Movement and Skill Analysis of Supercross Bicycle Motocross : The Journal of Strength & Conditioning Research

Secondary Logo

Journal Logo

Original Research

Movement and Skill Analysis of Supercross Bicycle Motocross

Cowell, John F.; McGuigan, Michael R.; Cronin, John B.

Author Information
Journal of Strength and Conditioning Research 26(6):p 1688-1694, June 2012. | DOI: 10.1519/JSC.0b013e318234eb22
  • Free



For strength and conditioning coaches to address the programming requirements of athletes, a knowledge of the physiological and/or mechanical characteristics of the sport is necessary. In many sports, a “needs analysis” has been performed where the specific and dominant movement patterns have been identified (6,8). This needs analysis is a key component to understanding the kinetic and kinematic requirements of a sport, and it is from the information this analysis provides that specific programming can be derived. With regards to the bicycle motocross (BMX), this understanding is in its infancy given that the sport was introduced as an Olympic sport in 2008 and therefore before this was of little interest to sport scientists and practitioners alike. The sport began in the U.S.A. in the late 1960s, and although it has matured since then, the main characteristics of the sport such as the dimensions of the bicycle, the 8-rider mass-start format and inclusion of jumps and banked turns remain. A traditional BMX track as specified by the American Bicycle Association ( is similar in length to a Supercross BMX (SBMX) track; however, certain features and obstacles such as the start ramp (2–3 m high) are much smaller in scale. Part of the reason for this is that although the SBMX track is contested solely by professional male and female athletes, a standard track must cater to racers of all ages, often as young as 5 or 6 years old. The SBMX racecourse has an 8-m-high start ramp with a downhill slope that is initially 18° and then progresses to 28° after approximately 5 m and the track tends to be between 300 and 400 m in length, 6–8 m wide, and negotiation of the course may take anywhere from 30 to 50 seconds to complete (2,5,7,9). The race is a 1-lap event that commences when a mechanical gate falls, releasing the (maximum of) 8 riders onto a course consisting of a start hill, jumps, and banked turns. In this configuration, the Olympic caliber racers are able to reach peak power and cadence within 6.0 m and 1.6 seconds (4). Because of the high relative contribution of technique to overall success, the rider who is able to first establish the best position and successfully negotiate the obstacles on the course without having to contend with other riders are both significant influences on the outcome (9).

Although the basic characteristics of an SBMX track are known, the physiological, mechanical, and technical demands fundamental to negotiating a track have not been investigated. For the strength and conditioning practitioner, an understanding of these demands in particular the movement patterns, their frequency, and duration serves as the foundation for developing a program for performance enhancement. Given this information, the purpose of this study was to quantify the frequency, duration, and type of movement patterns and skills used in Elite SBMX racing. From this knowledge, the basis for a specific training program that uses the identified characteristics of SBMX racing can be created.


Experimental Approach to the Problem

The study was descriptive in nature and consisted of collating data of multiple riders (n = 26) from video analysis of the Time Trial (TT) event of the 2010 BMX World Championships in Pietermaritzburg, South Africa. The riders negotiated the course solo at a maximum pace, and this provided data specific to the TT component of SBMX racing where the riders are ranked according to course navigation in the shortest elapsed time.


The riders to whom we had video access who were used in this study were Union Cycliste Internationale (UCI) categorized Elite (“Elite” defined as age ≥19 years) Men (n = 16) and Elite Women (n = 10) with a mean age of 22 ± 4 and 22 ± 3 years, respectively. The TT filming occurred at the 2010 World Championships in Pietermaritzburg, South Africa. Access to the timing data is in the public domain and available from the UCI (, so written consent was not supplied.


