Mobility and Flexibility

Mobility and Flexibility

Most of what is sold as mobility is a confusion of constructs. Flexibility, joint range of motion, stretch tolerance, tissue stiffness, and motor control are not the same quantity, do not change by the same mechanisms, and do not all need to be increased. The fitness industry collapses them; the experimental literature does not.

Compiled by South Beach Longevity · 20 August 2026 Copyright 2026 Series SBL-41 / SP-MOBILITY-AND-FLEXIBILITY · Register A scientific article Sources peer-reviewed human trials, consensus statements, meta-analyses, and labelled animal or in-vitro work · verified NCBI records Constraint This document describes published research. It is not medical advice. No human use, dose, route or schedule is recommended anywhere in this document.

How to read this document Every finding is labelled, in the sentence that reports it, by the kind of study that produced it. An acute stretch bout is not a twelve-week programme. Passive range of motion is not usable range under load. A soccer cohort is not an older-adult clinic. Stretching is not flexibility training, and range of motion is not flexibility (Afonso, Blazevich, and Behm, 2026). Where two results conflict, both are given. Durations and angles appear only as reported experimental parameters, always with the population attached. Nothing here is a recommendation. Findings are graded in place as established, strongly supported, emerging, plausible, or speculative. Two further labels mark careful absences rather than verdicts: not established, where the evidence is too thin to place a claim on the ladder at all — untested or insufficient, an absence of proof rather than disproof; and not supported, where the weight of evidence leans against a claim but stops short of a formal refutation.

01 What this document is, and four things it is not

This article is a research review of human flexibility, mobility, and range of motion (ROM): how those words are used, which quantities experiments actually measure, which interventions change which quantities, and which popular claims survive contact with randomised trials, meta-analyses, and in vivo muscle–tendon mechanics. It is written against a fitness market that treats “tight,” “locked,” “immobile,” and “needs more mobility” as diagnoses.

Four things follow immediately.

First, this is not a stretching programme. The American College of Sports Medicine’s 2011 position stand treats flexibility exercise as one component of a broader adult activity prescription, not as a sport, a religion, or a substitute for strength and neuromotor training (Garber et al., 2011). That paper is a society position, strongly supported as a description of what ACSM then judged the adult record to say. It is not used here as a licence to prescribe hold times or weekly frequencies for any person.

Second, it is not sports-injury prevention by another name. Stretching before or after activity has been tested as an injury intervention and has repeatedly failed to move incidence in a clinically useful way (Pope, Herbert, and Kirwan, 2000; Herbert and Gabriel, 2002; Thacker et al., 2004; Lauersen, Bertelsen, and Andersen, 2014). Neuromuscular training and strength training have a different record. Collapsing those two literatures is the most expensive error in this field.

Third, it is not a fascia-release manual. Foam rolling and “self-myofascial release” change sensation and, modestly, acute ROM. They have not been shown to remodel fascia at the loads used in gyms (Beardsley and Škarabot, 2015; Wiewelhove et al., 2019; Wilke et al., 2020). Treating a foam roller as a structural instrument is a category error.

Fourth, it is not medical advice and it is not a treatment for contracture, Ehlers–Danlos syndrome, or pain. Harvey, Katalinic, and Herbert, in an abridged Cochrane review of stretch for contracture, reported mean effects on the order of two degrees in neurological populations and about one degree in non-neurological populations — not a clinically meaningful change (Harvey, Katalinic, and Herbert, 2017). Hypermobility is a counterexample to the claim that more ROM is always better, not a flexibility goal (Beighton, Solomon, and Soskolne, 1973; Malfait et al., 2017).

This document describes published research. It recommends no exercise, stretch, yoga style, roller, or schedule for any person.

02 Flexibility, mobility, and joint ROM

Joint range of motion is an angle, or a pair of angles, at a named joint, measured under a named condition. It is an outcome. It is not a tissue property.

Flexibility, in the older sports-medicine sense, is a family of measures that can be static (end of ROM), dynamic-passive (stiffness or compliance under externally applied torque), or dynamic-active (stiffness with the muscle contracted). Gleim and McHugh said this in 1997 and noted that how flexibility is measured determines what “changed” appears to mean (Gleim and McHugh, 1997). That distinction is established as terminology and is still routinely ignored in commercial language.

Mobility, in contemporary coaching, is usually asked to mean active control of a range that a person can use under load. The word is not a laboratory quantity. It is a slogan covering active ROM, motor control, and sometimes strength at long muscle lengths. Afonso, Blazevich, and Behm, writing in 2026, argued that ROM and flexibility are not interchangeable, and that stretching is not the same intervention as flexibility training (Afonso, Blazevich, and Behm, 2026). A 2025 Delphi statement of stretching researchers reached a similar practical conclusion: recommendations have to name the construct being trained (Warneke, Thomas, Blazevich et al., 2025). Those papers are consensus and conceptual work. They are strongly supported as a warning against collapsed language. They are not new trials.

Three consequences follow.

A sit-and-reach centimetre is not hip flexion, is not hamstring stiffness, and is not lumbar extensibility. Mayorga-Vega, Merino-Marban, and Viciana’s meta-analysis found moderate criterion validity of sit-and-reach tests for hamstring extensibility and weaker validity for lumbar extensibility (Mayorga-Vega, Merino-Marban, and Viciana, 2014). Using the test as a whole-body mobility grade is a misuse of a moderately valid hamstring proxy.

A person with large passive ROM and poor active control is not “mobile” in the coaching sense. A person with limited passive ROM who can produce force through the range they have is not automatically injured. Witvrouw and colleagues found that reduced hamstring and quadriceps flexibility in male professional soccer players was associated with later muscle injury (Witvrouw et al., 2003). That is a prospective risk-factor study. It is strongly supported as evidence that low flexibility can mark risk in that sport. It is not evidence that stretching the marked players would have prevented the injuries. Risk factors are not automatically interventions.

There is no universal ideal ROM. Overhead throwing alters the glenohumeral rotation arc; a loss of internal rotation that exceeds the gain in external rotation, or a loss of total arc, is a different problem from a yoga-class shortfall (Kibler, Sciascia, and Thomas, 2012; Wilk et al., 2011). Older adults and people with generalised joint hypermobility sit on opposite sides of the same error: treating a single ROM number as a health target.

