Maximal dismounts from high bar

In men's artistic gymnastics the triple straight somersault dismount from the high bar has yet to be performed in competition. The present study used a simulation model of a gymnast and the high bar apparatus (J. Appl. Biomech. 19(2003a) 119) to determine whether a gymnast could produce the required angular momentum and flight to complete a triple straight somersault dismount. Optimisations were carried out to maximise the margin for error in timing the bar release for a given number of straight somersaults in flight. The amount of rotation potential (number of straight somersaults) the model could produce whilst maintaining a realistic margin for error was determined. A simulation model of aerial movement (J. Biomech.23 (1990) 85) was used to find what would be possible with this amount of rotation potential. The model was able to produce sufficient angular momentum and time in the air to complete a triple straight somersault dismount. The margin for error when releasing the bar using the optimum technique was 28 ms, which is small when compared with the mean margin for error determined for high bar finalists at the 2000 Sydney Olympic Games (55 ms). Although the triple straight somersault dismount is theoretically possible, it would require close to maximum effort and precise timing of the release from the bar. However, when the model was required to have a realistic margin for error, it was able to produce sufficient angular momentum for a double twisting triple somersault dismount.     

The margin for error when releasing the high bar for dismounts

In Men's Artistic Gymnastics the current trend in elite high bar dismounts is to perform two somersaults in an extended body shape with a number of twists. Two techniques have been identified in the backward giant circles leading up to release for these dismounts (J. Biomech. 32 (1999) 811). At the Sydney 2000 Olympic Games 95% of gymnasts used the "scooped" backward giant circle technique rather than the "traditional" technique. It was speculated that the advantage gained from the scooped technique was an increased margin for error when releasing the high bar. A four segment planar simulation model of the gymnast and high bar was used to determine the margin for error when releasing the bar in performances at the Sydney 2000 Olympic Games. The eight high bar finalists and the three gymnasts who used the traditional backward giant circle technique were chosen for analysis. Model parameters were optimised to obtain a close match between simulated and actual performances in terms of rotation angle (1.2 degrees ), bar displacements (0.014 m) and release velocities (2%). Each matching simulation was used to determine the time window around the actual point of release for which the model had appropriate release parameters to complete the dismount successfully. The scooped backward giant circle technique resulted in a greater margin for error (release window 88-157 ms) when releasing the bar compared to the traditional technique (release window 73-84 ms).     

Optimization of backward giant circle technique on the asymmetric bars

The release window for a given dismount from the asymmetric bars is the period of time within which release results in a successful dismount. Larger release windows are likely to be associated with more consistent performance because they allow a greater margin for error in timing the release. A computer simulation model was used to investigate optimum technique for maximizing release windows in asymmetric bars dismounts. The model comprised four rigid segments with the elastic properties of the gymnast and bar modeled using damped linear springs. Model parameters were optimized to obtain a close match between simulated and actual performances of three gymnasts in terms of rotation angle (1.5 degrees ), bar displacement (0.014 m), and release velocities (<1%). Three optimizations to maximize the release window were carried out for each gymnast involving no perturbations, 10-ms perturbations, and 20-ms perturbations in the timing of the shoulder and hip joint movements preceding release. It was found that the optimizations robust to 20-ms perturbations produced release windows similar to those of the actual performances whereas the windows for the unperturbed optimizations were up to twice as large. It is concluded that robustness considerations must be included in optimization studies in order to obtain realistic results and that elite performances are likely to be robust to timing perturbations of the order of 20 ms.     