The subjects were filmed navigating an entire SBMX course. Each rider traversed the course once in competition, and the footage captured provided a video record of the time the athletes were spending at various points on the track and the principal movement patterns. They were filmed from a perspective that allowed video capture for key strategic events along the course such as the start, jumps, and other obstacle sections that are typical of an SBMX course. Each rider was timed to specific points (based on film footage) on the track (beginning with the first movement of the bicycle and ending at the finish line) by means of individual video frame analysis where one frame represented 1/25 of a second. Video footage was captured and analyzed using Quicktime™ (Apple Inc., Cupertino, CA, USA) software and the joint angles corresponding to the respective movement patterns were quantified by VideoMotion© (Objectus Technology, PA, USA) software. The data on start gate reaction time, speed at the bottom of the start ramp, and total elapsed time were measured using Swiss Timing™ (Corgémont, Switzerland) and OMEGA Electronics™ (Biel, Switzerland) photo finish cell technology, and the times were accurate to 1/1,000 seconds and the speed to 1/10 km h. To determine reliability of data collected via notational analysis (NA), another 2 raters analyzed the video according to the procedures and criteria outlined.

The measuring points were defined for pedaling, pumping, and jumping and each section of the track. Pedaling was characterized by forward rotation of the cranks with a full revolution of the crank being one pedal stroke. Jumping was measured as the moment the front wheel leaves the ground to when the rear wheel touches down. If the riders were neither pedaling nor jumping, then they were either pumping or coasting. Although pumping, the riders were attempting to generate speed over obstacles while maintaining ground contact at all times. “Coasting” occurred just before or just after a jump. It was in this “transition” phase that the rider either sets up for the takeoff of the next jump or was moving from landing a jump to commencing pedaling. For the purposes of this study, these 2 categories were combined as “pumping.”

The start was outlined beginning with the first movement of the bicycle and ended the moment the front wheel touched the dirt surface of the track. The first straightaway was defined as the distance from where the start ramp meets the track surface to landing from the last jump. The second straightaway was measured from the end of the first corner to where the rider landed from the last jump. The third straightaway distance was quantified as starting at the end of the second corner extending to the top of the final obstacle (small jump). The fourth and final straightaway began as soon as the third corner was completed and extended to the finish line.

Statistical Analyses

Twenty-six riders were analyzed, and their respective times at specific points on the track were recorded. At any given position of the track, there was a corresponding frame number, and each frame number represented 1/25th of a second. The first movement of the bicycle at the start was defined as “Frame 0,” and the frame number at the finish line was associated with the recorded finish time of the lap. From these 2 numbers, we were able to create specific algorithms for the individual rider where a certain frame number could be converted into time. These points served as references to determine how the data in Table 1 (time to the drop/kink in the start ramp, to the bottom of the start ramp, jumping on the first straight, pedaling on the first straight, coasting/pumping on the first straight, total time on the first straight, jumping on the second straight, pedaling on the second straight, coasting/pumping on the second straight, total time on the second straight, jumping on the third straight, pedaling on the third straight, coasting/pumping on the third straight, total time on the third straight, jumping on the fourth straight, pedaling on the fourth straight, coasting/pumping on the fourth straight, and total time on the fourth straight) was measured.

Table 1:
Temporal characteristics of a Supercross BMX race (men, n = 16, women, n = 10).

The movement patterns at the start and during takeoff/landing/pumping and time spent pedaling, jumping and “pumping/coasting” were determined. Movement patterns were also categorized as hip flexion/extension, knee flexion/extension, and shoulder horizontal abduction/adduction. Each time one of these movement patterns occurred, it was quantified, and this calculation was based on the characteristics of the track. Each pedal stroke would have a corresponding knee action and each jump or obstacle would have a corresponding action at the torso. The movement at the start was unique, so it was categorized separately.

Means and SDs were used to represent centrality and spread of data. The statistical analysis was principally descriptive in nature. Reproducibility of the results was established by comparing the variables of interest as quantified by 3 separate raters. The interrater reliability was determined using a coefficient of variation (CV = [SD/mean] × 100) and the mean ± SD was calculated using Microsoft Excel™ (Microsoft Corporation, Redmond, WA, USA) for pedaling (seconds), jumping (seconds), and pumping/cornering (seconds). The CV was not determined for reaction time, speed, or total time because these figures were supplied by the UCI and therefore not calculated based on video footage.



As can be observed from Table 1, the CVs of all the variables of interest ranged from 0.00 to 5.87%. In terms of the 3 skill levels of interest, which represented the cumulative data, CVs for pedaling, jumping, and coasting for the men were 0.71, 0.96, and 2.14%, and 1.75, 1.56, and 0.87% for the women, respectively.