03 Passive ROM, active ROM, stiffness, compliance, and stretch tolerance

Passive ROM is the excursion reached when an external torque moves the joint, typically to a sensory or investigator-defined end. Active ROM is the excursion the person can reach by their own muscle action. The gap between them is not fascia. It is often strength, motor control, or pain (Figure 4).

Stiffness is the slope of torque against angle, or force against displacement, in a specified range. Compliance is its inverse. Neither is “tightness.” A muscle–tendon unit can be stiff in one range and not in another; Magnusson and colleagues showed that so-called tight hamstrings differed from “normal” hamstrings in stretch tolerance, with higher stiffness only when compared in a common range, not as a simple material diagnosis (Magnusson, Simonsen, Aagaard et al., 1997). That human in vivo work is established.

Muscle–tendon compliance is a series property. Tendon, aponeurosis, and fascicles share the displacement. Kubo’s group measured tendon structures in vivo and showed that acute static stretching can reduce tendon stiffness and hysteresis, and that short-term stretch training can change hysteresis without a matching stiffness change (Kubo, Kanehisa, Kawakami, and Fukunaga, 2001; Kubo, Kanehisa, and Fukunaga, 2002). Those are ultrasound and dynamometry experiments in defined lower-limb sites. They are strongly supported for those sites and timescales. They are not a whole-body fascia theorem.

Stretch tolerance is the sensory and affective limit: the angle or torque at which the person stops because of discomfort, not because the tissue has reached a structural end. Weppler and Magnusson reviewed the mechanical and sensory accounts of increased extensibility and concluded that much of the chronic ROM gain after intermittent stretching is a change in sensation, not a durable increase in muscle length (Weppler and Magnusson, 2010). Folpp, Deall, Harvey, and Gwinn tested that claim directly. After four weeks of intensive hamstring stretching in healthy adults with limited extensibility, muscle extensibility changed by −1° (95% confidence interval −4 to 3) while stretch tolerance rose by 8° (5 to 12) (Folpp et al., 2006). That trial is strongly supported evidence that a large, uncomfortable stretching diet can move the stop without moving the tissue. It is not evidence that no stretching programme ever changes architecture. Later sections treat the high-volume exceptions.

McHugh, Kremenic, and Fox separated mechanical and neural restraints to passive ROM and showed that both can operate (McHugh, Kremenic, and Fox, 1998). Reiner, Tilp, Nakamura and colleagues asked, more recently, whether muscle stiffness is even a determinant of leg ROM, and found the relationship too weak to treat stiffness as the explanation of flexibility scores (Reiner et al., 2024). Grade: emerging as a quantitative caution; consistent with the older tolerance literature.

04 Joint structure and motor control

Joint ROM is bounded first by geometry: bony congruence, labrum and meniscus, capsule, ligament, and the finite length of periarticular connective tissue. No amount of hamstring stretching will make a cam-type hip into a dancer’s hip. That is anatomy, not a training failure.

Motor control is the other bound that stretching programmes pretend not to need. Active ROM requires the person to produce force, and to inhibit antagonists, at joint angles they may rarely visit. Resistance training through a full or long-muscle-length range changes that capacity in ways passive stretching does not automatically copy (Pallarés et al., 2020; Schoenfeld and Grgic, 2020; Pedrosa et al., 2022). Nordic hamstring and hip-extension work change biceps femoris architecture; a toe-touch does not make the same claim (Bourne et al., 2017; Timmins, Shield, Williams et al., 2016).

Lederman’s critique of “core stability” as a clinical mythology is relevant here as a caution about motor-control slogans, not as a flexibility trial (Lederman, 2010). Stability, mobility, and tightness are three marketing words that have been asked to do the work of torque, angle, and task.

05 Measurement

A ROM number is only as good as the instrument, the landmark, the torque, and the instruction. Goniometry and inclinometry are the clinical default. Isokinetic dynamometry with ultrasound is the research default when the question is stiffness or fascicle length rather than a clinic angle. Sit-and-reach remains the field default because it is cheap. Cheap is not valid.

Mayorga-Vega and colleagues’ meta-analysis is the load-bearing statement on sit-and-reach: moderate criterion validity for hamstrings, poorer for lumbar spine (Mayorga-Vega, Merino-Marban, and Viciana, 2014). Sex differences in sit-and-reach appear in paediatric samples and should not be moralised as training gaps (Nuzzo, 2026). Beighton scoring for generalised joint hypermobility is a nine-point screen, not a flexibility curriculum; its measurement properties are mixed across cut-offs and methods (Beighton, Solomon, and Soskolne, 1973; Juul-Kristensen et al., 2017).

Thomas, Bianco, Paoli and colleagues related stretching typology and duration to ROM outcomes and found that the protocol details — static, dynamic, proprioceptive neuromuscular facilitation (PNF), and how long the stretch is held — change the ROM effect (Thomas et al., 2018). Comparing two studies that both “stretched hamstrings” without naming torque, end-feeling, and duration is how folklore survives a literature search.

This document treats ROM as condition-specified. Acute post-stretch ROM is not chronic ROM. Passive ROM is not active ROM. A pain-limited stop is not a structural stop.

06 Acute viscoelastic behaviour

Muscle–tendon tissue is viscoelastic. Under a held stretch it creeps; under a constant length the force relaxes. Magnusson’s hamstring series is the human reference: repeated static stretches produce stress relaxation; the mechanical change is real during the bout and does not, by itself, explain a lasting flexibility identity (Magnusson, Simonsen, Aagaard et al., 1995; Magnusson et al., 1996; Magnusson, Simonsen, Aagaard et al., 1996). Viscoelastic stress relaxation can occur with negligible EMG, which is why a “relaxed stretch” is still a mechanical event (Magnusson, Simonsen, Dyhre-Poulsen et al., 1996). Grade: established.

