Background: In the present study, we aimed to clarify the impact of the
Assisted swim (A-swim; towing from propulsion direction) on front crawl
performance at maximum and sub-maximum effort levels. Methods: Fourteen
male collegiate swimmers (age, 21.0
Assisted swim (A-swim) refers to pulling a swimmer to assist their propulsion from the direction of travel using a towing device or a rubber tube, and it is considered sprint training in competitive swimming [1, 2, 3]. Towing swim with towing power greater than the swimmers’ propulsion allows swimmers to reach faster velocities than that achieved by the swimmers themselves . Therefore, it is more familiar training for short-distance swimmers than long-distance swimmers.
Swim velocity (V) is obtained as a product of stroke rate (SR) and stroke length (SL) , which is used to analyze swim performances. Generally, SR and SL are in a tradeoff relationship with V (Eqn. 1).
Therefore, even if SR could be increased, anticipated V improvement cannot be achieved if SL decreases. For example, A-swim reportedly increased V, SL, and SR, and decreased the hand depth . However, when A-swim was incorporated in training for 3 weeks, V did not change while SR increased dramatically and SL decreased dramatically . Although A-swim increased V and SL, propulsive force (mechanical power) exerted by the swimmer to push the water was low . Considering the abovementioned fact, compared to normal swim (N-swim), the training effect of A-swim for a certain period remains unclear; however, A-swim would be expected to increase in swimmers’ V, SR, and SL without depending on their propulsive forces to keep up with the towing by the device .
As a similar index of swimmers’ arm movements, there is a report that broke down arm strokes into several phases and analyzed periodic fluctuations in the velocity of each stroke cycle and the ratio of time required for each phase [6, 7, 8, 9]. An attempt to clarify the impact of A-swim on SR and SL by breaking down the stroke movements into phases may lead to a new challenge in swim assessment. Since resistance during underwater propulsion increases in proportion to square ~ cube of V [10, 11]. A-swim causes higher resistance on the swimmer’s body than N-swim. Therefore, it impacts the swim velocity fluctuations within one stroke cycle, and as a result, a hypothesis holds that the stroke index and duration of each movement in A-swim are different from N-swim. In actual training, A-swim refers to attaching a rubber tube onto a belt around the swimmer’s waist where someone pulls the rubber tube from the ground [2, 12]. In addition to the rubber tube stretching, the force with which coaches pull the swimmer, and the level of effort by each swimmer vary. In other words, A-swim practiced currently relies on a subjective view of instructors and swimmers in terms of the load and points of focus and it cannot be confirmed that it is an established training method with universal significance. In spite of the considerable degree of evidence on resistance swim similar to this towing swim [13, 14, 15, 16, 17, 18, 19, 20, 21], findings on A-swim are rare; thus, detailed reports on the subject are warranted.
Under such a circumstance, in the present study, we aimed to clarify the impact of A-swim on performance at two levels of effort: maximum (Max) and sub-maximum (Submax). We hypothesized that swimming under A-swim conditions can promote a faster V, faster SR, and longer SL at Max and Submax.
Fourteen well-trained male collegiate swimmers voluntarily participated in the
present study. Their mean age, height, body mass, and FINA points at long course
50 m freestyle were 21.0
The trial was a 25-m front crawl, consisting of A-swim in which swimmers were
towed in the propulsion direction using a towing device (Torrent E-Rack Swim
Power Trainer; Trilabs Inc., Walnut Creek, CA, USA) installed above the pool wall
and N-swim in which swimmers swam without any towing [3, 8, 14, 23] (Fig. 1).