Movement-Skill Analysis

The specific values (times and speed) for each of the variables of interest for men and women can be observed in Table 1. The Elite Men riders averaged 39.67 ± 0.81 seconds and 30.45 ± 3.20 pedal strokes to complete the SBMX track. Of this time, the least amount of time was spent on the fourth and final straightaway, which was also the shortest, and the volume of pedaling, jumping, and coasting/pumping decreased proportionally as detailed in Figure 1A. The quantity of pedaling on a straightaway gradually decreased from approximately 7 to approximately 0.65 seconds. This same descending trend was similar for jumping with the riders spending approximately 3 seconds in the air on the first straight and <0.5 seconds on the fourth straight. Coasting/pumping on the other hand gradually increased as the race progressed (from ∼3.7 seconds on first straight to ∼7.4 seconds on the third straight).

Figure 1:
Time breakdown of skill spent per straightaway for (A) male (n = 16) and (B) female (n = 10) riders.

The Elite Women riders competed on a modified track where their second and third straightaways were different from that used for the men, thus allowing them approximately 3 more total pedal strokes than for the men. Despite this difference, a similar trend in skill use was observed as depicted in Figure 1B. The women pedaled for approximately 5.49 seconds on the first straight and approximately 0.49 seconds on the third straight. They spent no time at all jumping on the third straight compared with approximately 3 seconds on the second straight. The third straight was predominantly pumping/coasting with a minor amount of pedaling (∼1.7 seconds).

The relative and unique movement of the body at the start is represented in Figure 2 (hip extension and shoulder adduction) and predominant movement patterns for the lower and upper body were hip and knee extension/flexion (∼30 times per lap) and horizontal shoulder abduction/adduction (20 times per lap), respectively (Figure 3).

Figure 2:
Movement patterns in the start position of the track: A) setup and B) movement.
Figure 3:
Movement patterns in the takeoff/landing and pumping portions of the track. A) Ascent/setup, (B) Landing/descent.


With regards to the accuracy of the data presented, 3 raters analyzed the video footage with reference to the criteria described previously and detailed in Table 1. The CV was used as a representation of the percentage of variation among the 3 raters. In this study, the CV ranged from 0% to a maximum of 5.87% for the variables of interest (Table 1). It would seem that the procedures outlined for this NA are of an acceptable reliability.

An SBMX race is composed of 5 sections: the start, the first straight, the second straight, the third straight, and the finishing straight. Each of these sections requires a slightly different skill set, and the performance on one section is likely to influence performance on subsequent sections. The start in an SBMX race is the section of the race that sets the tone for all things to follow. The attempt to generate the most speed out of the start gate created a movement pattern that was unique to the entire remainder of the race and how well the start is executed appears foundational for overall success. In the TT, the start was crucial to overall elapsed time; in a race however where the rider position among his or her competitors is the determiner of success as opposed to total elapsed time, the rider was attempting to generate as much speed as quickly as possible to best establish position on the track against the other competitors.

After descending the start ramp, a distance of approximately 10 m exists before the take-off of the first jump, and roughly half of the riders chose to pedal there with no clear advantage or disadvantage in doing so as determined by the total elapsed time. The entire distance of the first straightaway was approximately 80 m and consisted of 3 jumps. The distances between the jumps and the scale of the jumps dictated that the riders spent almost as much time in the air as they did pedaling (∼36 vs. ∼43%, respectively). If the start ramp was included, the riders spent more time on the first straightaway than on any other.

The second straightaway consisted of 4 jumps and started with a 2- to 3-second effort of pedaling out of the first corner and to the first jump. Even though the jumps tended to be shorter in distance, they were higher thus sending the riders higher in the air than did the jumps on the first straightaway. Also, because the jumps were in quick succession, there was little time to pedal. The riders were jumping approximately 42% of the time (and another ∼30% of the time coasting or pumping) leaving only approximately 28% of the time for pedaling. Because of the increased time spent in the air during the second straightaway, jump setup, flight trajectory, and landing technique appeared to determine elapsed time. The second straightaway was unique in that it split into 2 paths. The Elite Men took 1 path, often called the “Pro Straight,” whereas the Elite Women took the other called “the Women's Section.” Despite their differences, both straightaways had similar skill requirements.

