (sin(θ) + C_rr × cos(θ)))

Then: time = distance / speed

And: VAM = elevation_gain / time_in_hours

Outputs:

• Predicted climb time
• VAM (m/hr) and category rating
• Gradient (%)
• Predicted speed (km/h)

VAM Categories

The calculator applies these thresholds:

Category	VAM (m/hr)
Elite Pro	>1800
Pro	1500-1800
Cat 1	1200-1500
Cat 2	1000-1200
Cat 3	800-1000
Recreational	<800

Practical Application

Example 1: Analyzing a training climb

A cyclist climbs 450m in 35 minutes over 6.4 km.

VAM = 450 / (35/60) = 450 / 0.583 = 771 m/hr  -> Recreational
gradient = (450 / 6400) * 100 = 7.0%
theta = arctan(0.070) = 3.99 degrees
v = 6.4 / 0.583 = 10.98 km/h = 3.05 m/s

W/kg = 9.8067 * sin(3.99 deg) * 3.05 + 0.005 * 9.8067 * cos(3.99 deg) * 3.05
     = 9.8067 * 0.0696 * 3.05 + 0.005 * 9.8067 * 0.9976 * 3.05
     = 2.082 + 0.149
     = 2.23 W/kg

To reach Cat 3 territory (VAM >800), the rider would need to complete the same climb in under 33:45.

Example 2: Predicting a race climb

A rider sustains 4.0 W/kg. Target climb: 1071m elevation gain, 13.8 km (Alpe d'Huez-scale).

gradient = (1071 / 13800) * 100 = 7.76%
theta = arctan(0.0776) = 4.44 degrees

speed = 4.0 / (9.8067 * (sin(4.44 deg) + 0.005 * cos(4.44 deg)))
      = 4.0 / (9.8067 * (0.07742 + 0.004987))
      = 4.0 / (9.8067 * 0.08241)
      = 4.0 / 0.8083
      = 4.95 m/s = 17.82 km/h

time = 13.8 / 17.82 = 0.775 hours = 46:30
VAM = 1071 / 0.775 = 1382 m/hr  -> Cat 1

Example 3: Comparing two climbs

Rider A: Mont Ventoux from Bedoin, 1610m gain in 1:45:00.
Rider B: Stelvio from Prato, 1808m gain in 2:05:00.

Rider A VAM = 1610 / 1.75 = 920 m/hr  -> Cat 3
Rider B VAM = 1808 / 2.083 = 868 m/hr  -> Cat 3

Rider A produced higher VAM. VAM normalizes for distance and elevation, making it a fair cross-climb comparison. Note that the Stelvio tops out at 2758m vs Ventoux at 1912m - the physiological cost of altitude is real but not captured by this model.

Why This Matters

On a steep gradient, aerodynamics recede and the dominant equation becomes power versus gravity. That is why watts per kilogram defines climbing performance.

VAM - Velocita Ascensionale Media, or average ascent speed in meters per hour - was developed by Italian sports physician Dr. Michele Ferrari as a practical way to compare climbing performances across different mountains without a power meter. You need only elevation gained and time elapsed.

For coaches and athletes without power meters, VAM provides a meaningful performance metric from basic GPS data. For those with power meters, it offers a cross-climb comparison that absolute power cannot.

The Research

Physics of cycling uphill

The foundational cycling power model was established by Martin, Milliken, Cobb, McFadden, and Coggan (1998), who validated a mathematical model predicting cycling power from measurable variables (R-squared = 0.97, standard error 2.7 watts). Their model decomposes total power into resistance components:

P_total = P_gravity + P_rolling + P_aero + P_drivetrain

Expanded:

P = m × g × sin(θ) × v + C_rr × m × g × cos(θ) × v + 0.5 × ρ × CdA × v³ + P_{drivetrain loss}

Where:

• m = total mass (rider + bike + gear), kg
• g = gravitational acceleration, 9.81 m/s²
• θ = road angle (arctan of gradient)
• v = velocity, m/s
• C_rr = rolling resistance coefficient (typically 0.004-0.008)
• ρ = air density (1.225 kg/m³ at sea level)
• CdA = aerodynamic drag area (typically 0.30-0.45 m² for road cycling)

On steep climbs (above 7%), gravity dominates. At 10% gradient and 15 km/h, gravity accounts for roughly 85-90% of total resistance, rolling resistance 5-8%, and aerodynamics 3-7%. This is why the calculator's simplified model (gravity + rolling resistance only) is a reasonable approximation for steep climbs - but becomes less accurate below 5-6%.

