Optimise your biomechanics with data: which quantities really count and how to derive real training decisions from AI, 3D tracking and wearables.
At pro level, the difference between good and great athletes is often the one between gut feeling and data. Modern analysis — AI-supported motion capture, 3D motion tracking, wearables — makes visible exactly the biomechanical quantities that carry your performance: the relative propulsive impulse that explains around 57% of sprint speed, your force direction and your force-velocity profile. Those who measure these quantities train the limiting factor specifically instead of generically.
Straight talk: technology doesn't replace training, but it shows you where you need to refine — and where not. This guide shows you which data actually tell you something, which only look like progress, and how to derive training decisions from them.
Not every number is worth the same. Prioritise the quantities with real explanatory value: the relative propulsive impulse (explains around 57% of sprint-speed variance), force direction/the ratio of force and ground force per contact. Vertical impulse, on the other hand, barely correlates with speed — optimising it is wasted attention. Measure what explains, not what's easy to measure.
Modern, partly markerless AI systems capture your joint angles, contact times and strike positions without a lab. With them you make visible whether your foot lands in front of the centre of mass (braking impulse) and how your posterior chain works just before ground contact — where horizontal force is created. Use these systems to objectively check force direction and strike instead of relying on feeling.
The F-V profile is your central control variable: F0 (maximal force), V0 (maximal velocity), Pmax (maximal power), derived from a sprint with splits. It tells you whether force or velocity limits you. A force-deficient profile you train with heavy resistance, a velocity-deficient one with maximal-speed runs — the stimulus thus hits the bottleneck, not the average.
Wearables deliver speed and load data from which you derive fatigue and risk. Since the hamstrings are the number-one injury and top-speed exposure means both protection and load at once, tracking helps you hit the optimum. Careful with interpretation: even seemingly strong predictors have limits — drill tests, for instance, explain only around 20 to 22% of sprint time.
Data without a decision are just numbers. Merge your sources (biomechanics, F-V profile, wearables) and read trends over weeks, not single values. For every metric, ask: what do I change if it moves? If your F-V deficit doesn't shift, the stimulus is wrong — then you change it. That's the core of data-driven training: measure, interpret, act.
Data-driven training must not tempt you into stimulus compression without recovery. Use the load data primarily to spot overfatigue early and steer the hamstring load — it remains the number-one injury. With pain or conspicuous fatigue markers, don't rely on "the numbers look okay", but reduce the load. Maximal sprint and jump tests only fully warmed up and rested.
Define an "if-then rule" for every metric in advance: if the horizontal force share rises, keep the stimulus; if the F-V deficit stagnates, change the stimulus; if reactive strength drops and resting heart rate rises, unload. That turns a data dashboard into a decision system — and you avoid the most common trap: measuring a lot, changing nothing.
The ones with real explanatory value: the relative propulsive impulse (explains around 57% of sprint speed), force direction or the ratio of force and ground force per contact. Vertical impulse, on the other hand, barely correlates with speed. So measure specifically the quantities that explain your speed instead of collecting every easily available number.
They make biomechanical quantities objectively visible without a lab: joint angles, contact times and strike positions. That way you recognise whether your foot lands in front of the centre of mass (braking impulse) and how your posterior chain works before ground contact — that's where horizontal force is created. The value lies in objective feedback instead of gut feeling, especially at pro level where cues alone no longer suffice.
You derive F0, V0 and Pmax from a sprint with splits and read your limiting factor from it. A force-deficient profile you train with heavy resistance, a velocity-deficient one with maximal-speed runs. Repeat the profile every 4 to 6 weeks under the same conditions — if your deficit doesn't move towards balance, the stimulus is wrongly chosen.
No — single values and single predictors deceive. Even seemingly strong predictors have limits: drill tests explain only around 20 to 22% of sprint time. So merge several data sources and read trends over weeks instead of single measurements. And set an action rule for every metric, or you'll collect numbers without benefit.
Leichtathletik: Sports Psychology