The third straightaway, often referred to as the “rhythm section,” was very similar to the second straightaway in that it was split into one section for the Elite Men and one for the Elite women. The third straightaway may have been the most physically demanding of the 4 because of the constant effort the riders exerted. Typically, the third straightaway had very little jumping and therefore very little time for recovery from pumping effort. The riders were pumping often as much as approximately 60% (for the men) and approximately 95% (for the women) of the time depending on the track and the rider. It has been suggested that jumping provides some degree of recovery and because the third straightaway had none or almost none, the athletes were working for 8.52 ± 1.13 seconds, the longest continuous and concentrated effort of the entire lap (1).

The finishing or fourth straightaway was the shortest in distance, and TT outcome did not appear to be altered during this section. The average time spent on the fourth straight was 4.78 ± 0.48 seconds, and this was broken down into approximate values of 54, 41, and 5% coasting/pumping, pedaling, and jumping respectively.

Three predominant movement patterns, hip flexion/extension (∼30 times per lap), knee flexion/extension (∼30 times per leg per lap), and shoulder horizontal abduction/adduction (∼20 times per lap) were observed throughout the TT. These movements were used often in conjunction with pedaling and in nonpedaling portions of the track. From the video analysis, it was found that the number of knee flexion/extensions that occurred was a by-product of the number of pedal strokes; therefore, there was interrider variation, whereas the number of hip flexion/extensions and horizontal shoulder abduction/adductions were determined by the quantity of jumps on the track, so there was no variation among the riders. Hip and knee extension/flexion and the horizontal shoulder abduction and adduction occurred during pedaling and during pumping. At the start, the riders were staged in a somewhat crouched position with their hips just behind the seat at an angle of approximately 90° to the torso as shown in Figure 2A. As the start cadence sounded and the gate dropped, the rider drove the hips forward, virtually meeting the handlebars at an angle of approximately 130° to the torso, and the knee angle remained at a fairly constant approximately 165° through this movement (Figure 2B).

The movement of the arms in relation to the torso while in the jumping and pumping sections of the track is illustrated in Figure 3. On the ascent of a jump, the arms were tucked with the hands close to the shoulder at an angle of approximately 60–70° of abduction to the torso (Figure 3A). Upon landing or descending the jump, the rider further abducted the shoulders to an angle of approximately 90–125° (Figure 3B) to the torso, depending on the type of jump or landing. In the “Rhythm” section (third straightaway) of the track where the rider was pumping, the objective of the rider was to generate as much speed as possible by alleviating gravitational forces on the ascent (by “absorbing” the take-off) and accentuating the gravitational forces (by pushing down on the bike) on the descent. In other words, the rider allowed the bike to come close to him or her in an attempt to absorb the jump and then pushed the bike away on the landing to accelerate down the backside of the jump.

The riders were pedaling 32.7 ± 2.7 pedal strokes per lap and despite the approximate 8% difference among the riders and this skill being only used for approximately 30% of the track, this appeared to have a significant impact on the other approximately 70% of the race. Furthermore, because of the complex nature and technical demands of the sport, not only was it important to be able to demonstrate the transgression of skills during a lap, as shown in Figure 1, but also to understand more deeply the underlying statistics as shown in Table 1, for it is from these figures that the reader can gain the deepest understanding of the underlying processes involved in a lap of an SBMX track.

Certain limitations of this particular study are (a) data were derived from one specific SBMX track, which will most likely have unique properties; (b) the analysis was of a TT where the rider is solo and does not have to contend with the milieu of as many as 7 other riders and he or she will have to contend with a maximum of 7 other riders, and the corresponding riding behavior will most certainly have to adapt accordingly; (c) the 25-Hz video recording may not have been precise enough to capture the video data in such detail as to identify greater differences in technique or expose greater variation in interrater reliability; and (d) because much of the video was obtained from an external source, the consistency of the means by which it was captured cannot be completely verified.

