The National Strength and Conditioning Association (NSCA) defines ROM as “the amount of rotation available at a joint” (p. 276) [1]. Similarly, the 11th edition of the American College of Sports Medicine (ACSM) Guidelines for Exercise Testing and Prescription initially defines flexibility as “the ROM available at a joint” (p. 28Footnote 1) [2], and then as “the ability to move a joint through its complete, pain-free ROM” (p. 104) [2]. In the same vein, the National Academy of Sports Medicine (NASM) defines flexibility as “the ability to move a joint through its complete ROM” (p. 162) [3]. The NSCA, however, establishes mobility as the ability to move through full ROM, and flexibility as the extensibility of the periarticular soft tissues, considering flexibility as only one of several factors limiting ROM (p. 275) [1]. Given the very heterogeneous use of the word “mobility,” which can be very broad (e.g., the timed up-and-go test [4]), we will not further discuss this concept here. Similar to the NSCA, peer-reviewed publications often use “flexibility” and “extensibility” interchangeably [5,6,7]. Finally, a very popular book on training methodology defines flexibility as “the ROM of a joint or set of joints” (p. 236) [8]. The World Health Organization was not cited because their latest guidelines for physical activity do not address these topics [9], and we also failed to find relevant material from the International Olympic Committee, which was surprising.

Across its guidelines, the ACSM interchangeably mentions measurement of flexibility and measurement of ROM, and at a certain point states that flexibility can be quantified in terms of ROM expressed in degrees (p. 105) [2]. The NSCA has a similar approach (despite the previously mentioned distinction between mobility and flexibility), and goes as far as equating ROM tests to measuring muscle length (p. 286) [1]. Bompa and Buzzichelli [8] (p. 326) also state that muscle length is what limits ROM (if only it were that simple…). However, the ACSM guidelines also state that “flexibility training is important to enhance ROM” (p. 262) [2], implying that flexibility can be trained and is therefore modifiable. Similarly, flexibility has been considered a trainable biomotor ability [8], meaning that motor learning and not only physiological or biomechanical adaptations can take place. This raises the question: Is ROM fully modifiable, or may it be limited by non-modifiable factors (e.g., bone and joint structure), and if so, should it be equated with flexibility? Or should flexibility be considered only a modifiable (and therefore trainable) part of ROM?

First, it should be acknowledged that there are non-modifiable components of ROM, namely bone and joint structure [1], which have relevant inter- and even intra-individual variations that affect ROM [10] (e.g., knee and elbow extension, or less evident features, such as an os talus secundarius, limiting subtalar ROM and causing restriction and pain in sports activities [11]). The case of elbow extension can be easily observed, as variation in bone shape and size limits this action differently from person to person [10]. A typical case affecting athletes is femoroacetabular impingement syndrome, which (regardless of being a natural anatomic variation or acquired due to repetitive movements) negatively affects hip flexion, adduction, and/or internal rotation and commonly requires arthroscopic intervention [12, 13].

Body-on-body contact (e.g., knee flexion, where the movement will stop upon contact of the posterior aspects of the leg and thigh) will also limit ROM, acting as a physical barrier [1] regardless of flexibility levels. Body size ratios may influence distance-based assessments (e.g., inhibition of sit-and-reach distance by a protuberant or distended waistline) but likely play a minor role in angle-based assessments (e.g., through goniometry). If ROM is influenced by both modifiable and non-modifiable factors, and flexibility is a modifiable factor [1,2,3, 8], then it is reasonable not to equate flexibility with ROM, and instead consider it only as one of its components.

However, flexibility as the extensibility of soft tissues [1, 5,6,7] (and considered trainable by most accounts [1,2,3, 8]) is not the only modifiable component limiting ROM. The ACSM considers flexibility a health-related physical fitness component, while it catalogs balance under skill-related physical fitness components (p. 28) [2]. However, a study asked 20 healthy young adults (14 women) to perform the toe-touch test (also known as the stand-and-reach) under standard conditions versus using a device restraining the participant against forward falls [14]. When restrained (i.e., when risk of falling was removed through use of a Velcro strap attached to a platform support), the participants showed superior ROM to the traditional toe-touch test (− 4.1 ± 4.9 cm versus − 15.2 ± 6.5 cm, respectively), and therefore the authors suggested that balance influenced flexibility [14]. Alternatively, it could be stated that fear of falling (and not balance per se) was responsible for the differences, or that using the Velcro strap may have induced postural changes and influenced ROM. Regardless, it suffices to establish that “flexibility” is not the single modifiable factor limiting ROM.

The NASM defines optimal control of movement through full ROM as dynamic ROM, and states that it depends on a mixture of flexibility and neural control of movement (p. 163) [3], and this includes muscle recruitment and balancing strength production across agonists, antagonists, synergists, and stabilizers (p. 163) [3]. Therefore, achieving full ROM relies not only on flexibility (i.e., soft tissue extensibility [1, 5,6,7]), but also on accurately managing force production, which are the typical tenets of strength training. Perhaps unsurprisingly, strength training (at least when performed to relatively long muscle lengths) has been shown to improve ROM with a magnitude comparable to that induced by stretching protocols [15,16,17], and suggests that ROM is limited by much more than soft tissue extensibility. In fact, active stretching through eccentric resistance training seems quite effective for generating ROM gains [18], with the “active” component again hinting at a large role played by neural modulation and not merely tissue extensibility.

Beyond these factors, pain can also negatively affect ROM acutely [19], which may affect athletes’ ROM after injury or surgery, for example. There is substantial variability between people in their pain tolerance and even within individuals over time (e.g., before or after exercise). Thus, is ROM primarily dependent upon how much or little discomfort someone can tolerate? It is not our intention to be exhaustive in the context of this Current Opinion article, but merely to illustrate that ROM is a wider concept than “flexibility” (in agreement with the NSCA’s proposal [1]), which is just one piece of the ROM puzzle that is more related to soft tissue (e.g., muscle, tendon, nerves) properties. Several other modifiable and non-modifiable factors influence ROM.

This distinction between flexibility and ROM has implications for the reporting of scientific findings. ROM is a measurable property—there is a motion, and that motion can be performed throughout a given range. ROM is thus an objective parameter that can be assessed externally and can be limited by factors unrelated to flexibility (e.g., bone on bone contact, typical in elbow extension). On occasion, ROM may be inferior to the available flexibility: beyond elbow extension (limited by the configuration of bony articular surfaces), a deep squat is a movement that will inevitably stop when the thigh contacts the lower leg, even if the person would theoretically have the flexibility to perform a deeper movement. We propose that studies should report ROM, as this is the objective measure that is most commonly assessed (e.g., using goniometers or inclinometers [20]) and reported (e.g., in degrees or radians). Flexibility would require assessments of soft tissue extensibility, such as ultrasound testing of fascicle length elongation [21, 22]. However, these assessments require laboratorial settings and may be impossible to perform in vivo. All in all, most assessments will in practice reflect ROM, not necessarily flexibility, thus this term should probably be used by default.