The practical implication is temporal. An immediate ROM gain after a stretch session is partly creep and relaxation. Weppler and Magnusson noted that the length increase during stretch is transient, with magnitude and duration depending on the protocol (Weppler and Magnusson, 2010). Konrad and Tilp tracked muscle–tendon properties after three and five minutes of static stretching and described a time course of function and structure that returns toward baseline rather than remaining as a new anatomy (Konrad and Tilp, 2020; Konrad, Reiner, Thaller et al., 2019). Kubo, Kanehisa, Kawakami, and Fukunaga found that a single static-stretch session reduced tendon stiffness and hysteresis in vivo (Kubo et al., 2001). Three weeks of stretch training, in a related paper, changed hysteresis more than stiffness (Kubo, Kanehisa, and Fukunaga, 2002). Acute tendon mechanics and chronic tendon adaptation are different claims.

Gajdosik, Vander Linden, and McNair measured viscoelastic properties of short calf muscle–tendon units in older women under slow and fast passive dorsiflexion (Gajdosik, Vander Linden, and McNair, 2005). Age changes the material; it does not convert stretching into a different physics.

07 Neural tolerance

A long passive stretch does not only unload collagen. It changes the nervous system’s willingness to allow the angle, and it can reduce force. Guissard and Duchateau reviewed neural aspects of muscle stretching: reduced stretch-reflex excitability and altered motoneuron drive are part of the acute picture (Guissard and Duchateau, 2006). Trajano, Nosaka, and Blazevich reviewed neurophysiological mechanisms of stretch-induced force loss (Trajano, Nosaka, and Blazevich, 2017). Trajano and Blazevich later argued that static stretching reduces motoneuron excitability, with a possible neuromodulatory contribution (Trajano and Blazevich, 2021). Avela, Kyröläinen, and Komi showed that repeated prolonged passive stretching of triceps surae deteriorated maximal voluntary contraction, EMG, and reflex sensitivity (Avela, Kyröläinen, and Komi, 1999). Fowles, Sale, and MacDougall imposed about thirty minutes of cyclical maximal passive stretch of the plantarflexors in ten young adults and reported an immediate MVC reduction of 28%, with incomplete recovery at 60 minutes (Fowles, Sale, and MacDougall, 2000). Those experiments are established for prolonged stretch. They are not a 15-second warm-up.

Pulverenti, Trajano and colleagues found that stretch-induced force loss was not accompanied by altered corticospinal excitability in one protocol, and lacked cortical or Ia-afferent spinal pathway involvement in another (Pulverenti et al., 2019; Pulverenti, Trajano, Walsh et al., 2020). The neural story is not a single pathway. What is strongly supported is the performance consequence of long static stretch, not a complete circuit diagram.

Behm, Kay, Trajano and colleagues located the practical caveat: impairments after prolonged static stretching are most obvious when the stretch is not embedded in a comprehensive warm-up (Behm, Kay, Trajano et al., 2021). Chaabene, Behm, Negra and colleagues attempted to clarify acute strength and power caveats and reached a similar duration- and context-dependent picture (Chaabene et al., 2019). Short static stretch inside a full warm-up is a different exposure from Fowles’s 30-minute laboratory bout. Mixing those two is how a real finding becomes a gym superstition, and also how a superstition ignores a real finding.

08 Fascicle length and tendon adaptation

If stretching “lengthened muscle,” ultrasound should show longer fascicles, and dynamometry should show a shifted torque–angle curve that survives after tolerance is accounted for. Ordinary 3–8 week static-stretch programmes often fail that test.

Freitas, Mendes, Le Sant and colleagues’ systematic review asked whether chronic stretching changes muscle–tendon mechanical properties and did not find a consistent architectural explanation for the ROM gain at ordinary training doses (Freitas et al., 2018). Konrad and Tilp trained static stretching and reported ROM rising from 30.9° to 36.3° without alteration of fascicle or tendon stiffness measures that would explain the gain (Konrad and Tilp, 2014). Panidi, Donti, Konrad and colleagues’ later meta-analysis of static-stretch training on muscle architecture found only trivial-to-small fascicle-length effects overall; a high-volume adolescent protocol from the same group is the exception that proves the dose point, not the gym default (Panidi, Donti, Konrad et al., 2023; Panidi, Bogdanis, Terzis et al., 2021). Their ballistic-stretch and PNF-stretch training papers similarly found ROM change without matching architectural change in muscle and tendon structures (Konrad and Tilp, 2014; Konrad, Gad, and Tilp, 2015). An acute comparison of static, ballistic, and PNF stretch showed immediate tissue-property differences that still do not license a chronic fascia narrative (Konrad, Stafilidis, and Tilp, 2017). Grade for typical gym stretching as architecture: not established; often refuted at that dose.

High-volume or loaded stretching is the remaining mechanical candidate. Nunes, Schoenfeld, Nakamura and colleagues reviewed whether stretch training induces hypertrophy in humans and found that stretching at a person’s own ROM, without additional load, did not; loaded or apparatus stretching was the setting in which hypertrophy sometimes appeared (Nunes, Schoenfeld, and Nakamura, 2020). Warneke’s group compared long-lasting static stretch with hypertrophy training in plantarflexors, and in a randomised trial of long-lasting static stretch reported morphological and functional changes that ordinary brief stretching does not reproduce (Warneke, Wirth, Keiner et al., 2023; Warneke, Keiner, Wohlann et al., 2023). Their 2024 meta-analysis of chronic static stretching found a small hypertrophy effect (on the order of d ≈ 0.20) only at high dose, with strength and power effects remaining small beside a large flexibility effect (Warneke, Lohmann, Behm et al., 2024). Arntz, Markov, Behm and colleagues’ multilevel meta-analysis of chronic static stretching reported standardised mean differences of 0.21 for strength, 0.19 for power, and 0.96 for flexibility (Arntz, Markov, Behm et al., 2023). A companion hypertrophy meta-analysis from the same orbit treated muscle size as a possible, dose-dependent byproduct, not the purpose of stretching (Arntz, Markov, Schoenfeld et al., 2024). Grade: emerging that very high-volume or loaded stretch can move morphology; established that ordinary stretching moves ROM far more than it moves strength or size.

