Assisted sprinting: what overspeed towing actually changes

The paper: Gleadhill, S., Jiménez-Reyes, P., van den Tillaar, R., & Nagahara, R. (2024). Comparison of kinematics and kinetics between unassisted and assisted maximum speed sprinting. Journal of Sports Sciences, 41(24), 2169–2175 (open access). doi.org/10.1080/02640414.2024.2314866

Assisted, or “overspeed,” sprinting uses a tow to pull an athlete beyond the speed they can reach unaided, on the theory that training above normal top speed teaches the body to move faster. This study placed that theory on a force platform and asked a direct question: when the tow makes an athlete faster, is the athlete producing the extra speed, or is the rope?

The study

Fifteen sub-elite sprinters — regional to national level, 100 m bests around 11.3 s — performed maximal 50 m sprints in two conditions: unassisted, and assisted by a 5 kg pull from a robotic towing device. Both were run over a 54-plate force runway sampling at 1000 Hz, so every step’s forces were measured directly rather than inferred.

The analysis compared the top-speed phase of each condition. To separate the effect of the tow from the effect of simply running faster, it also compared the two conditions at matched running speeds.

What they found

The tow raised top speed by roughly 9%, through longer steps and shorter ground contact. Step length increased and contact time shortened, with a small rise in step rate. On the surface, an unambiguous improvement.

Beneath it, the athlete’s own force production moved the wrong way. Propulsive force fell — mean force by 4.3%, impulse by 12.4% — while braking force rose sharply, by 34%. The net horizontal force, propulsion minus braking, reversed: from driving the athlete forward to, on balance, holding them back. The added speed came from the tow, not from the athlete generating more force.

Change from unassisted to assisted running at top speed (values from the paper’s Table 1). Speed rises through longer steps and shorter contact, but the athlete’s own propulsive force falls and braking rises — the net horizontal force reverses, from +0.82 to −0.10 N/kg.

The only variable that correlated with an individual’s speed increase was contact time: the athletes who shortened their ground contact the most gained the most speed.

Strengths and limitations

Strengths

• Forces measured directly on a 54-plate runway, during genuine assisted running, with a device that controls the tow precisely.
• A matched-speed comparison — a real attempt to separate the pull of the tow from the act of running faster.
• The inconvenient result is reported plainly: the athletes’ own propulsion fell rather than rose.

Limitations

• The study is acute. It measures what happens during a towed repetition, not whether a block of assisted training improves unaided sprinting — there is no follow-up.
• Fifteen sub-elite sprinters is a small sample.
• Only one tow load (5 kg) was tested, so there is no dose-response — a lighter or heavier pull may behave differently.
• The matched-speed comparison confounds sprint phase: the assisted trials reached the matched speed early, still accelerating (around 20 m), while the control top speed occurred later (around 40 m), so part of that difference reflects acceleration versus top-speed running rather than the tow itself.

Practical takeaways

• The extra speed is the tow’s, not the athlete’s. During an assisted repetition propulsive force falls and braking rises, so a faster time under tow is not evidence that the athlete is producing more force.
• Its value, if any, is exposure and technique — not force development. Reaching a higher speed may let an athlete rehearse faster turnover and shorter contacts; in this data it does not build the force behind them.
• Watch contact time. Shortening ground contact was the one change that tracked with a genuine speed increase, so it is the variable to cue and monitor if assisted work is used.
• Treat it as experimental. The acute picture is now clear; whether assisted training transfers to faster unaided sprinting is not, so it warrants caution rather than wholesale adoption.