Knees Over Toes / ATG: Every Claim Checked

“Knees must not go past the toes” was, for decades, one of the best-known technique rules in the gym.

Ben Patrick — KneesOverToesGuy — made the counter-argument popular: knee travel is not inherently dangerous, deep joint angles can be trained, and robustness should be built up step by step rather than avoided wholesale.

In many respects, that core idea fits the modern training and rehab literature surprisingly well.

It does not mean that every single ATG exercise, and every health claim derived from it, is equally well supported.

Which is exactly why it is worth going through this claim by claim.

Who is Ben Patrick — and what is ATG?

Ben Patrick is a US coach and the founder of ATG. He became known above all through videos on deep split squats, backward walking and sled work, tibialis raises, and training with a lot of knee travel.

His system today is broader than just “knees over toes” and includes upper-body, mobility and athletic work as well.

For a scientific assessment, though, something else is decisive:

A brand is not an intervention.

Research mostly examines individual exercises or principles. If a study shows that Nordic hamstring curls reduce injury rates, that demonstrates the value of that exercise — not automatically the value of every other ATG component.

So in what follows, the overall marketing gets separated from the evidence for each individual piece.

The core claims — checked against the evidence

1. “Knees past the toes is inherently dangerous” — there is no good basis for this

Knee travel changes the biomechanics of a squat.

When the knees come further forward, the knee extensor moment typically rises. So the quadriceps and the knee joint take on more of the work.

That is not the same thing as:

more load = damage.

A classic study by Fry et al. (2003) compared squats with normal knee travel against a variation in which the knees were artificially blocked from travelling forward. Restricting them lowered the knee moment, but raised the hip moment very substantially.

Newer biomechanical work shows the same basic principle: knee, hip and trunk position distribute the demands differently.

So the sensible conclusion is:

Knee travel is a loading variable, not a red line.

A healthy lifter has no need to keep their knees artificially behind the toes. With certain complaints or rehab goals, on the other hand, it can make a lot of sense to steer knee loading through technique and ROM for a while.

Which is exactly why individual squat technique matters more than a universal toe line.

2. Deep squats are not automatically worse for the knees

Here too, the popular rule is too crude.

As knee flexion increases, shear, compressive and patellofemoral forces change. Those loads do not simply rise in a straight line forever: contact areas and the distribution of the forces change as well.

The review by Hartmann et al. (2013) concluded that deep squats, performed with appropriate technique and progressive loading, do not in principle represent an increased injury risk for the passive structures of the knee joint.

That does not mean:

The deeper, the healthier.

And it does not mean:

Everyone has to squat ATG.

Deeper squats can train a larger range of motion and support muscle development and strength across that range. At the same time, the depth has to suit your anatomy, your goal, what you can tolerate and the exercise variation.

“Deep squats are dangerous” is too absolute.

“Deep squats are always better” would be too.

3. Nordic hamstring curls really do have a strong evidence base

The Nordic hamstring curl is one of the best-studied components that appear in the ATG context.

A large meta-analysis by van Dyk et al. (2019) covered 8,459 athletes. Prevention programs that included the Nordic hamstring exercise cut the hamstring injury rate roughly in half compared with programs without Nordics.

An earlier meta-analysis by Al Attar et al. arrived at a similar result.

The precise statement matters:

Programs containing Nordic hamstring exercises reduce hamstring injuries.

Not:

• Nordics prevent every injury.
• Everyone needs them.
• They prove the ATG system as a whole.

For sports with a high sprinting load, though, they are without question a very well supported tool.

4. Backward walking has rehab evidence — reverse sleds are not the same thing

ATG frequently uses backward walking, and reverse sled work in particular, as a low-threshold way to load the knee.

For backward walking there really is clinical research.

A 2025 meta-analysis pooled 13 randomised trials in people with knee osteoarthritis. Backward walking on top of conventional rehabilitation improved pain and function substantially compared with the respective control conditions.

That is an interesting result.