The margin for error when releasing the asymmetric bars for dismounts

It has previously been shown that male gymnasts using the "scooped" giant circling technique were able to flatten the path followed by their mass center, resulting in a larger margin for error when releasing the high bar (Hiley and Yeadon, 2003a). The circling technique prior to performing double layout somersault dismounts from the asymmetric bars in women's artistic gymnastics appears to be similar to the "traditional" technique used by some male gymnasts on the high bar. It was speculated that as a result the female gymnasts would have margins for error similar to those of male gymnasts who use the traditional technique. However, it is unclear how the technique of the female gymnasts is affected by the need to avoid the lower bar. A 4-segment planar simulation model of the gymnast and upper bar was used to determine the margins for error when releasing the bar for 9 double layout somersault dismounts at the Sydney 2000 Olympics. The elastic properties of the gymnast and bar were modeled using damped linear springs. Model parameters, primarily the inertia and spring parameters, were optimized to obtain a close match between simulated and actual performances in terms of rotation angle (1.2 degrees), bar displacement (0.011 m), and release velocities (<1%). Each matching simulation was used to determine the time window around the actual point of release for which the model had appropriate release parameters to complete the dismount successfully. The margins for error of the 9 female gymnasts (release window 43-102 ms) were comparable to those of the 3 male gymnasts using the traditional technique (release window 79-84 ms).     

Maximising forward somersault rotation in springboard diving

Performance in the flight phase of springboard diving is limited by the amounts of linear and angular momentum generated during the takeoff phase. A planar 8-segment torque-driven simulation model combined with a springboard model was used to investigate optimum takeoff technique for maximising rotation in forward dives from the one metre springboard. Optimisations were run by varying the torque activation parameters to maximise forward rotation potential (angular momentum × flight time) while allowing for movement constraints, anatomical constraints, and execution variability. With a constraint to ensure realistic board clearance and anatomical constraints to prevent joint hyperextension, the optimised simulation produced 24% more rotation potential than a simulation matching a 2½ somersault piked dive. When 2 ms perturbations to the torque onset timings were included for the ankle, knee and hip torques within the optimisation process, the model was only able to produce 87% of the rotation potential achieved in the matching simulation. This implies that a pre-planned technique cannot produce a sufficiently good takeoff and that adjustments must be made during takeoff. When the initial onset timings of the torque generators were unperturbed and 10 ms perturbations were introduced into the torque onset timings in the board recoil phase, the optimisation produced 8% more rotation potential than the matching simulation. The optimised simulation had more hip flexion and less shoulder extension at takeoff than the matching simulation. This study illustrates the difficulty of including movement variability within performance optimisation when the movement duration is sufficiently long to allow feedback corrections.     

Optimised performance of the backward longswing on rings

Many elite gymnasts perform the straight arm backward longswing on rings in competition. Since points are deducted if gymnasts possess motion on completion of the movement, the ability to successfully perform the longswing to a stationary final handstand is of great importance. Sprigings et al. (1998) found that for a longswing initiated from a still handstand the optimum performance of an inelastic planar simulation model resulted in a residual swing of more than 3 degrees in the final handstand.For the present study, a three-dimensional simulation model of a gymnast swinging on rings, incorporating lateral arm movements used by gymnasts and mandatory apparatus elasticity, was used to investigate the possibility of performing a backward longswing initiated and completed in handstands with minimal swing. Root mean square differences between the actual and simulated performances for the orientations of the gymnast and rings cables, the combined cable tension and the extension of the gymnast were 3.2 degrees, 1.0 degrees, 270N and 0.05m respectively.The optimised simulated performance initiated from a handstand with 2.1 degrees of swing and using realistic changes to the gymnast's technique resulted in 0.6 degrees of residual swing in the final handstand. The sensitivity of the backward longswing to perturbations in the technique used for the optimised performance was determined. For a final handstand with minimal residual swing (2 degrees) the changes in body configuration must be timed to within 15 ms while a delay of 30 ms will result in considerable residual swing (7 degrees).     

Influence of optimization constraints in uneven parallel bar dismount swing simulations