Eccentric training is a different architecture stimulus. Timmins, Bourne, Shield and colleagues showed, in elite soccer, that biceps femoris long-head fascicles shorter than 10.56 cm carried a relative risk of 4.1 for later hamstring strain, and that eccentric knee-flexor strength below 337 N carried a relative risk of 4.4 (Timmins et al., 2016). Bourne, Duhig, Timmins and colleagues showed that Nordic hamstring and hip-extension exercise change hamstring architecture and morphology (Bourne et al., 2017). Those are injury-risk and training-adaptation papers, not stretching papers. They are the reason this article refuses to treat “more flexibility” as the hamstring-injury programme.

Resistance training through a long range is the other architectural route. Bloomquist and colleagues randomised men to 12 weeks of deep squat (0–120° knee flexion) versus shallow squat and found greater muscle and functional adaptations with the deep range (Bloomquist et al., 2013). Pallarés, Cava, Courel-Ibáñez and colleagues reported that full squat produced greater neuromuscular and functional adaptations and lower pain than partial squats after prolonged training (Pallarés et al., 2020). Pedrosa, Lima, Schoenfeld and colleagues found that partial ROM at long muscle lengths elicited favourable muscular adaptations compared with short-length partials (Pedrosa et al., 2022). Schoenfeld and Grgic’s systematic review of ROM during resistance training is the map of that literature (Schoenfeld and Grgic, 2020). Grade: strongly supported that the range used under load changes hypertrophy and function; not a claim that everyone should squat to a named angle.

09 Sensory change versus tissue change

The sensory theory is not a consolation prize. It is the best-supported explanation of why people get better at sit-and-reach after a few weeks of stretching.

Weppler and Magnusson’s review is the conceptual load-bearing paper: muscle length increases during stretch because the tissue is viscoelastic; the lasting increase in extensibility after intermittent stretching is, in many protocols, a modified sensation (Weppler and Magnusson, 2010). Folpp et al. (2006) is the cleanest experimental split of extensibility versus tolerance, cited in section 03. Konrad and Tilp (2014) is the cleanest ultrasound split of ROM versus structure after static-stretch training.

Klinge, Magnusson, Simonsen and colleagues combined strength and flexibility training and measured EMG, stiffness, and viscoelastic stress relaxation, showing that the training cocktail does not reduce to a single stiffness story (Klinge et al., 1997). Magnusson, Aagaard, Simonsen and colleagues measured passive tensile stress and energy of human hamstrings in vivo and treated those as mechanical quantities, not as diagnoses of character (Magnusson, Aagaard, Simonsen et al., 2000).

The remaining honest sentence is conditional. If stretch volume, load, and duration are high enough, architecture can move (Nunes, Schoenfeld, and Nakamura, 2020; Warneke, Lohmann, Lima et al., 2023; Warneke, Lohmann, Behm et al., 2024). Most commercial stretching is not that intervention. Treating a 30-second hamstring stretch as a sarcomere programme is speculative. Treating it as a way to change what a person will tolerate is strongly supported.

10 Static stretching

Static stretching is a held position at or near end range, with the muscle relatively relaxed. It is the oldest gym habit in this document and the best studied.

Acute ROM. Static stretching increases ROM in the minutes after the bout. Thomas et al. (2018) treated duration and typology as determinants of that effect. Konrad’s time-course papers show that the tissue and functional responses after a few minutes of static stretch are not a new permanent length (Konrad, Reiner, Thaller et al., 2019; Konrad and Tilp, 2020). Grade: established for acute ROM; not established as lasting architecture.

Chronic ROM. Bandy, Irion, and Briggler randomised adults with limited hamstring flexibility to different hold times and frequencies over six weeks and found that 30 seconds of static stretching was as effective as 60 seconds for knee-extension ROM, with once-daily stretching sufficient in that protocol (Bandy, Irion, and Briggler, 1997). That trial is strongly supported as a duration comparison in young adults with limited hamstrings. It is not a hypertrophy protocol and not an older-adult function trial. Feland, Myrer, Schulthies and colleagues tested 15, 30, and 60 second hamstring stretches in people aged 65 years or older (mean age 84.7 in the reported sample) and found duration still mattered in that group (Feland et al., 2001). Stathokostas, Little, Vandervoort and colleagues’ systematic review of flexibility training and functional ability in older adults found inconsistent translation from ROM to function (Stathokostas, Little, Vandervoort et al., 2012).

Acute force and power. Kay and Blazevich’s systematic review is the load-bearing acute-performance paper: stretches under 30 seconds produced a mean force change of about −1.1%; 30–45 seconds about −1.9%; longer durations produced larger impairments (Kay and Blazevich, 2012). Simic, Sarabon, and Markovic’s meta-analysis of pre-exercise static stretching reported strength −5.4%, power −1.9%, and explosive performance −2.0%, with the worst picture when static stretch was the sole warm-up (Simic, Sarabon, and Markovic, 2013). Behm and Chaouachi’s narrative review of acute static versus dynamic stretching reached the same practical split (Behm and Chaouachi, 2011). Yamaguchi and Ishii compared 30 seconds of static stretching with dynamic stretching on leg-extension power and favoured dynamic work for power (Yamaguchi and Ishii, 2005). Winchester, Nelson, Landin, Young, and Schexnayder showed that adding static stretching after a dynamic warm-up still impaired sprint performance in collegiate track athletes (Winchester et al., 2008). Wong, Chaouachi, Lau and colleagues found that short static stretches combined with dynamic stretching did not impair repeated sprints and agility in their protocol (Wong, Chaouachi, Lau et al., 2011). The literature is therefore duration- and context-dependent, not a single prohibition. Grade: established that prolonged static stretch can reduce maximal force and power; strongly supported that short static stretch inside a full warm-up is a smaller, sometimes trivial, effect (Behm, Kay, Trajano et al., 2021).

Chronic strength and power. Kokkonen, Nelson, Eldredge, and Winchester reported that a 10-week static-stretching routine, without other training, improved several performance tests in previously inactive volunteers (Kokkonen et al., 2007). That is a small stretch-only trial against a non-training control. Later meta-analyses put the chronic performance effect in a much smaller box than the flexibility effect (Arntz, Markov, Behm et al., 2023; Warneke, Lohmann, Behm et al., 2024). Shrier’s 2004 systematic review already distinguished acute impairment from a possible chronic benefit and warned against spending the acute literature as if it were the chronic one (Shrier, 2004). Grade for chronic static stretch as a strength method: emerging and small; inferior to resistance training for those outcomes.

