Sprint Force Model: The Physics of Running

1. Running as a series of collisions

Unlike a car, a sprinter has no continuous engine pushing them forward. Instead, each foot strike is a brief, violent impulsive contact with the ground lasting only 80–100 ms during the drive phase. Newton's second law still governs everything:

F_GRF,x is the horizontal component of the ground reaction force (GRF) — the force the track pushes back on the athlete's foot, by Newton's third law, in reaction to the athlete pushing backward and down against it. Everything a sprinter does technically — arm drive, forward lean, stride frequency — exists to maximize the horizontal component of this force while it's available.

2. Ground reaction force

During stance, the leg acts like a stiff spring (the spring-mass model of running). The vertical GRF rises and falls roughly sinusoidally over the contact time t_c, while the horizontal GRF flips from braking (negative, early stance) to propulsive (positive, late stance):

The net horizontal impulse per stride (area under the F_x(t) curve) determines the change in velocity per step. Elite sprinters minimize the braking impulse in early stance by landing with the foot nearly under the hip (rather than reaching far forward), which reduces the negative-impulse penalty.

3. Hill's force-velocity muscle model

A muscle cannot produce arbitrarily large force at arbitrarily high contraction speed — this is the single biggest constraint on sprint performance. A. V. Hill's 1938 hyperbolic relation still underlies modern sprint models:

As shortening velocity v increases, available force F drops hyperbolically. At the theoretical maximum velocity, F → 0 — the muscle can no longer produce net propulsive force at all, which is exactly why a sprinter's speed plateaus rather than growing indefinitely.

4. Acceleration, max-velocity and deceleration phases

A 100 m race decomposes into three biomechanically distinct phases, each governed by a different balance of forces:

• Drive phase (0–30 m): Large forward lean (~45° off the blocks), long ground contact, near-maximal horizontal force ratio (F_x/F_total). Acceleration is highest here because velocity — and therefore drag and Hill's v-penalty — are still low.
• Max-velocity phase (30–60 m): Torso upright, contact time shortest, stride frequency near its ceiling. Acceleration → 0 as propulsive GRF impulse balances drag and the Hill-model force deficit at high v.
• Speed-maintenance / deceleration (60–100 m): Neuromuscular fatigue reduces peak force faster than technique can compensate; even elite sprinters lose 0.3–0.5 m/s of top speed in the final 20–30 m of a 100 m race.

5. Why the block angle isn't 45°

For a projectile in a vacuum, 45° maximizes range — a fact often (wrongly) generalized to sprinting. A sprinter is not a projectile: they stay in continuous contact with the ground and re-apply force every stride, so the relevant optimization is entirely different.

What matters at the start is maximizing the horizontal component of force while keeping enough vertical impulse to clear the next stride. Biomechanics studies consistently find optimal block/shin angles around 42–45° from horizontal for the front leg and a steeper rear leg angle, but the underlying logic is force direction, not projectile range — a coincidence of the numbers, not the same physics.

6. Air resistance at sprint speed

At sprint speeds (up to ~12 m/s for elite males), aerodynamic drag is small but not negligible — it's estimated to cost a sprinter roughly 3–6% of maximum velocity. The familiar drag equation applies directly:

This is exactly the same drag equation used for projectiles and vehicles — see our companion article on drag coefficients and streamlined shapes for how C_d is measured and why body posture changes it. At v = 10 m/s, F_D ≈ 0.5 · 0.9 · 1.225 · 0.45 · 100 ≈ 24.8 N — small compared to peak GRF (2,000–3,500 N), but it accumulates over the whole race and explains why tailwinds above +2.0 m/s disqualify sprint records.

7. A simple numerical sprint model

Combining the impulse model, a simplified Hill-style velocity cap, and drag gives a compact ODE that reproduces real 100 m splits surprisingly well:

function stepSprint(state, dt):
  // state: { v, x } — velocity, distance

  // 1. Propulsive force falls off as v approaches v_max (Hill-style)
  f_prop = F0 * (1 - state.v / v_max)

  // 2. Aerodynamic drag opposes motion
  f_drag = 0.5 * Cd * rho * A * state.v ** 2

  // 3. Net force → acceleration
  a = (f_prop - f_drag) / mass

  // 4. Semi-implicit Euler integration
  state.v += a * dt
  state.x += state.v * dt

  return state

With F0 ≈ 820 N, v_max ≈ 11.8 m/s, mass ≈ 82 kg, this simple model reproduces a 100 m time within a few tenths of a second of world-class performances — a good demonstration of how far basic Newtonian mechanics gets you before you need full musculoskeletal modeling.