Your health comes first. This educational article provides general information. Discuss your individual circumstances, pregnancy, illness or symptoms with your clinician; the article cannot replace an individual assessment.

One run has a lower maximum ground-reaction force, yet force rises more steeply after landing. A contradiction? No: peak force and loading rate answer different questions. The first describes the magnitude of a value, the second how force increases over time. Calling both simply ‘impact’ can easily hide the essential distinction.

In a running-laboratory report, sensor app or shoe study, therefore, first look for the metric’s definition. Which force was measured, during which interval and with what processing? The following explanation teaches you to read a force–time curve. The original numerical examples do not provide targets for your technique; they show how several different metrics can describe the same movement.

Peak force is a height; rate is a slope

A force–time graph shows time on its horizontal axis and force on its vertical axis. The peak is the largest value within the selected section. Rate describes how much force changes per unit of time. If we compared the curve to a staircase, the peak would be its height and the slope would describe the steepness of the ascent. The same height can be reached along a gentler or steeper route.

An original simple example: force increases from 200 to 1200 newtons over a selected interval imagined as a straight line. The change is 1000 newtons. If it takes 0.020 seconds, the average rate of increase is 1000/0.020 = 50,000 newtons per second. If the same change takes 0.040 seconds, the rate is 25,000 newtons per second. The final force is identical, while its average speed of increase is halved.

A real curve is not necessarily straight. The average slope between the endpoints of a selected window and the steepest local slope within it are different numbers. A short steep section can yield a high instantaneous rate while the wider section’s average is lower. Alongside ‘loading rate’, therefore, look for ‘average’ and ‘instantaneous’ or ‘maximum’.

Which peak are we talking about?

The full ground-contact force curve may contain an early local rise and a later larger peak. Not every curve has a clearly separate early rise. If a calculation defines the rising interval relative to that early point, it must use some rule to handle its absence. If two analyses select different peaks or estimate their location differently, they may summarise the same raw curve differently.

An original study comparing four loading-rate calculations treated the average and maximum slopes associated with an impact peak separately from rates calculated over a specified portion of stance time. Both method and running speed influenced the values. The particular calculation window is therefore part of the data’s meaning, not a software setting that can later be omitted.

When comparing two reports, add a short explanation beside the peak name: the largest vertical force over the whole stance, an early local peak or the maximum of a predefined interval. ‘Peak’ can appear in all three cases. Likewise, interpreting rate requires knowing which force range or percentage of stance time supplied the calculation. Precise naming directly affects the comparison here.

A runner undergoes a treadmill measurement wearing a respiratory-gas analysis mask under professional observation in a sports-physiology laboratory.Enlarge image
AI-generated editorial illustration of laboratory assessment of running economy; not a documentary photograph.

A smaller peak can accompany a higher rate

Imagine two fictional rising intervals. In case A, force rises from 200 to 1400 newtons in 0.040 seconds: a change of 1200 newtons and an average rate of 30,000 newtons per second. In case B, it rises from 200 to 1200 newtons in 0.020 seconds: the change is only 1000 newtons, yet the rate is 50,000 newtons per second. B has a lower endpoint but a steeper rise.

Original illustrative case Force change Duration Average rate
A 1200 N 0.040 s 30,000 N/s
B 1000 N 0.020 s 50,000 N/s

We have not called either case in the table healthier. We have not even modelled a complete running step, only one selected rising interval. The example shows why a smaller peak does not automatically imply a lower rate. Equally, a reduced rate does not reveal the full force curve’s shape, stance duration or internal muscle-force magnitudes.

If both metrics change after new shoes or a technique instruction, first state the observations separately. ‘The selected peak became lower, and the early loading rate became higher.’ That is more precise than one overall statement that ‘my running became softer’. Further interpretation also requires speed, measurement settings and other movement data.

Newtons, body-weight ratios and seconds: align the units too

Force can be expressed in newtons or relative to the body’s weight force. Body mass in kilograms and weight force in newtons are different quantities: in a simple calculation near Earth’s surface, weight force is m × g. For a 70-kilogram person, using approximately 9.81 metres per second squared gives 686.7 newtons. An external force of 1500 newtons is about 2.18 times that. This is merely normalisation, not a clinical threshold.

Dividing a rate of 40,000 newtons per second by the same weight force gives approximately 58.3 body weights per second. The same absolute force may correspond to a different relative value in another person. If one report uses N/s and another BW/s, the raw numbers cannot be compared directly. BW means force has been expressed relative to the selected body-weight reference.

