The Science Behind Running Assessments: What Can Running Gait Analysis Really Tell You?

Written by: Feroz Mansoor

MSK Physiotherapist

Stride Lab London

Clinically reviewed: August 2026

Running assessments have become increasingly sophisticated.

A modern running assessment may include slow-motion video, treadmill analysis, wearable sensors, cadence measurement, strength testing and evaluation of training history.

But what does all of this data actually tell us?

Can a running assessment identify why your knee hurts after 10 km? Can it tell you whether you are overstriding? Can changing your cadence reduce load through a painful knee? And perhaps most importantly:

Can a running assessment predict whether you will become injured?

The scientific answer is more nuanced than many runners are led to believe.

Running biomechanics can provide useful information — particularly when they are interpreted alongside a runner’s symptoms, injury history, strength and training load.

However, there is no single “perfect” running style and no movement pattern that reliably identifies who will or will not become injured.

A good running assessment should therefore not simply produce a list of biomechanical “faults”.

It should answer a more useful question:

Is there anything about this runner’s strength, training or running mechanics that appears relevant to their symptoms or goals — and can we meaningfully change it?

What Is a Running Assessment?

A comprehensive running assessment looks at the runner rather than simply watching their feet on a treadmill.

Depending on the reason for assessment, it may include:

previous injury history
current symptoms
weekly running mileage
recent changes in training
running frequency and intensity
race goals
footwear
strength testing
functional movement assessment
treadmill running
slow-motion video analysis
cadence or step rate
ground-contact characteristics
stride length
asymmetry
wearable sensor data
response to changes in running technique.

This distinction is important.

Running gait analysis is one component of a running assessment — not the whole assessment.

A runner with knee pain following a rapid increase in marathon mileage may require a very different intervention from an injury-free runner trying to improve running efficiency.

Why Start With the Runner's History?

Before looking at video or sensor data, perhaps the most important question is:

What changed?

Running injuries are multifactorial.

Training volume, intensity, previous injury, recovery, strength, sleep, nutrition and biomechanics can all contribute.

One of the most consistently identified risk factors for future running injury is actually previous injury, rather than a particular foot strike or running style.

A systematic review of prospective research into running-related injury risk factors found previous injury to be the most frequently identified risk factor.

A later systematic review examining overuse injuries in short- and long-distance runners similarly identified previous running-related injury as one of the strongest risk factors in long-distance runners.

This is why a running assessment should begin with questions such as:

Have you had this injury before?
Has your mileage recently increased?
Have you introduced speed work?
Has your long run suddenly become longer?
Have you returned after illness or injury?
Has strength training changed?
When during the run do symptoms begin?

The treadmill comes afterwards.

Training Load Matters — But There Is No Magical 10% Rule

Runners are frequently told never to increase their mileage by more than 10% per week.

The science is not that simple.

A systematic review examining changes in training load and running-related injury found some evidence that sudden changes in running distance, speed or frequency may be associated with injury, but overall evidence was limited.

For example, increasing from 20 to 22 miles per week may represent a 10% increase.

But adding:

a faster track session
hills
a heavy gym session
and a longer Sunday run

at the same time could substantially increase the overall physical demand without producing a dramatic change in weekly mileage.

A good running assessment therefore considers what the runner has recently been exposed to, rather than blindly applying one percentage rule.

Knee Pain Stopping You From Running?

If your knee pain repeatedly returns when mileage increases, prevents you reaching your normal running distance or you are unsure whether the problem is runner’s knee, IT band pain or another condition, a running-specific assessment can help establish the likely cause.

Contact Stride Lab London to book a running physio assessment.

What Does Video Running Analysis Measure?

Slow-motion video allows a clinician to observe movement frame by frame.

Depending on camera position and software, measurements may include:

foot strike pattern
step rate
stride characteristics
ankle position at initial contact
knee flexion
hip position
pelvic movement
trunk position
vertical displacement.

Video analysis can therefore reveal patterns that may be difficult to appreciate when watching someone run at normal speed.