Swimmers were instructed to swim at Max and Submax. For Submax, we referred to a
previous study  and fluctuations in front crawl short-distance race  and
set SR at 80% of Max. Therefore, all swimmers swam four lengths of 25-m front
crawl with a 30-minute break in between. Swimmers swam A-swim after N-swim to
smoothen the progress of this experiment; however, the order of level of efforts
(setting of SR) was randomized. The swimmers repeated all the trials six times
and were instructed to not breathe. Before the experiment, it was confirmed that
there was no obvious difference in the rhythm of left-right strokes following a
visual observation of the three experienced swimmers (
Apparatus used for the measurements. (A) Assisted swimming (A-swim) in which swimmers were towed in the propulsion direction using a towing device. (B) Free swimming (N-swim) in which the swimmers were not towed (zero load) by a towing device. 1 = Personal Computer, 2 = towing device.
We connected a non-stretchable nylon cable to a belt on swimmers’ abdomen. Swimmers were towed by the towing device at the end of the cable installed at the edge of the pool. During N-swim, swimmers swam as they pulled non-stretchable cables of the device (Fig. 1). The SR of A-swim at Submax was controlled in the range below 80%–90% of N-swim at Max. The SR was measured using a small audible waterproof metronome (Tempo Trainer Pro; FINIS, Inc., Tracy, CA, USA) which swimmers wore under their cap .
The towing device was connected to a computer using a USB cable for control.
Swim velocity and travel distance was input into the computer at the sampling
frequency of 100 Hz using analytical software, Dasylab (Version 13.0, Measurement
Computing Corporation, Norton, MA, USA). To analyze V (m
In the present study, we referred to the coordination index of stroke movements reported by Chollet et al.  and Seifert et al.  and divided one stroke cycle of the left arm with the camera into the following four phases; we then calculated the duration of each stroke phase: (1) Glide phase: from the moment the hand enters the water to when the hand starts moving opposite to the direction of travel, (2) Pull phase: from the moment the hand begins to move opposite to the direction of travel to when it reaches right below the shoulder, (3) Push phase: from the moment the hand is right below the shoulder to when it leaves the water near the thigh, and (4) Recovery phase: from the moment the hand leaves the water to when it enters the water. For the duration of each phase, the mean of two-stroke cycles was analyzed.
All values were expressed either as mean
Significant interaction was confirmed in V (p = 0.001, F = 18.887,
|Mean (SD)||95% CI||Mean (SD)||95% CI|
||Submax||1.54 (0.08)||[1.50, 1.59]||2.22 (0.07)||[2.18, 2.26]|
|Max||1.64 (0.08)||[1.59, 1.68]||2.26 (0.06)||[2.23, 2.30]|
|SR (Hz)||Submax||0.84 (0.04)||[0.81, 0.86]||0.84 (0.04)||[0.82, 0.86]|
|Max||0.98 (0.05)||[0.95, 1.01]||1.03 (0.07)||[0.99, 1.07]|
|SL (m)||Submax||1.84 (0.13)||[1.77, 1.92]||2.64 (0.15)||[2.56, 2.73]|
|Max||1.69 (0.13)||[1.61, 1.76]||2.20 (0.17)||[2.10, 2.30]|
|V, swimming velocity; SR, stroke rate; SL, stroke length; Submax, submaximal effort; Max, maximal effort; N-swim, swimming without towing device; A-swim, swimming with towing device; 95% CI, 95% confidence interval.|
Duration of each stroke phase confirmed significant interaction at Submax
(p = 0.019, F = 4.103,
|Mean (SD)||95% CI||Mean (SD)||95% CI|
|Glide||Submax||0.37 (0.09)||[0.32, 0.42]||0.37 (0.07)||[0.33, 0.41]|
|Max||0.24 (0.06)||[0.21, 0.27]||0.24 (0.05)||[0.21, 0.27]|
|Pull||Submax||0.27 (0.06)||[0.23, 0.30]||0.24 (0.04)||[0.21, 0.26]|
|Max||0.25 (0.03)||[0.21, 0.27]||0.23 (0.04)||[0.21, 0.26]|
|Push||Submax||0.28 (0.03)||[0.26, 0.30]||0.25 (0.02)||[0.24, 0.27]|
|Max||0.27 (0.02)||[0.26, 0.29]||0.23 (0.03)||[0.21, 0.25]|
|Recovery||Submax||0.29 (0.04)||[0.27, 0.32]||0.34 (0.04)||[0.31, 0.36]|
|Max||0.27 (0.03)||[0.28, 0.30]||0.29 (0.02)||[0.28, 0.30]|
|Glide, glide phase; Pull, pull phase; Push, push phase; Recovery, recovery phase; Submax, submaximal effort; Max, maximal effort; N-swim, swimming without towing device; A-swim, swimming with towing device; 95% CI, 95% confidence interval.|
Duration of each stroke phase (s). (A) Submax and Max effort
during N-swim (solid line) and A-swim (dotted line). (B) Data are means
In the present study, we aimed to clarify the impact of A-swim at Max and Submax effort levels on performance. The main finding of this study was that the higher values were confirmed in V, SR, and SL in A-swim compared to those of N-swim at both of Max and Submax. These results supported previous studies in the literature [3, 5] and verified our hypotheses.