N/kg/s is another unit again: it uses mass in kilograms as the divisor. Converting to BW/s also introduces the factor of gravitational acceleration. A graph’s axis label is therefore not decoration. If absent or unclear, ask about it first rather than placing the number within another study’s range. A normalised metric supports comparison but does not replace aligning speed and method.

Signal processing also shapes the slope

Rate divides a force difference by a time difference. In a short window, even a small signal fluctuation can produce a larger change in the quotient. In an original example, a difference of 10 newtons over 0.001 seconds means 10,000 N/s, while over 0.010 seconds it means 1000 N/s. This does not describe that instrument’s error; it illustrates why slope is sensitive to how data are handled over time.

Filtering can damp some rapid fluctuations, but in doing so may also change brief curve details and slopes calculated from them. An original sensor study estimated peak force and contact time from a sacral inertial sensor in thirty collegiate cross-country runners; the chosen low-pass filtering influenced the relationship with reference values. This is a finding about that particular estimation procedure, not a prescription for selecting a filtering frequency for every sensor.

Sampling frequency and filter cutoff frequency are separate settings. One indicates how frequently the signal is recorded; the other concerns which fluctuations over time processing retains or dampens. ‘High sampling’ therefore does not automatically mean two systems report the same rate. The signal threshold, event selection and calculation window also belong in the methods description.

Ground-reaction force is external; tissue loading is a separate question

A force-measuring surface supplies data on the interaction between ground and runner. Understanding internal forces in bone, tendon or joint surfaces also requires movement, geometry and muscle effects. The official OpenSim documentation likewise distinguishes the model’s calculated net intersegmental force from the force between joint surfaces. One external force peak therefore cannot directly reveal a particular tissue’s full loading.

The original tibial model investigating step width connected movement and force data to further biomechanical calculations to estimate internal stresses. External force was one input to the analysis. The final result also required the model’s structure and assumptions. This illustrates how many further steps separate a curve measured at the floor from an estimated quantity inside a bone.

For an original thought experiment, imagine pushing a door with the same force at the handle or near the hinge: the turning effect differs. The body is more complex, but this still illustrates the relationship between force and geometry. External force, moment and internal tissue estimates therefore deserve separate lines on a runner’s report. They should not all be grouped under ‘joint impact’.

During a consultation, a runner indicates the precise location of lower-leg symptoms with the leg supported, without demonstrating an examination or treatment method.Enlarge image
AI-generated editorial illustration; not a documentary photograph.

As one metric improves, work may shift elsewhere

Several joints in the movement system perform the task together. Cigoja and colleagues studied thirteen recreational male runners using a version of the same shoe modified with an insert to increase bending stiffness. Positive joint work changed differently at the knee and metatarsophalangeal joints. This does not establish an identical effect for every carbon-plated shoe; it is a specific example of joint-work redistribution.

If reducing one number is your goal, it is easy to overlook what happens to other quantities. Running speed, stance duration, muscle activity and load distribution also belong to the task. A smaller number is not always a better number in isolation. Make clear beforehand which practical question and observation the change concerns.

Loading rate’s role in predicting injury is not self-evident either. In the previously mentioned study examining prospective data from 79 collegiate cross-country runners, baseline measured rate showed no significant association with later injuries using any of the calculation methods. The study concerned a defined group of athletes. This is sufficient reason not to make a personal injury prediction from one rate, but it does not claim that running mechanics are generally irrelevant.

A good results sheet explains how the number was produced

When viewing laboratory or app results, first identify the source: direct force-measurement data, a value estimated from a wearable sensor or a model calculated from video? Next look at units and the calculation interval. Finally check the running situation: were speed, surface and shoes the same when the comparison was made? These three checks resolve many misunderstandings before conclusions are drawn.

  • Peak: the largest value of which force, over which interval?
  • Rate: average or maximum slope, from which time window?
  • Unit: N, N/s, BW/s or data normalised to mass?
  • Processing: which sampling, filtering and event definition?
  • Interpretation: does it report external force or an internal quantity from a further model?

Peak force and rate become useful when we stop making them compete to describe the same truth. They show different properties of the curve. With precise names, identical measurement conditions and traceable calculations, both can contribute to understanding running movement. A decision attached to a number is well-founded only when it also considers the whole task, the runner’s goal and other relevant data.

Sources

Helpful? Pass it on.