A widely cited clinical paper by Richard Souza described an evidence-based approach to video running biomechanics analysis, demonstrating how relatively accessible two-dimensional video can be used to evaluate clinically relevant running characteristics.

However, there is an important limitation.

Seeing a movement pattern does not prove that it caused an injury.

How Accurate Is 2D Running Analysis?

The most sophisticated biomechanics laboratories use three-dimensional motion capture systems.

These systems use multiple cameras and reflective markers to calculate movement in three dimensions.

Clinical running assessments more commonly use two-dimensional video because it is faster, more accessible and considerably less expensive.

How reliable is it?

A 2022 systematic review examining 2D video analysis during running concluded that reliability varied according to the measurement being assessed and suggested that categorising movement patterns may sometimes be more useful than attempting extremely precise measurements of joint angles.

A more recent systematic review and meta-analysis comparing 2D and 3D running measurements found that 2D video had generally good-to-excellent reliability for many measures, but validity compared with 3D analysis was often only low to moderate. This is particularly relevant to movements occurring outside the sagittal plane. In other words:

2D video is clinically useful — but it should not be treated as a laboratory-grade measurement of every joint angle.

Good clinical reasoning is more valuable than simply collecting more numbers.

Can Wearable Sensors Improve Running Assessment?

Increasingly, running assessments use wearable sensors containing technologies such as:

accelerometers
gyroscopes
inertial measurement units (IMUs)
pressure sensors.

These can measure variables such as:

cadence
ground-contact time
stride length
tibial acceleration
symmetry
vertical movement.

This has one major advantage.

Video primarily tells us how the runner moves.

Wearable sensors can provide additional quantitative information about when and how movement occurs over repeated strides.

A large systematic review of wearable technology for running gait analysis examined 131 studies.

IMUs and accelerometers were the most commonly studied technologies, and wearables were generally found to provide valid and reliable measures of several running-gait variables.

But again, more data does not automatically mean a better diagnosis.

A number such as:

Ground-contact time: 238 ms

has little clinical meaning by itself.

The important questions are:

Is it unusual for this runner?

Does it differ substantially between sides?

Does it change with speed or fatigue?

Does modifying it alter the runner’s symptoms?

Technology should support clinical reasoning rather than replace it.

What About Treadmill Running?

Often, some running can be maintained.

The answer depends on the severity and behaviour of the symptoms.

If you can run with:

relatively mild discomfort
no significant limp
no substantial swelling
stable rather than progressively increasing pain
symptoms that settle reasonably afterwards,

then temporarily modifying rather than completely stopping running may be appropriate.

For example, a runner whose knee begins hurting after 50 minutes might temporarily reduce running to 30–35 minutes while strength and running tolerance improve.

The objective is not to continually test how much pain you can tolerate.

It is to create a manageable training stimulus that can gradually increase.

What About Treadmill Running?

Some runners worry that treadmill assessment is irrelevant because they normally run outdoors.

There are differences between treadmill and overground running — but overall the two are more similar than many people assume.

A substantial systematic review and meta-analysis comparing treadmill and overground running biomechanics included 33 studies and 494 participants.

Most spatiotemporal, kinematic and kinetic measures were broadly comparable between conditions, although some differences were observed in variables such as foot-ground angle, knee movement and contact time.

This makes treadmill analysis a useful clinical tool, particularly because it allows:

repeated observation
consistent speed
lateral and posterior filming
immediate replay
testing of different running speeds
real-time technique modification.

Where possible, the assessment speed should resemble the pace at which the runner experiences symptoms or normally trains.

A runner who only develops knee pain at marathon pace may look very different when jogging slowly for an assessment.

Can a Running Assessment Predict Injury?

This is perhaps the most important scientific limitation.

A common marketing claim is that running analysis can identify biomechanical “faults” before they cause injury.

Current evidence does not justify that level of certainty.

A 2022 systematic review and meta-analysis of prospective studies examined more than 100 biomechanical and musculoskeletal measurements as potential predictors of running-related injuries.

Most were not consistently associated with future injury.