On the body of swimmers moving forward in the water, in addition to their thrust for propulsion [26, 27], there exists resistance that increased in proportion to square to the cube of the propulsion velocity [10, 11]. To illustrate, among the total resistance on the body, the thrust required to propel forward becomes propulsion [26, 27, 33]. Conversely, in A-swim, swimmers constantly swim with thrust. Therefore, in order to increase V, propulsion more than towing must be created by the limbs. In the present study, swimmers had significantly higher V at Max than Submax for both N-swim without the towing device and A-swim with the towing device. In conclusion, it is evident that the towing force of 24 lbs is a load that male competitive swimmers with a FINA POINT of approximately 500 who participated in this study can adjust their V in voluntary.
The arm movements in front crawl can be classified into three patterns based on the coordination of both arms ; however, in all patterns, the velocity of the front crawl was not constant, wherein acceleration and deceleration repeat dependent on the movements of strokes [26, 27]. Focusing on the changes brought about by the level of efforts with/without towing revealed the presence of a significant interaction between V, SR, and SL. In addition, even if SR is controlled, V and SL were drastically higher when towed. This result indicated that with towing, even in phases with no acceleration and those in which the swimmer might decelerate, constant towing from the front results in a difference in acceleration within one stroke cycle between A-swim and N-swim. In other words, in A-swim, the towing cable can move the swimmers forward constantly, regardless of swimmers’ arm-stroke position, even in the recovery phase when no propulsion is generated against the water.
We then examined the changes in the duration of each stroke phase from the point of towing. Narita et al.  reported that when swimmers’ SR was controlled using a small audible waterproof metronome, swim itself became different from N-swim in some cases. In other words, as observed in the present study, in A-swim at Submax, as swimmers synchronized their arm movements to the metronome used to recreate the SR of N-swim, it might have impacted the difference in the duration of the recovery phase.
In terms of A-swim at Max, Williams et al.  reported that the position of stroke was shallower than that of N-swim, increasing SR. In the present study, although we cannot discuss changes in the stroke arm positions, the results supported those reported by a previous study . High SR in A-swim might be explained by the difference in duration of the recovery phase at Submax and time difference in the push phase at Max. For example, by adjusting the coordination of arm strokes and kicking, the point at which the arm leaves the water might move closer to the head. However, even when comparing SL at Max between A- and N-swim, SL is considerably longer under towing. Therefore, details of the movement difference by SL cannot be determined. Prior studies [1, 34, 35] that examined the arm angles of competitive swimmers from the side reported that swimmers moved their hands, forearms, and upper arms perpendicular to the floor to push the water and gain thrust. Therefore, the swimmers gained an increased area of the forward project in the direction of propulsion. In other words, in A-swim, by being towed from the propulsion direction, water collides with the surface of the arm in the propulsion direction, possibly promoting the acceleration of the push phase. If this inference is correct, when synchronizing the stroke with the metronome rhythm, adjustments were being made in the recovery phase. To verify this inference properly, a difference of both arms in N-swim and A-swim should be confirmed, and pressure difference in both sides of hands should be quantified using a pressure sensor [26, 27, 36, 37]. In any case, A-swim contributes to increasing swimmers’ SR as well as SL and V.