Page’s 2012 clinical commentary remains a useful map of stretching types for exercise and rehabilitation, with the usual limitation of a commentary: it organises evidence, it does not replace it (Page, 2012).

11 Dynamic and ballistic stretching

Dynamic stretching moves the joint through range under control, typically as part of a warm-up. Ballistic stretching uses bouncing or momentum to the end range. The two are not the same intervention, though commercial language often treats “dynamic” as a moral upgrade of “ballistic.”

Yamaguchi and Ishii (2005) and Yamaguchi, Ishii, Yamanaka and colleagues (2007) found acute dynamic stretching more favourable for concentric power than static stretching. Little and Williams tested static stretching, dynamic stretching, or no stretching in the warm-up of eighteen professional soccer players: dynamic stretching better preserved high-speed capacities; static stretching was the least friendly to sprint and agility in that sample (Little and Williams, 2006). McMillian, Moore, Hatler, and Taylor compared a dynamic warm-up with a static-stretching warm-up and no warm-up in military cadets and found the dynamic protocol superior for power and agility tests (McMillian et al., 2006). Turki, Chaouachi, Drinkwater and colleagues reported that ten minutes of dynamic stretching was sufficient to potentiate vertical-jump characteristics in their sample (Turki et al., 2011). O’Sullivan, Murray, and Sainsbury compared warm-up, static stretching, and dynamic stretching on hamstring flexibility in previously injured subjects (O’Sullivan, Murray, and Sainsbury, 2009). Takeuchi, Nakamura, Matsuo and colleagues later showed that speed and amplitude of dynamic stretching change flexibility and strength outcomes, and that combining static and dynamic stretching has its own stiffness and strength profile (Takeuchi, Nakamura, Matsuo et al., 2022; Takeuchi et al., 2024). Grade: strongly supported that dynamic stretching is the warmer neighbour of high-speed performance; not a claim that dynamic stretching is an injury vaccine.

Ballistic stretching can increase ROM. Konrad and Tilp (2014) trained ballistic stretching and still did not find the architectural change that would make it a tissue-lengthening method. Matsuo, Takeuchi, Nakamura and colleagues’ 2025 systematic review and meta-analysis of acute dynamic and ballistic stretching on flexibility is the recent quantitative map (Matsuo, Takeuchi, Nakamura et al., 2025). Grade: emerging as a pooled acute-flexibility estimate; still not a fascia story.

12 Proprioceptive neuromuscular facilitation

PNF stretching pairs a stretch with a voluntary contraction of the target muscle or its antagonist (contract–relax, hold–relax, and related methods). Sharman, Cresswell, and Riek reviewed mechanisms and clinical implications and treated autogenic and reciprocal inhibition as the usual neural story, with stretch tolerance again doing a large share of the ROM work (Sharman, Cresswell, and Riek, 2006). Hindle, Whitcomb, Briggs and colleagues similarly reviewed PNF effects on ROM and muscular function (Hindle et al., 2012). Konrad, Gad, and Tilp trained PNF stretching and, as with static stretching, increased ROM without a matching change in muscle and tendon structures (Konrad, Gad, and Tilp, 2015). Konrad, Seiberl, Tilp and colleagues later showed that isolated PNF of quadriceps or triceps surae, followed by post-stretching activity, alters tissue stiffness and jump performance in an acute design (Konrad, Seiberl, Tilp et al., 2024). Grade: strongly supported that PNF is an efficient way to gain ROM in the short term; not established that it lengthens muscle; emerging and protocol-dependent for acute power effects.

Practical commentaries often mention low contraction intensities and brief contractions. Those numbers, when they appear in trials, are experimental parameters. They are not a prescription in this document.

13 Loaded stretching and eccentric training

Loaded stretching applies an external load at long muscle length. It sits closer to resistance training than to a yoga hold. Nunes, Schoenfeld, and Nakamura (2020) is the human hypertrophy review: self-stretch at existing ROM, no; loaded stretch, sometimes. Warneke, Lohmann, Lima and colleagues’ narrative review of stretch-mediated hypertrophy and strength treats high-volume loaded stretch as a physiological candidate, not as a replacement for lifting (Warneke, Lohmann, Lima et al., 2023). Wohlann, Warneke, Kalder and colleagues compared eight weeks of supervised static stretching of pectoralis major with resistance training and measured maximal strength, thickness, and ROM (Wohlann et al., 2024). Warneke, Aragão-Santos, Alizadeh and colleagues asked whether acute foam-rolling effects are simply dynamic-warm-up effects in disguise (Warneke, Aragão-Santos, Alizadeh et al., 2023). Grade: emerging. High-volume loaded stretch is not the 30-second doorway stretch, and citing one as if it were the other is a bait-and-switch.

Eccentric training is included here because the market sells it as “lengthening.” The relevant length is fascicle length under load, not a stretching sensation. Timmins et al. (2016) and Bourne et al. (2017) are the hamstring architecture papers. Van der Horst, Smits, Petersen, Goedhart, and Backx published a randomised-trial protocol for the Nordic hamstring exercise as injury prevention in amateur soccer; the protocol paper is not the outcome paper and is cited as such (van der Horst et al., 2014). Eccentric work is a strength intervention with architectural side effects. It is not a flexibility class.

14 Resistance training through full range of motion

If the goal is usable range, loading the range is the intervention with the least folklore per millimetre of ultrasound.

Schoenfeld and Grgic (2020) reviewed ROM during resistance-training interventions and found that full-range protocols generally favour muscle development compared with partial-range protocols, with important exceptions when the partial is performed at long muscle length. Pedrosa et al. (2022) and Pedrosa, Simões, Figueiredo and colleagues (2023) are the long-length versus short-length partials contrast. Pallarés et al. (2020) and Bloomquist et al. (2013) are the squat-depth trials. Grade: strongly supported.