The authors concluded that biomechanical and musculoskeletal measurements alone should not currently be used to make injury-prevention recommendations for non-elite runners.

This is fundamental.

A runner may display:

pelvic drop
knee valgus
rearfoot strike
low cadence
asymmetry

and remain completely injury-free.

Another runner may look biomechanically “excellent” but develop an injury following a large training increase.

A running assessment should therefore not be sold as a crystal ball for predicting injury.

Its value lies elsewhere.

Where Running Assessment Becomes Particularly Useful: The Injured Runner

Running biomechanics become much more meaningful when combined with a specific clinical problem.

Imagine a runner with patellofemoral knee pain.

They report:

pain only when running
symptoms starting after approximately 40 minutes
no pain with normal walking
discomfort during longer downhill runs.

During assessment, increasing step rate slightly reduces pain and changes knee loading.

Now the biomechanical information has clinical context.

It is not:

“Your cadence is wrong.”

It is:

“Changing this variable appears to reduce the load associated with your symptoms.”

That is a much more defensible use of running analysis.

Can Changing Running Technique Reduce Load?

Yes — certain running characteristics can be modified.

Perhaps the best-studied example is step rate or cadence.

A systematic review and meta-analysis examining changes in running step rate found that increasing cadence generally reduced or did not increase several loading-related variables around the ankle, knee and hip.

A modest cadence increase usually shortens stride length and places the foot closer to the body’s centre of mass.

However, this does not mean everyone should run at 180 steps per minute.

There is no universally optimal cadence.

For some runners, a 5–10% increase from their natural cadence may be trialled where there is a clinically relevant reason.

The key word is individualised.

Does Gait Retraining Actually Work?

There is growing evidence that running technique can be modified.

A systematic review and meta-analysis of gait retraining in distance runners examined 19 trials involving 673 runners.

Gait retraining interventions were able to change variables including:

step rate
knee kinematics
vertical loading rates.

Importantly, these changes did not appear to impair running performance overall.

The evidence for reducing actual injuries remains less certain.

However, in symptomatic populations there is growing evidence that targeted changes can help selected runners.

For example, a 2026 umbrella review examining gait retraining for runners with patellofemoral pain found the most consistent evidence supported modest cadence increases and softer-landing strategies for short-term improvements in certain biomechanical measures, pain and function.

The evidence for long-term injury prevention remains much weaker.

This reinforces an important principle:

Change running technique because there is a specific clinical reason — not simply because someone’s technique looks different.

Should Everyone Change From Heel Strike to Forefoot Strike?

No.

Foot strike is one of the easiest characteristics to identify on video and therefore attracts considerable attention.

But intentionally changing from a rearfoot strike to a forefoot strike changes where load is distributed rather than simply removing load.

Generally, a forefoot strike can shift greater demand towards the ankle and calf while reducing some loading around the knee.

A systematic review and meta-analysis examining the consequences of changing foot strike found that different strike patterns alter biomechanics, but evidence does not support universally converting runners to a particular foot strike for injury prevention or performance.

This is why:

heel strike ≠ bad running.

For one runner, changing foot strike may have a clinical rationale.

For another, it may simply transfer stress from the knee towards the calf or Achilles tendon.

What Does Strength Testing Add?

Running requires repeated force production from:

calves
quadriceps
hamstrings
gluteals
trunk musculature.

Strength testing can identify whether a runner has difficulty producing or tolerating force in certain movements.

Useful tests may include:

calf raises
knee-extension strength
split squats
single-leg squat
hopping
hip strength
single-leg functional testing.

But strength measurements must also be interpreted carefully.

The same prospective systematic review of biomechanical and musculoskeletal injury predictors found that physical measurements generally performed poorly as standalone injury predictors.

So a runner who demonstrates weakness should not automatically be told:

“This weakness caused your injury.”

A better interpretation is:

“This is a modifiable capacity that may be useful to improve given the demands of your running.”

Can Running Assessment Improve Performance?

Potentially — but the relationship between running technique and performance is more complex than simply correcting a few visible “faults”.