But a transfer creeps in here that the studies did not test directly:

Backward walking ≠ a heavy reverse sled.

The movements share properties, but load, speed and mechanical demands differ.

So the clean statement would be:

• Backward walking is promising as a rehab component in knee osteoarthritis and is by now reasonably well studied.
• Reverse sleds are a plausible, robust training progression, but their specific effect on knee pain has been studied far less directly.

5. Tibialis raises are logical — but no proven knee or shin splint cure

The tibialis anterior lifts the foot and works actively during walking and running.

Training it is, in principle, a normal strength training decision.

The jump from there to:

A stronger tibialis prevents shin pain and makes your knees healthy.

is not scientifically covered, though.

A recent meta-analysis on preventing medial tibial stress syndrome (MTSS) found evidence for certain neuromuscular programs, and in individual populations for insoles — not for isolated tibialis raises as an established prevention method.

That does not mean tibialis work is useless.

It only means:

The direct prevention claim is stronger than the available evidence.

If tibialis raises agree with you and you want to train your dorsiflexors: entirely legitimate. You just do not get a health guarantee with it.

6. Jefferson curls: the honest answer is “specific evidence is missing”

Jefferson curls combine controlled spinal flexion with an external load.

In the fitness world there are two extreme positions on this:

• “Loaded spinal flexion is dangerous.”
• “The spine has to be trained heavy in maximal flexion, otherwise it stays fragile.”

For Jefferson curls themselves there is no long-term evidence base comparable to the one for squats or Nordic curls.

At the same time, the blanket claim that any loaded lumbar flexion causes injury is scientifically too crude as well. A systematic review of lifting tasks found no convincing prospective evidence that more lumbar flexion is in itself a risk factor for back pain.

That does not mean load is irrelevant.

Tissue adapts to loading, but dose, training history, speed, fatigue and individual symptoms all influence tolerance.

So for Jefferson curls, a cautious classification is more appropriate than a green or a red safety seal:

Specific long-term data are thin. Anyone who wants to use the movement should load it gradually and not treat it as necessary therapy for back health. With existing back complaints, the decision belongs in an individual context.

The evidence table

The most important distinction in this table is the one between exercise and claim.

An exercise can be worthwhile even when the marketing claim attached to it is phrased too big.

What ATG popularised correctly, in my view

Not avoiding knee loading on principle

Robustness does not come from permanently working around every demanding position.

In rehab and training alike, gradual exposure within tolerated ranges is a central principle.

Treating range of motion as a trainable ability

Deep positions can be trained, as long as they are loaded in a controlled and progressive way.

That fits today’s evidence better than blanket bans on particular joint angles.

Making unsexy exercises visible

Nordics, calf work, backward walking and controlled split squat variations were not invented by ATG. But ATG made them visible to a large audience.

That is a real practical contribution — regardless of who the exercise is originally credited to.

Where ATG claims can go too far

”Bulletproof” is not a scientific category

No exercise makes a joint impossible to injure.

Injury risk depends on load, exposure, sport, prior damage, recovery and many other factors.

Even highly effective prevention programs reduce risk — they do not eliminate it.

Evidence for one component is not evidence for the whole package

Nordic curls can be very well supported while tibialis claims stay thin at the same time.

So a program should not be written off wholesale as either “scientifically proven” or “bro science”. Its components can sit at completely different levels of evidence.

Pain is not a general fitness problem

Knee pain can have many causes and many diagnoses.

That an exercise works in studies in one particular population does not mean the same exercise is the right solution for every person with “knee problems”.

That applies especially with persistent pain, swelling, instability, trauma, or after surgery.

The big picture: the strongest ATG idea is actually the least spectacular

The most valuable thought behind “knees over toes” is not one particular slant board exercise.

It is the move away from a binary rule:

This position is dangerous and must be avoided.

towards:

Loading can be dosed and built up step by step.

That fits modern training theory very well.

But the very same logic also demands that ATG itself not be treated as dogma.