Since SL of A-swim is longer than that of N-swim, its effectiveness as a training to improve swimmers’ technique has been indicated [5, 12]. In contrast, the present result showed that A-swim significantly increases SR of swimmers compared to N-swim. Futhermore, Girold et al.  reported that 3 weeks of A-swim training increased SR and strength at a high velocity. Therefore, A-swim could also be an effective training method to improve the rotation of arms. However, even if SR is increased with A-swim, SL might decrease . Additionally, it has been reported that A-swim may change the stroke pattern [3, 12], slow down the hand speed, and/or cannot maintain the kinetics of the hand movement . Thus, A-swim may be masking the decrease in propulsive force. Actually, in semi-tethered swimming, there are concerns that the coordination of strokes (coordination pattern) may be changed  and that the propulsive force exerted by the swimmer may be decreased [3, 22]. Regardless, when swimmers and coaches are implementing A-swim, it is imperative to focus on changes in stroke movements in addition to SR and SL.
As an assessment method of physical strength elements (powers that impact V), the tethered resisted swim is a major knowledge; however, A-swim might be considered a new evaluation method in addition to being a training method. Meaning comparing the difference in acceleration and deceleration of stroke movements using towing can clearly show the propulsion phase and deceleration phase of swimmers. In the future, a more detailed analysis of kinetics which especially include measuring the thrust of arm stroke  and kinematics of stroke movements in A-swim is anticipated. Confirming the towing force setting may also be useful. Depending on the swimmers’ performance level, it could exceed the velocity of towing.
Swimmers can achieve a faster V in A-swim compared with N-swim. For recreational swimmers who swim mainly out of interest or for health maintenance, they can train for longer distances in a fixed time through faster swimming and improved skills efficiency with A-swim. In other words, A-swim can serve as a useful training for both competitive and recreational (health conscious) swimmers. As training for health maintenance is different from performance, the use of A-swim to keep training at maximal velocity with lower energy cost, compared with N-swim, can help to reduce the physiological effort relation to recovering strategies and to avoid overtraining. This will also allow better control over training intensities.
Sex difference, given that V is a product of SR and SL, it is likely that the benefits derived from A-swim in females will be similar to males.
We conclude that A-swim with a load of 24 lbs increases SR in addition to the V and SL of swimmers compared with N-swim, even considering the impact of propulsion by towing. Furthermore, from the changes in the duration of the push phase, impact on stroke movements was indicated. There is a possibility, however, A-swim negatively change kinetics and kinematics. Therefore, for coaches and swimmers, understanding the characteristics of A-swim and carefully pay attention to changes in swimming performance due to it could be needed if introducing A-swim into training.
SIM, YW and KM conceptualized and designed the study. SIM, KM, and YT performed the research. SIM wrote the manuscript. SIM and KM analyzed the data. YT, YW, JEM, PF, HPN, and DAM reviewed and edited the manuscript. All authors read and approved the final manuscript.
All swimmers signed the informed consent from prior to participating in the study. This study was approved by the Institutional Ethics Committee of Tokyo Gakugei University, Japan (approval number 494). All human research procedures were in accordance with the ethical standards of the committee responsible for human experimentation (institutional and national), and with the Helsinki Declaration of 1975 as revised in 2013.
We would like to thank Associate professor Shunsuke Yamaji at the Fukui University for the advice on statistical matters.
This work was supported by the JSPS KAKENHI Grant Number JP21K11360 and through the Portuguese Foundation for Science and Technology (FCT), I.P., under project UIDB04045/2020.
The authors declare no conflict of interest. PF, HPN, and DAM are serving as one of the Guest editors of this journal. DAM is are serving as one of the Editorial Board members. We declare that PF, HPN, and DAM had no involvement in the peer review of this article and has no access to information regarding its peer review. Full responsibility for the editorial process for this article was delegated to AT.
Publisher’s Note: IMR Press stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.