Run-up, drive step, dive: how to improve your skeleton push start through metrics instead of feel — with target values, video analysis and load control.
If your start has stalled, it is almost never a lack of strength — it is that you treat it as one single sprint instead of three separate jobs: the drive step, the acceleration steps and the dive. Biomechanically the skeleton start is a hybrid of sprint acceleration and a deliberate leap onto the sled — and it is exactly at that seam that most athletes lose their time [1].
The measurements make it tangible. The drive step is markedly shorter at roughly 1.60 metres than the following step at about 2.09 metres [1] — nearly half a metre of difference back to back. That is not a flaw, it is the pattern: you push short and hard, then you open up. Make that first step too long and you kill precisely the impulse you are chasing.
And impulse is the right word here. Measurements show maximum horizontal forces around 827 newtons and a mean horizontal impulse of roughly 264 newton-seconds [1]. Impulse is force times time — so it is not only how hard you push, but how long you stay in contact with the ice while pushing.
The honest framing: a faster push guarantees you nothing at the finish. The relationship between push time and finish time is strongly track-dependent, ranging from roughly 0.57 to −0.14 [5]. You are working here on the phase you can control — not on the whole run.
Record three things: push time over the official start section, splits at 5, 10 and 20 metres, and a side-on video [2][4]. Only once you can see exactly where you lose time — drive step, middle section or dive — does the guessing stop. Without splits you optimise a number you cannot decompose.
The first step should not be long. The reference from start analysis: roughly 1.60 metres in the drive step against 2.09 metres in the following step [1]. Train it with marked start lines on the ground and check on video whether in the first two steps you really push under your centre of mass rather than reaching ahead of it.
The decisive point for the skeleton start is holding maximal power at the highest sustainable step rate — which is why diagnostics record the decline in power over ten seconds alongside sprint and jump values [4]. In practice: if your splits collapse between 10 and 20 metres, your problem is not maximum strength but your ability to hold power over time.
The dive — the deliberate leap onto the sled — is a movement in its own right, not a continuation of the sprint. You go from a full run into the prone position without decelerating the sled [3]. Break it down: hand position first, then take-off, then settling. Film every repetition from the side and look for the moment the sled loses speed [2].
Compared with unresisted sprinting, resisted sprints with a sled or belt demonstrably improve force, power and times over 5, 10 and 20 metres [6]. Heavy towing at around 80 percent of body weight markedly increased horizontal ground reaction forces in studies [6]. The optimal load remains scientifically contested [6] — so work with a load range rather than a magic number, and judge the effect by your splits.
The push section is the part of skeleton you can train under control — the dive at full speed is not. Practise it exclusively on purpose-built facilities and under your coach's supervision, with ramping speed instead of instant full throttle.
And the boundary stays clear: this guide covers run-up and dive, not driving in the ice channel. For line, cornering and steering there is deliberately no how-to here — you learn that exclusively with the federation, a club and qualified coaches on the ice. Also watch your hamstrings: maximal sprints under added load are a classic trigger for muscle injuries when the warm-up is cut short.
Film your start once per training block from the exact same position — same height, same distance, same frame rate. Markerless motion analysis works in the field precisely because it receives comparable footage [2]. Overlay the videos and look for exactly one difference per block. One detail changed cleanly beats five building sites at once — and you will see it in your splits before you feel it.
A short, powerful drive step of roughly 1.60 metres, a markedly longer following step around 2.09 metres, and a dive that does not decelerate the sled [1][3]. Behind that sit high horizontal forces around 827 newtons and a horizontal impulse around 264 newton-seconds — force that acts long enough [1].
The run-up covers roughly 30 to 40 metres in spiked shoes, followed by the dive onto the sled; the first 50 metres or so take about five seconds at over 40 km/h [3]. The exact take-off point is individual and is set with your coach through video control [2].
Through push time plus splits at 5, 10 and 20 metres from a three-point start [4] and a high frame-rate side-on video. Markerless motion analysis captures step rates, angles and trunk position directly in the field, making technical change demonstrable rather than merely perceptible [2].
More important for your confidence than for the clock. The correlation between push time and finish time is strongly track-dependent and ranges from roughly 0.57 to −0.14 [5]. A fast start puts you in the race, but track knowledge and line decide at least as much.
Yes. Compared with unresisted sprinting, resisted sprints improve force, power and times over 5, 10 and 20 metres, and heavy loads around 80 percent of body weight markedly increase horizontal ground reaction forces [6]. The optimal load remains contested, though — work with a range and check the effect against your splits [6].
Starting Without Ice: Push Simulation, Metrics and What Really Transfers