The W/kg estimation model used here

Calculate mode estimates W/kg from ride data using gravity + rolling resistance:

W/kg ≈ g × sin(θ) × v + C_rr × g × cos(θ) × v

This simplified model excludes aerodynamic drag, which is negligible above ~8% gradient but accounts for 10-15% of resistance at 5%.

Predict mode uses the full power balance equation including aerodynamic drag, solved with Newton's method:

P = 0.5 × CdA × ρ × v³ + C_rr × m × g × cos(θ) × v + m × g × sin(θ) × v

Using CdA = 0.4 (hoods/climbing position), ρ = 1.225 kg/m³. This prevents unrealistic speed predictions at shallow gradients where aerodynamic drag dominates.

This is not the Ferrari VAM formula. Ferrari's published approximation uses a gradient factor (2 + percent_grade/10) to convert VAM directly to W/kg. This calculator does not use that formula. Instead it applies the physics model directly to the measured speed.

VAM categories - context

Professional race data and coaching benchmarks produce these approximate W/kg ranges on a typical 8% climb:

Category	VAM (m/hr)	Approximate W/kg (8% gradient)
World Tour GC contender	1700-1900	6.1-6.8
World Tour domestique	1500-1700	5.4-6.1
Professional continental	1350-1500	4.8-5.4
Elite amateur / Cat 1	1200-1400	4.3-5.0
Strong club cyclist	1000-1200	3.6-4.3
Recreational cyclist	600-1000	2.1-3.6
Untrained rider	<600	<2.1

The W/kg ranges above use the full physics model for reference. The calculator's simplified model will give slightly lower W/kg estimates because it excludes the aero term.

Historical note: In the EPO era (mid-1990s to early 2000s), Tour de France mountain stages regularly produced VAM values exceeding 1800-1900 m/hr on major climbs. Post-biological passport values have generally decreased to 1650-1800 m/hr for stage-winning performances.

Limitations

1. Aero drag excluded: The W/kg estimate omits aerodynamic drag. At gradients above 8% this is a reasonable simplification. Below 5%, the error is 10-15% or more. In predict mode, gradients below 3% trigger a recommendation to use the Power-to-Speed calculator instead (which includes the full aerodynamic model), while gradients between 3-5% include the aerodynamic model but note that results are more reliable above 5%.

2. Variable gradients: VAM averages over the entire climb. A climb with a 10% ramp, a 2% false flat, and another 10% ramp will produce a misleadingly low VAM compared to a consistent 7% climb at the same average gradient, because the rider travels faster on the flat sections, contributing less vertical gain per unit time.

3. Altitude not modeled: Reduced air density at altitude slightly affects the physics, but the physiological cost of altitude is significant and not captured by any power estimate.

4. Wind: Headwinds and tailwinds are not accounted for but can shift power requirements by 10-20% on exposed climbs.

5. Drafting: In group riding, drafting reduces aerodynamic drag, allowing faster climbing at the same power. The effect is smaller than on flat terrain but measurable.

6. C_rr assumed constant: The calculator fixes C_rr at 0.005. Road surface, tire pressure, and tire choice can shift this between 0.003 and 0.010, affecting W/kg estimates by up to 5%.

References

Ferrari, M. (2003). Velocita Ascensionale Media and cycling performance. Published on personal website (archived).

Martin, J. C., Milliken, D. L., Cobb, J. E., McFadden, K. L., & Coggan, A. R. (1998). Validation of a mathematical model for road cycling power. Journal of Applied Biomechanics, 14(3), 276-291.

Mognoni, P., & di Prampero, P. E. (2003). Gear, inertia, and rolling resistance in road cycling. European Journal of Applied Physiology, 89(1), 47-51.

Padilla, S., Mujika, I., Cuesta, G., & Goiriena, J. J. (1999). Level ground and uphill cycling ability in professional road cycling. Medicine and Science in Sports and Exercise, 31(6), 878-885.

Olds, T. S. (2001). Modelling human locomotion: Applications to cycling. Sports Medicine, 31(7), 497-509.

di Prampero, P. E., Cortili, G., Mognoni, P., & Saibene, F. (1979). Equation of motion of a cyclist. Journal of Applied Physiology, 47(1), 201-206.

Cycling Power Lab. (2023). VAM and relative power. Retrieved from http://www.cyclingpowerlab.com/VAM.aspx

VeloViewer. (2023). VAM and relative power - what's that all about then? Retrieved from https://blog.veloviewer.com/vam-and-relative-power-whats-that-all-about-then/