Forward dynamics simulations of a dismount preparation swing on the uneven parallel bars were optimized to investigate the sensitivity of dismount revolution potential to the maximum bar force before slipping, and to low-bar avoidance. All optimization constraints were classified as 1-anatomical/physiological; limiting maximum hand force on the high bar before slipping, joint ranges of motion and maximum torques, muscle activation/deactivation timing and 2-geometric; avoiding low-bar contact, and requiring minimum landing distance. The gymnast model included torso/head, arm and two leg segments connected by a planar rotating, compliant shoulder and frictionless ball-and-socket hip joints. Maximum shoulder and hip torques were measured as functions of joint angle and angular velocity. Motions were driven by scaling maximum torques by a joint torque activation function of time which approximated the average activation of all muscles crossing the joint causing extension/flexion, or adduction/abduction. Ten joint torque activation values, and bar release times were optimized to maximize dismount revolutions using the downhill simplex method. Low-bar avoidance and maximum bar-force constraints are necessary because they reduce dismount revolution potential. Compared with the no low-bar performance, optimally avoiding the low bar by piking and straddling (abducting) the hips reduces dismount revolutions by 1.8%. Using previously reported experimentally measured peak uneven bar-force values of 3.6 and 4.0 body weight (BW) as optimization constraints, 1.40 and 1.55 revolutions with the body extended and arms overhead were possible, respectively. The bar-force constraint is not active if larger than 6.9 BW, and instead performances are limited only by maximum shoulder and hip torques. Bar forces accelerate the mass center (CM) when performing muscular work to flex/extend the joints, and increase gymnast mechanical energy. Therefore, the bar-force constraint inherently limits performance by limiting the ability to do work and reducing system energy at bar release.     

Maximising somersault rotation in tumbling

Performing complex somersaulting skills during the flight phase of tumbling requires the generation of linear and angular momenta during the approach and takeoff phases. This paper investigates how approach characteristics and takeoff technique affect performance with a view to maximising somersault rotation in tumbling. A five-segment planar simulation model, customised to an elite gymnast, was used to produce a simulation which closely matched a recorded performance of a double layout somersault by the elite gymnast. Three optimisations were carried out to maximise somersault rotation with different sets of initial conditions. Using the same initial linear and angular momentum as the double layout somersault and varying the joint torque activation timings allowed a double straight somersault to be performed with 19% more rotation potential than the actual performance. Increasing the approach velocity to a realistic maximum of 7 ms(-1) resulted in a 42% reduction in rotation potential when the activation timings were unchanged but allowed a triple layout somersault to be performed with an increase of 31% in rotation potential when activation timings were re-optimised. Increasing also the initial angular momentum to a realistic maximum resulted in a 4% reduction in rotation potential when the activation timings were unchanged but allowed a triple straight somersault to be performed with a further increase of 9% in rotation potential when activation timings were re-optimised. It is concluded that the limiting factor to maximising somersault rotation is the ability to generate high linear and angular velocities during the approach phase coupled with the ability to adopt consonant activation timings during the takeoff phase.     

Evaluation of a subject-specific female gymnast model and simulation of an uneven parallel bar swing

A gymnast model and forward dynamics simulation of a dismount preparation swing on the uneven parallel bars were evaluated by comparing experimental and predicted joint positions throughout the maneuver. The bar model was a linearly elastic spring with a frictional bar/hand interface, and the gymnast model consisted of torso/head, arm and two leg segments. The hips were frictionless balls and sockets, and shoulder movement was planar with passive compliant structures approximated by a parallel spring and damper. Subject-specific body segment moments of inertia, and shoulder compliance were estimated. Muscles crossing the shoulder and hip were represented as torque generators, and experiments quantified maximum instantaneous torques as functions of joint angle and angular velocity. Maximum torques were scaled by joint torque activations as functions of time to produce realistic motions. The downhill simplex method optimized activations and simulation initial conditions to minimize the difference between experimental and predicted bar-center, shoulder, hip, and ankle positions. Comparing experimental and simulated performances allowed evaluation of bar, shoulder compliance, joint torque, and gymnast models. Errors in all except the gymnast model are random, zero mean, and uncorrelated, verifying that all essential system features are represented. Although the swing simulation using the gymnast model matched experimental joint positions with a 2.15cm root-mean-squared error, errors are correlated. Correlated errors indicate that the gymnast model is not complex enough to exactly reproduce the experimental motion. Possible model improvements including a nonlinear shoulder model with active translational control and a two-segment torso would not have been identified if the objective function did not evaluate the entire system configuration throughout the motion. The model and parameters presented in this study can be effectively used to understand and improve an uneven parallel bar swing, although in the future there may be circumstances where a more complex model is needed.     