One important concept is running economy.

Running economy describes how much oxygen or energy a runner requires to maintain a given submaximal speed.

Two runners may have very similar VO₂ max values but perform differently because one uses less energy to maintain the same pace.

Biomechanics can contribute to running economy, but there is no single running style that guarantees better economy.

A systematic review and meta-analysis examining running biomechanics and running economy analysed more than 50 studies involving over 1,100 participants.

The review found that some biomechanical variables were associated with running economy, but many commonly discussed measures showed weak or inconsistent relationships when considered individually.

For example, variables such as:

cadence
vertical oscillation
leg stiffness
contact characteristics
stride mechanics

may influence economy, but no single factor explains performance.

This is clinically important.

A runner should therefore be cautious if told:

“Changing your running form will automatically make you faster.”

Performance is influenced by many factors including:

aerobic capacity
lactate threshold
running economy
training history
strength
fatigue resistance
body mass
- recovery
- pacing
- psychology.

However, running assessment can still contribute to performance optimisation.

For example, it may identify:

Excessive Braking

A runner whose foot lands substantially ahead of their centre of mass may demonstrate greater braking forces.

A small technique change or cadence adjustment may sometimes reduce this braking demand.

Unnecessary Vertical Movement

Some runners move more vertically than is necessary for their running speed.

Reducing excessive vertical oscillation in selected runners may theoretically reduce wasted movement, although this should not be treated as a universal correction.

Cadence and Stride Characteristics

A runner may benefit from testing whether a modest cadence adjustment produ

ces:

smoother running
reduced braking
reduced symptoms
improved comfort at race pace.

Fatigue-Related Changes

A runner may look very different after 20 minutes compared with the first two minutes.

Performance assessments may therefore become particularly useful when testing:

marathon pace
threshold pace
faster running
running under fatigue.

The goal is not simply to create aesthetically “perfect” movement.

It is to determine whether a change:

reduces unnecessary loading, improves comfort or potentially improves running efficiency without creating problems elsewhere.

Importantly, any biomechanical modification should also be assessed against performance.

A technically different stride is not automatically better if it substantially increases perceived effort or energy cost.

Good performance-focused running assessment therefore involves testing rather than assuming.

Can Running Assessment Help With Fatigue?

Potentially.

Running mechanics do not remain completely constant as a runner becomes fatigued.

Some runners may demonstrate changes in:

cadence
ground-contact time
trunk position
hip control
knee movement
stride length.

This can be relevant because many running injuries do not appear in the first ten minutes of a session.

They appear after:

45 minutes

10 kilometres

during the final third of a long run.

Assessing only the first minute of comfortable treadmill running may therefore miss the context in which symptoms normally occur.

Where practical, assessment should try to reproduce:

the relevant pace
sufficient running duration
the circumstances in which symptoms usually develop.

Is Asymmetry Always a Problem?

No.

Human beings are not perfectly symmetrical.

Differences may exist between the right and left side in:

strength
ground-contact time
step length
joint angles
loading.

Finding an asymmetry does not automatically mean something needs correcting.

The important questions are:

How large is the difference?

Is it consistent?

Does it relate to symptoms?

Does it persist at different speeds?

Is there a plausible reason to modify it?

A runner who has been asymmetrical for years without injury does not necessarily need to be made perfectly symmetrical.

Again:

measurement requires interpretation.

Can Running Assessment Help Choose Running Shoes?

Potentially, but running assessments should not be presented as a perfect shoe-prescription system.

Traditional shoe selection often focused heavily on foot posture and pronation.

The scientific relationship between pronation, footwear and injury is more complicated than the traditional:

flat foot → stability shoe

model.

Comfort, previous experience, symptoms, training demands and personal preference all matter.

Running assessment can still help by observing how a runner responds to different footwear where relevant.

For example:

Does the shoe feel comfortable?
Does pain change?
Does it substantially alter stride characteristics?
Is it appropriate for the runner’s training and race goals?

But choosing trainers should not be reduced to one visual measurement of pronation.