This is also the cleanest answer to “everyone needs more mobility.” A person who cannot control the bottom of a squat has a strength-and-control problem at that angle. Stretching them into a deeper passive position, then asking them to load it, skips the variable that the hypertrophy and function trials actually manipulated.

15 Yoga

Yoga is a mixed intervention: postures, isometric holds at end range, breathing, and, in many trials, a social and attentional package. It is not a purified stretching experiment, which is why a yoga ROM gain cannot be spent as evidence that stretching lengthened fascia.

Cramer, Lauche, Haller and colleagues’ systematic review and meta-analysis of yoga for low back pain, and Anheyer, Haller, Lauche and colleagues’ later update, treat pain and function as the outcomes, not sit-and-reach (Cramer, Lauche, Haller et al., 2013; Anheyer, Haller, Lauche et al., 2022). Cramer, Lauche, Hohmann and colleagues randomised yoga against home-based exercise for chronic neck pain (Cramer, Lauche, Hohmann et al., 2013). Those are pain trials. They are strongly supported as evidence that some yoga programmes can change back- and neck-pain scores in defined samples. They are not mobility architecture.

In older adults, Tiedemann, O’Rourke, Sesto, and Sherrington’s 12-week Iyengar yoga pilot improved balance and mobility measures in community-dwelling older people (Tiedemann et al., 2013). Youkhana, Dean, Wolff, Sherrington, and Tiedemann’s systematic review and meta-analysis reported a moderate effect on balance (Hedges’ g = 0.40) in people aged 60 years and over (Youkhana et al., 2016). Grade: emerging to strongly supported for balance and mobility tests in older samples; still a mixed intervention.

Cramer’s other meta-analyses — cardiovascular risk factors, hypertension, heart disease, COPD — exist in the reviewed record and are not spent here as flexibility evidence (Cramer, Lauche, Haller et al., 2014; Cramer, Haller, Lauche et al., 2014; Cramer, Lauche, Haller et al., 2015; Cramer, Haller, Klose et al., 2019). Dropout in yoga trials is itself a measured quantity (Cramer, Haller, Dobos et al., 2016). A class that people leave is not a universal ROM prescription.

16 Warm-up

A warm-up is a temperature, neural, and skill preparation for the task that follows. Stretching is one optional ingredient. Treating “warm-up” as synonymous with static stretching is how the acute-force literature became a culture war.

Fradkin, Gabbe, and Cameron asked whether warming up prevents sport injury and found the randomised evidence thin (Fradkin, Gabbe, and Cameron, 2006). McCrary, Ackermann, and Halaki’s systematic review of upper-body warm-up found performance effects more often than a clean injury effect (McCrary, Ackermann, and Halaki, 2015). Soligard, Myklebust, Steffen and colleagues’ cluster-randomised trial of a comprehensive warm-up in young female footballers — later widely known as FIFA 11+ — reduced overall injury (rate ratio 0.68), overuse injury (0.47), and severe injury (0.55) (Soligard et al., 2008). Compliance modified the benefit (Soligard, Nilstad, Steffen et al., 2010). Al Attar, Soomro, Pappas and colleagues later found that adding a post-training 11+ block further reduced injury in male amateurs (Al Attar et al., 2017). Emery, Roy, Whittaker, Nettel-Aguirre, and van Mechelen’s meta-analysis of neuromuscular training in youth sport is the broader injury-prevention map (Emery et al., 2015). Grade: established that a structured neuromuscular warm-up can reduce injury in football and related youth sports; not established that the stretching minutes inside those programmes are the active ingredient.

Dynamic stretching, as section 11 recorded, is the performance-friendlier neighbour of a high-speed warm-up (Yamaguchi and Ishii, 2005; Little and Williams, 2006; McMillian et al., 2006; Behm and Chaouachi, 2011). Static stretching as the sole warm-up is the setting in which Simic, Sarabon, and Markovic (2013) found the largest performance cost. Short static stretch after or inside a dynamic preparation is a smaller exposure (Wong, Chaouachi, Lau et al., 2011; Behm, Kay, Trajano et al., 2021). McHugh and Cosgrave’s “to stretch or not to stretch” review treated injury and performance as separate questions and did not license a single answer (McHugh and Cosgrave, 2010).

17 Foam rolling and fascia claims

Foam rolling is sold as self-myofascial release: a structural intervention on a tissue most users cannot name. The trial literature measures ROM, pain-pressure threshold, sprint, and jump. It does not measure plastic deformation of fascia at gym loads.

Beardsley and Škarabot’s systematic review of self-myofascial release found acute ROM increases and mixed recovery and performance effects, and did not demonstrate a structural fascia mechanism (Beardsley and Škarabot, 2015). Cheatham, Kolber, Cain and colleagues reached a similar applied conclusion (Cheatham, Kolber, Cain et al., 2015). Škarabot, Beardsley, and Štirn compared foam rolling with static stretching on ankle ROM in adolescent athletes (Škarabot, Beardsley, and Štirn, 2015). Macdonald, Button, Drinkwater, and Behm tested foam rolling as recovery after intense activity (Macdonald et al., 2014). Pearcey, Bradbury-Squires, Kawamoto and colleagues tested foam rolling for delayed-onset muscle soreness and recovery of dynamic performance (Pearcey et al., 2015). Cheatham, Kolber, and Cain compared video-guided, live-instructed, and self-guided rolling on knee ROM and pressure-pain threshold (Cheatham, Kolber, and Cain, 2017).

Wiewelhove, Döweling, Schneider and colleagues’ meta-analysis of foam rolling on performance and recovery reported a small pre-roll sprint benefit (about +0.7%), a flexibility increase (about +4%), and a jump change of about −1.9% (Wiewelhove et al., 2019). Wilke, Müller, Giesche and colleagues’ multilevel meta-analysis found an acute ROM standardised mean difference of 0.74 versus no intervention, not larger than stretching (Wilke et al., 2020). Konrad, Alizadeh, Anvar and colleagues compared chronic static-stretch training with foam-rolling training: over periods up to about four weeks, stretching increased ROM more than foam rolling (Konrad, Alizadeh, Anvar et al., 2024). Behm, Alizadeh, Hadjizadeh Anvar and colleagues’ clinical commentary on foam-rolling “prescription” is a commentary, not a mechanism paper (Behm, Alizadeh, Hadjizadeh Anvar et al., 2020). Kaya, Cug, and Behm found that rolling during a simulated half-time attenuated some soccer-specific decrements (Kaya, Cug, and Behm, 2021). Warneke, Aragão-Santos, Alizadeh and colleagues asked whether acute rolling effects are simply dynamic-warm-up effects (Warneke, Aragão-Santos, Alizadeh et al., 2023).