Coping with perturbations to a layout somersault in tumbling

Tumbling is a dynamic movement requiring control of the linear and angular momenta generated during the approach and takeoff phases. Both of these phases are subject to some variability even when the gymnast is trying to perform a given movement repeatedly. This paper used a simulation model of tumbling takeoff to establish how well gymnasts can cope with perturbations of the approach and takeoff phases. A five segment planar simulation model with torque generators at each joint was developed to simulate tumbling takeoffs. The model was customised to an elite gymnast by determining subject specific inertia and torque parameters and a simulation was produced which closely matched a performance of a layout somersault by the gymnast. The performance of a layout somersault was found to be sensitive to the approach characteristics and the activation timings but relatively insensitive to the elasticity of the track and maximum muscle strength. Appropriate variation of the activation timings used during the takeoff phase was capable of coping with moderate perturbations of the approach characteristics. A model of aerial movement established that variation of body configuration in the flight phase was capable of adjusting for takeoff perturbations that would lead to rotation errors of up to 8%. Providing the errors in perceiving approach characteristics are less than 5% or 5 degrees and the errors in timing activations are less than 7ms, perturbations in the approach can be accommodated using adjustments during takeoff and flight.     

Effect of hip flexibility on optimal stalder performances on high bar

In the optimisation of sports movements using computer simulation models, the joint actuators must be constrained in order to obtain realistic results. In models of a gymnast, the main constraint used in previous studies was maximum voluntary active joint torque. In the stalder, gymnasts reach their maximal hip flexion under the bar. The purpose of this study was to introduce a model of passive torque to assess the effect of the gymnast's flexibility on the technique of the straddled stalder. A three-dimensional kinematics driven simulation model was developed. The kinematics of the shoulder flexion, hip flexion and hip abduction were optimised to minimise torques for four hip flexion flexibilities: 100°, 110°, 120° and 130°. With decreased flexibility, the piked posture period is shorter and occurs later. Moreover the peaks of shoulder and hip torques increase. Gymnasts with low hip flexibility need to be stronger to achieve a stalder; hip flexibility should be considered by coaches before teaching this skill.     

Mechanical energetic processes during the giant swing exercise before dismounts and flight elements on the high bar and the uneven parallel bars.

The purposes of this study were as follows: (1) To study the energy exchange between the body of the gymnast and the high bar and uneven parallel bars during forward and backward giant swings. (2) To examine the differences between the mechanical energy produced and the mechanical energy absorbed by the muscles during forward and backward giant swings on the high bar and the uneven parallel bars. The data were gathered during the gymnastic world championships in 1994. The experimental set up consisted of two video cameras (50 Hz) and two force measurement bars (500 Hz). A total of 101 giant swings before dismounts and flight elements performed by 33 male and 34 female gymnasts were analyzed. There are characteristically two main phases during forward and backward giant swings before dismounts and flight elements. During the first phase energy is transferred from the gymnast's body into the bar. During this phase of the backward giant swing the energy of the system decreases because the amount of energy decrease of the gymnast's body is more than the energy transferred into the high bar. An exception can be seen during the giant swings in which the gymnast used the power technique. During forward giant swings the energy of the system increases during the first phase. This occurs through active flexion of the hipjoint which produced the extra muscular energy. During the second phase energy is transferred from the bar back into the gymnast's body whose total energy increases. An increase in the energy of the system can only be achieved through muscular work. During the second phase of the backward giant swing the energy of the system increases. The forward giant swings performed on the uneven parallel bars showed a large energy loss during this phase. The energy deficit seen during the first phase of the backward giant swing can be improved by using the power technique. To achieve this the athlete must be in a bent position at the start of the giant swing exercise. Through extension at the shoulder and hip joints muscular energy can be put into the system.     

The influence of touchdown conditions and contact phase technique on post-flight height in the straight handspring somersault vault

In vaulting the gymnast must generate sufficient linear and angular momentum during the approach and table contact in order to complete the rotational requirements in the post-flight phase. This study investigated the effects of touchdown conditions and contact technique on peak post-flight height of a straight handspring somersault vault. A planar seven-segment torque-driven computer simulation model of the contact phase in vaulting was evaluated by varying joint torque activation time histories to match three performances of a straight handspring somersault vault by an elite gymnast. The closest matching simulation was used as a starting point to optimise peak post-flight height of the mass centre for a straight handspring somersault. It was found that optimising either the touchdown conditions or the contact technique increased post-flight height by 0.1 m whereas optimising both together increased post-flight height by 0.4 m above that of a simulation matching the recorded performance. Thus touchdown technique and contact technique make similar contributions to post-flight height in the straight handspring somersault vault. Increasing touchdown velocity and angular momentum lead to additional post-flight height although there was a critical value of vertical touchdown velocity beyond which post-flight height decreased.     