It is apparent from our analysis and the status of research into SBMX that there is a need for further research using more precise means of video NA and better technology to capture cadence, speed, accelerometer, and power data. Optimizing strength and conditioning programming necessitates that certain variables pertaining to performance in SBMX must be quantified. With accurate capture of power output, and when the respective characteristics of power output as they pertain to performance (peak power, time to peak power, repeat sprint/power output ability, fatigue, power:cadence relationships) have been reliably quantified, then the interventions aiming to increase performance in these characteristics can be appropriately evaluated. For example, when attempting to determine whether pedaling in a certain portion of the track is advantageous, the pedaling cadence to power output and acceleration relationship would be beneficial to evaluate the efficacy of such a strategy.

As the ability to understand the relationship between these variables increases, specific on-the-bike coaching can develop. Furthermore, with the dominant movement patterns exhibited by the riders having been identified, will increasing strength and/or power in these movements increase the performance benefit of the respective skill? It is suggested that future investigation into developing strength and power of the riders as demonstrated in movements specific to SBMX to overall performance be conducted.

Practical Applications

An entire SBMX race is completed in ≤40 seconds. Within this timeframe, the efforts of the riders are intermittent and varied. Unlike other cycling events where the riders are pedaling virtually the whole time, SBMX racers are only pedaling for 30–38% of the race, that is, approximately 10–14 seconds. With regard to the principle of specificity, it is the belief of the authors that the conditioning of an SBMX athlete should be specific to these physiological requirements and although aerobic conditioning may have some benefit particularly in recovery between multiple races (∼5–6 races per day), we believe that the conditioning program should focus almost primarily on anaerobic alactic and lactic energy system development (3).

Exercises that strengthen the hip and knee extensors such as the squat or deadlift (start and pedaling) and the muscles responsible for horizontal shoulder abduction and adduction, for example, the bench press and barbell row (pumping and jumping sections) may be beneficial to SBMX performance. Strength and conditioning specialists may want to pay particular attention to bar movement velocity because it pertains to the respective skill, be it pumping or the start. Specific attention to eccentric strength for landing the jumps and hypertrophy for crash-induced injury resistance may also be indicated. Further specificity of programming would benefit from an understanding of when, how, and how often these actions take place.


1. Allen H, Coggan A. Training and Racing with a Power Meter (2nd ed.). Boulder, CO: Velo Press, 2010. pp. 254.
2. Campillo P, Doremus T, Hespel JM. Pedaling analysis in BMX by telemetric collection of mechanic variables. Braz J Biomotricity 1: 15–27, 2007.
3. Delecluse C, Van Coppenolle H, Willems E, Van Leemputte M, Diels R, Goris M. Influence of high-resistance and high-velocity training on sprint performance. Med Sci Sports Exerc 27: 1203–1209, 1995.
4. Herman C, McGregor SJ, Allen H, Bollt EM. Power capabilities of elite bicycle motocross (BMX) racers during field testing in preparation for 2008 Olympics. Med Sci Sports Exerc 41: 306–307, 2009.
5. Hodgkins T, Slyter M, Adams K, Berning J, Warner S. A comparison of anaerobic power and ranking among professional BMX racers. Med Sci Sports Exerc 33: S246, 2001.
6. Iwai K, Okada T, Nakazato K, Fujimoto H, Yanamoto Y, Nakajima H. Sport-specific characteristics of trunk muscles in collegiate wrestlers and judokas. J Strength Cond Res 22: 350–358, 2008.
7. Slyter M, Pinkham KK, Adams KJ, Durham MP, Moss CK, Wenger TE. Comparison of lower body power output between expert and professional bicycle motor cross racers. Med Sci Sports Sci 33: S157, 2001.
8. White D, Olsen P. A time motion analysis of bouldering style competitive rock climbing. J Strength Cond Res 24: 1356–1360, 2010.
9. Zabala M, Sanchez-Munoz C, Mateo M. Effects of the administration of feedback on performance of the BMX cycling gate start. J Sports Sci Med 8: 393–400, 2009.

video; cycling; jumping; pedaling

Copyright © 2012 by the National Strength & Conditioning Association.