Grade: strongly supported that rolling can raise acute ROM and blunt soreness ratings; not established as structural fascia change; speculative as a remodeling tool. Sensation and stretch tolerance are sufficient to explain the ROM. They are also less marketable.

18 Acute versus chronic

The same ROM number is not the same adaptation. Acute gains are viscoelastic and neural, largely reversible after the session (Magnusson, Simonsen, Aagaard et al., 1995; Weppler and Magnusson, 2010; Konrad and Tilp, 2020). Chronic gains over weeks are commonly stretch-tolerance changes (Folpp et al., 2006; Konrad and Tilp, 2014). Architectural change is possible at high volume or under load (Nunes, Schoenfeld, and Nakamura, 2020; Warneke, Lohmann, Behm et al., 2024). Injury and performance effects do not travel automatically with the ROM number (Herbert and Gabriel, 2002; Simic, Sarabon, and Markovic, 2013; Lauersen, Bertelsen, and Andersen, 2014).

19 Strength and power

Acute static stretch can cost force and power. Kay and Blazevich (2012) and Simic, Sarabon, and Markovic (2013) are the quantitative maps; Fowles, Sale, and MacDougall (2000) and Avela, Kyröläinen, and Komi (1999) are the long-bout physiology. Dynamic stretching is the warmer neighbour for high-speed tasks (Yamaguchi and Ishii, 2005; Little and Williams, 2006). Chronic static stretch has a large flexibility effect and small strength and power effects (Arntz, Markov, Behm et al., 2023). Kokkonen et al. (2007) remains a small stretch-only performance trial that later pooling shrunk. Shrier (2004) is still the right warning: do not spend acute impairment as if it cancelled chronic training, and do not spend chronic ROM as if it were a sprint method.

Resistance training through range, and eccentric hamstring work, are the strength interventions. Stretching is a ROM intervention that sometimes leaks a little strength when the dose is extreme.

20 Pain

Pain limits ROM. Stretching and rolling can change pain-pressure thresholds and soreness ratings without proving a structural repair.

Herbert and de Noronha’s Cochrane review, and the 2011 update with Kamper, found stretching before or after exercise did not produce a clinically important reduction in delayed-onset muscle soreness (Herbert and de Noronha, 2007; Herbert, de Noronha, and Kamper, 2011). Jamtvedt, Herbert, Flottorp and colleagues’ pragmatic randomised trial of stretching before and after physical activity found an all-injury hazard ratio of 0.97 and a soreness odds ratio of 0.69 — a possible small soreness effect without an injury effect (Jamtvedt et al., 2010). Pearcey et al. (2015) and Macdonald et al. (2014) are the rolling-and-soreness papers. Yoga’s back-pain meta-analyses are pain literature, not fascia literature (Cramer, Lauche, Haller et al., 2013; Anheyer, Haller, Lauche et al., 2022).

Pallarés et al. (2020) reported lower pain with full squat than with partial squat after prolonged training — a loading finding, not a stretching finding. Grade for stretching as a DOMS preventer: not established; the Cochrane synthesis is the load-bearing negative. Grade for stretching as an analgesic ritual: plausible as a sensory intervention; not a tissue diagnosis.

21 Injury prevention

This is the claim that has cost the most training time.

Pope, Herbert, Kirwan and colleagues randomised 1,538 Australian Army recruits to a pre-exercise stretching programme or to a control warm-up without the stretches. The hazard ratio for lower-limb injury was 0.95 (95% confidence interval 0.77 to 1.18) (Pope, Herbert, and Kirwan, 2000). Herbert and Gabriel’s systematic review of stretching before and after exercise for soreness and injury risk did not find a useful preventative effect (Herbert and Gabriel, 2002). Thacker, Gilchrist, Stroup and colleagues’ systematic review of stretching and sports-injury risk reached the same practical conclusion (Thacker et al., 2004). Lauersen, Bertelsen, and Andersen’s meta-analysis of exercise interventions to prevent sports injuries is the contrast that matters: stretching relative risk 0.963, strength training 0.315, proprioception 0.550 (Lauersen, Bertelsen, and Andersen, 2014). Jamtvedt et al. (2010) is the large pragmatic negative on injury. McHugh and Cosgrave (2010) left room for a strain-specific question that still needed better trials.

Witvrouw et al. (2003) found tightness associated with later muscle injury in professional soccer. That is a risk factor. Timmins et al. (2016) found short biceps femoris fascicles and eccentric weakness associated with later hamstring strain. That is a different risk factor, and one that eccentric training actually targets (Bourne et al., 2017). Soligard et al. (2008) and Emery et al. (2015) are neuromuscular-training positives. Fradkin, Gabbe, and Cameron (2006) found the generic warm-up-for-injury randomised record too thin to treat “warm up” as a completed proof.

Grade: established that generic stretching programmes do not meaningfully reduce overall injury incidence in the large randomised and pooled record. Strongly supported that strength and neuromuscular training do, in defined sports. Plausible that a sport-specific ROM deficit that blocks a required task can contribute to load error; that is not the same as “stretch to prevent injury.”

22 Older adults

Older adults lose ROM, lose strength, and lose balance. Those are not one problem.

Feland et al. (2001) showed that hamstring-stretch duration still moves ROM in people in their eighth and ninth decades. Stathokostas, Little, Vandervoort et al. (2012) found that flexibility training’s translation into functional ability is inconsistent. Stathokostas, Theou, Little and colleagues’ scoping review of physical-activity-related injuries in older adults is a different question again (Stathokostas, Theou, Little et al., 2013). Gajdosik, Vander Linden, and McNair (2005) described calf viscoelasticity in older women. Harvey, Katalinic, and Herbert (2017) found stretch for contracture clinically unimpressive. Garber et al. (2011) still include flexibility and neuromotor exercise in the ACSM adult position stand — a society judgement, not a function RCT.