Not every person needs maximum knee-travel exercises. Not every person needs tibialis raises. Not every person has to do Jefferson curls.

The exercise is a tool. The goal decides whether you need it.

Who can ATG elements make sense for?

Healthy lifters

If you want to train larger knee and ankle angles, improve your single-leg strength or add extra hamstring work, several ATG-adjacent exercises slot into normal strength training without trouble.

Athletes

Nordics and targeted robustness in deep positions can be highly relevant depending on the sport. What decides it, though, is the overall plan across strength, sprinting, jumping, sport practice and recovery.

People with knee pain

Here the answer gets more individual.

Backward walking can be worthwhile for certain populations, and gradual strength work is central to many rehab concepts. But a social media program is no substitute for a diagnosis when complaints persist or are complex.

Conclusion

ATG is neither a scientific miracle program nor pure hype.

Some of its best-known components have a very good evidence base. Others are plausible but barely studied directly. And some statements are simply bigger than what the studies actually show.

The cleanest classification:

• Knees past the toes: not a general safety problem; it changes how load is distributed.
• Deep squats: not dangerous in principle for healthy, suitably prepared people.
• Nordic curls: a very well supported tool for reducing hamstring injuries in athletic populations.
• Backward walking: good rehab evidence in knee osteoarthritis; not automatically proof for reverse sleds.
• Tibialis raises: a legitimate exercise, but no proven shin splint or knee insurance.
• Jefferson curls: specific long-term data thin; neither blanket panic nor a blanket safety guarantee is justified.

ATG is most useful when you treat the exercises as options and check the marketing claims separately.

And for the squat, the same rule still stands: your anatomy and your goal decide what execution makes sense. More on that in the article on individual squat technique.

Sources

• Hartmann H, Wirth K, Klusemann M (2013). Analysis of the Load on the Knee Joint and Vertebral Column with Changes in Squatting Depth and Weight Load. Sports Medicine, 43(10):993–1008.
• Schoenfeld BJ (2010). Squatting Kinematics and Kinetics and Their Application to Exercise Performance. Journal of Strength and Conditioning Research, 24(12):3497–3506.
• Fry AC, Smith JC, Schilling BK (2003). Effect of knee position on hip and knee torques during the barbell squat. Journal of Strength and Conditioning Research, 17(4):629–633.
• Straub RK, Powers CM (2024). A Biomechanical Review of the Squat Exercise: Implications for Clinical Practice. International Journal of Sports Physical Therapy, 19(4):490–501.
• Wolf M, Androulakis-Korakakis P, Fisher JP et al. (2023). Partial Vs Full Range of Motion Resistance Training: A Systematic Review and Meta-Analysis. International Journal of Strength and Conditioning, 3(1).
• van Dyk N, Behan FP, Whiteley R (2019). Including the Nordic hamstring exercise in injury prevention programmes halves the rate of hamstring injuries: a systematic review and meta-analysis of 8459 athletes. British Journal of Sports Medicine, 53(21):1362–1370.
• Al Attar WSA et al. (2017). Effect of Injury Prevention Programs that Include the Nordic Hamstring Exercise on Hamstring Injury Rates in Soccer Players. Sports Medicine, 47(5):907–916.
• Lin LH, Peng YL, Yen LW et al. (2025). Effectiveness of backward walking exercises combined with conventional rehabilitation programs on managing pain intensity and disability in patients with knee osteoarthritis: A systematic review and meta-analysis of randomized controlled trials. Physiotherapy Theory and Practice.
• Marques TBT et al. (2025). Preventive interventions for medial tibial stress syndrome: Systematic review and meta-analysis. Gait & Posture, 122:92–98.
• Saraceni N, Kent P, Ng L et al. (2020). To Flex or Not to Flex? Is There a Relationship Between Lumbar Spine Flexion During Lifting and Low Back Pain? A Systematic Review With Meta-analysis. Journal of Orthopaedic & Sports Physical Therapy, 50(3):121–130.