What Should a Good Running Assessment Actually Do?

A scientifically grounded running assessment should combine several pieces of information.

1. Understand the Runner

Why are they attending?

Injury?
Recurring niggle?
Return to running?
Marathon preparation?
Performance?
General curiosity about their running mechanics?

The goal determines the assessment.

2. Review Training

Mileage, frequency, pace, hills, long runs, intervals and recent changes should all be considered.

3. Assess Clinically Relevant Capacity

Strength and functional testing should be selected according to the runner rather than performed as a generic checklist.

4. Observe Running

Video and/or sensor technology can identify potentially relevant characteristics.

5. Test Hypotheses

If something appears relevant, modify it.

For example:

increase cadence by 5%

then reassess:

symptoms
stride characteristics
loading
comfort
perceived effort.

Consider Performance

For performance-focused runners, determine whether a modification actually improves the desired outcome rather than simply making the running pattern look different.

7. Create an Actionable Plan

The runner should leave understanding:

what appears relevant
what probably does not matter
what to work on
how to progress training
whether running technique needs changing at all.

What a Running Assessment Should NOT Do

A running assessment should not automatically tell every runner:

“Your glutes aren’t firing.”

“You pronate too much.”

“You heel strike.”

“Your pelvis drops.”

“Your cadence is too low.”

“You’re going to get injured.”

Those statements imply a level of certainty the scientific evidence does not support.

Running injuries are multifactorial.

Training, previous injury, tissue capacity, health, recovery and biomechanics can all contribute.

The best running assessment therefore combines technology with clinical reasoning.

Who Might Benefit From a Running Assessment?

Running assessment may be particularly useful if:

pain only occurs while running
pain starts at a predictable distance
injuries repeatedly return when mileage increases
you have recurrent knee, Achilles, calf, hip or shin symptoms
you are returning after injury
you are increasing mileage for a marathon
you want objective information about your running mechanics
previous rehabilitation has not restored comfortable running
you want to investigate running economy or technique
symptoms only occur at particular running speeds.

It can also be useful for an injury-free runner who simply wants to better understand their running.

But an ethical assessment should be willing to say:

“There is nothing here that needs fixing.”

Not every unusual running style needs correcting.

Running Assessments: The Key Message

The science behind running assessment is both more interesting — and more cautious — than the idea of simply identifying a “bad running technique”.

Video analysis can reliably assess several clinically useful running characteristics.

Wearable sensors can provide objective information on variables such as cadence, ground-contact time and stride characteristics.

Running gait can also be deliberately modified, and changes such as modest increases in cadence can alter lower-limb loading.

Running assessments may also have value for performance, particularly when examining running economy, braking, cadence, fatigue-related changes and whether technique modifications improve comfort or efficiency.

But biomechanics should never be interpreted in isolation.

Research does not currently support using a running assessment to reliably predict who will become injured.

Instead, its greatest value is combining:

the runner’s story

training history

strength and physical capacity

running biomechanics

symptom response

performance goals

to develop an individual plan.

The goal is not to make every runner look the same.

It is to identify whether anything measurable and modifiable appears relevant to that runner’s injury, training or performance goals.

Running Assessment in London

At Stride Lab London, our running assessments combine clinical reasoning with objective running data.

Depending on your needs, an assessment may include:

injury and training-history review
strength and functional assessment
treadmill running
video running analysis
sensor-based running data
cadence and stride characteristics
assessment of relevant asymmetries
real-time testing of running modifications
rehabilitation recommendations
performance-focused recommendations where appropriate.

Whether you are dealing with a recurring running injury, returning after rehabilitation, preparing for a race or simply want to better understand your running mechanics, the aim is to provide useful information you can actually apply to your training.

Learn more about the Stride Lab London Running Assessment

Contact Stride Lab London to arrange your assessment.

This article is intended for general educational purposes and does not replace individual medical or physiotherapy assessment. Seek urgent medical assessment for significant or progressive neurological symptoms, changes in bladder or bowel function, saddle numbness, significant trauma or other concerning symptoms.