Yoga and balance-challenging exercise have a clearer functional story in older samples than sit-and-reach training does (Tiedemann et al., 2013; Youkhana et al., 2016). Grade: strongly supported that ROM can still be trained late in life; emerging to mixed that isolated stretching improves the tasks people actually fail; strongly supported that balance-challenging activity is the more relevant functional lever.

23 Hypermobility

Greater ROM is not a universal good. Beighton, Solomon, and Soskolne described articular mobility in an African population and gave the field its nine-point screen (Beighton, Solomon, and Soskolne, 1973). The 2017 international classification of the Ehlers–Danlos syndromes distinguishes thirteen subtypes; hypermobile EDS is a clinical diagnosis, not a yoga achievement (Malfait, Francomano, Byers et al., 2017; Bloom, Byers, Francomano et al., 2017). Castori, Tinkle, Levy and colleagues offered a framework that separates joint hypermobility as a trait from related syndromes (Castori et al., 2017). Tinkle, Castori, Berglund and colleagues described the natural history of hypermobile EDS (Tinkle et al., 2017). Juul-Kristensen, Schmedling, Rombaut, Lund, and Engelbert reviewed measurement properties of GJH classification methods and found them imperfect (Juul-Kristensen et al., 2017). Russek and Errico found generalised joint hypermobility in 26.2% of a college sample, which is not the same prevalence as joint-hypermobility-syndrome symptoms (Russek and Errico, 2016). Scheper, de Vries, Verbunt and colleagues treated chronic pain in hypermobility syndrome and hypermobility-type EDS as a clinical challenge, not a flexibility deficit (Scheper et al., 2015).

Grade: established that hypermobility is a trait and, in some people, a syndrome. Stretching a hypermobile person toward a universal athletic ROM is the wrong intervention wearing the right marketing.

24 Sport specificity

Sport decides which ranges matter.

Throwing shoulders remodel. Glenohumeral internal-rotation deficit and total-arc loss, not “tight pecs” as a brand, are the quantities Wilk, Macrina, Fleisig and colleagues related to injury in professional pitchers (Wilk et al., 2011). Kibler, Sciascia, and Thomas described pathogenesis and the acute throwing response (Kibler, Sciascia, and Thomas, 2012). Soccer has a hamstring-strain problem that tracks fascicle length and eccentric strength more cleanly than it tracks sit-and-reach (Timmins et al., 2016; Witvrouw et al., 2003; Soligard et al., 2008). Sprint and jump sports are the setting in which long pre-effort static stretch is most likely to be a bad bargain (Winchester et al., 2008; Simic, Sarabon, and Markovic, 2013). Dance and gymnastics live near the hypermobility border; their ROM demands are occupational, not a public-health target.

Gleim and McHugh (1997) already said there is no scientifically based universal flexibility prescription. Thirty years of trials have not produced one. They have produced sport-specific risk factors, task-specific warm-ups, and a large ROM literature that is weaker on injury and performance than the market requires.

25 Tight-muscle folklore

“Tight muscles” is a folk diagnosis that collapses stiffness, shortness, stretch tolerance, guarding, and untrained end-range strength. Magnusson, Simonsen, Aagaard et al. (1997) showed that people labelled tight differed in tolerance, with stiffness higher only in a common range. Folpp et al. (2006) increased tolerance without increasing extensibility. Konrad and Tilp (2014) increased ROM without changing the ultrasound architecture that would make “the muscle got longer” true. Reiner et al. (2024) found muscle stiffness a poor determinant of leg ROM.

A protective spasm in pain is not a short sarcomere. A person who dislikes the stretch sensation is not a fascia patient. A soccer player with a short biceps femoris fascicle has a measurable architectural risk (Timmins et al., 2016) — and that is still not the same as “tight hamstrings” on a treatment card. Grade for “tight muscles” as an explanation of ordinary ROM limits: weak, except where a named tissue property has been measured.

26 Posture-based flexibility

The claim that a postural shape — forward head, rounded shoulders, anterior pelvic tilt — is caused by tight muscles that stretching will release, after which pain and performance will follow, is a chain of unforced inferences. Lederman’s critique of core-stability mythology is the nearest harvested paper to this commercial posture religion: motor-control slogans outran the evidence (Lederman, 2010). Stretching a chest wall because a photograph looks “closed” is not the same experiment as changing glenohumeral rotation in a pitcher (Wilk et al., 2011; Kibler, Sciascia, and Thomas, 2012). Sit-and-reach is not lumbar posture (Mayorga-Vega, Merino-Marban, and Viciana, 2014).

Grade: speculative as a causal account of pain and performance; not established that stretching the attributed “tight” muscle corrects a postural photograph in a way that changes clinical outcomes.

27 Universal mobility ideals

Not everyone needs more ROM. Hypermobility is the existence proof (Beighton, Solomon, and Soskolne, 1973; Malfait et al., 2017; Russek and Errico, 2016). Harvey, Katalinic, and Herbert (2017) showed that even when ROM is pathologically reduced, stretch as a contracture treatment moves angles by amounts that fail clinical importance. Older adults need function more than centimetres (Stathokostas, Little, Vandervoort et al., 2012). Athletes need the ranges their sport loads, and they need to own those ranges under force (Pallarés et al., 2020; Timmins et al., 2016). ACSM’s inclusion of flexibility exercise in an adult activity stand is a public-health inclusion, not a command to chase splits (Garber et al., 2011).

Afonso, Blazevich, and Behm (2026) and Warneke, Thomas, Blazevich et al. (2025) are the recent language police: stop equating ROM with flexibility and stretching with flexibility training. This article’s thesis is the same split applied to marketing. Clinically meaningful limitation is a task the person cannot perform, or a sport demand they cannot meet without a costly compensation. It is not a failed Instagram pose.

28 Claim ledger

The red-team that generated these verdicts, and the resolution of each folklore question, is filed with the apparatus, not reprinted as a second article.

References

References are compiled from NCBI records verified 20 August 2026. In-text citations are author–year. The numbered list is generated at build from the verified store.