A mathematical high bar–human body model for analysing and interpreting mechanical-energetic processes on the high bar

The aims of this study were:

1. To study the transfer of energy between the high bar and the gymnast.

2. To develop criteria from the utilisation of high bar elasticity and the utilisation of muscle capacity to assess the effectiveness of a movement solution.

3. To study the influence of varying segment movement upon release parameters.

For these purposes a model of the human body attached to the high bar (high bar–human body model) was developed. The human body was modelled using a 15-segment body system. The joint-beam element method (superelement) was employed for modelling the high bar. A superelement consists of four rigid segments connected by joints (two Cardan joints and one rotational–translational joint) and springs (seven rotation springs and one tension–compression spring). The high bar was modelled using three superelements. The input data required for the high bar–human body model were collected with video-kinematographic (50 Hz) and dynamometric (500 Hz) techniques. Masses and moments of inertia of the 15 segments were calculated using the data from the Zatsiorsky et al. (1984) model. There are two major phases characteristic of the giant swing prior to dismounts from the high bar. In the first phase the gymnast attempts to supply energy to the high bar–human body system through muscle activity and to store this energy in the high bar. The difference between the energy transferred to the high bar and the reduction in the total energy of the body could be adopted as a criterion for the utilisation of high bar elasticity. The energy previously transferred into the high bar is returned to the body during the second phase. An advantageous increase in total body energy at the end of the exercise could only be obtained through muscle energy supply. An index characterising the utilisation of muscle capacity was developed out of the difference between the increase in total body energy and the energy returned from the high bar. A delayed and initially slow but even reduction of hip and shoulder angles provided more advantageous release conditions. The total body energy could be improved by up to 15%, the vertical CM release velocity by up to 10% and the angular momentum by up to 35%.     

Timing accuracy in human throwing

This study examines the precision required in the timing of muscle activations and projectile release to hit a target of 20 cm in diameter oriented horizontally either 6 or 8 m away. Over-arm throws, constrained to the sagittal plane, were simulated using a muscle-actuated, two-segment model representing the forearm and hand plus projectile. The parameters defining the modeled muscles and the anthropometry were specific to two male subjects. An objective function specified that throws must be both fast and accurate. Once an optimal solution had been found, the sensitivity of these timings was investigated. The times of activation or release were changed and the simulation model re-run with the new timings, and it was determined whether the projectile would still have struck the target. For one set of simulations, to hit the target at 8 m, the optimal throw was achieved with a time delay between the onset of wrist activation and elbow extensor activation [Proximal-distal (PD) delay] of 49 ms and a release time of 83.4 ms. At this optimal point in the solution space, the launch window was 1.2 ms (assuming the original PD delay). The launch window was the time available within which the projectile must be released and still strike the target. The window during which the wrist flexors could be activated was 10. 41 ms (assuming the projectile was released at the pre-planned optimal time). The control scheme which required the least timing precision had a PD delay of 56 ms and a release time of 89.4 ms. Errors in timing could occur in activation and release simultaneously under this scheme, the timing windows were 4 ms in PD delay and 2.4 ms in release. Similar results were found for a second set of simulations. These simulations revealed the precise timings required in muscle activations and release required for fast accurate throws.     

Reducing ground reaction forces in gymnastics’ landings may increase internal loading

The aim of this study was to use a subject-specific seven-link wobbling mass model of a gymnast, and a multi-layer model of a landing mat, to determine landing strategies that minimise ground reaction forces (GRF) and internal forces. Subject-specific strength parameters were determined that defined the maximum voluntary torque/angle/angular velocity relationship at each joint. These relationships were used to produce subject-specific ‘lumped’ linear muscle models for each joint. Muscle activation histories were optimised using a Simplex algorithm to minimise GRF or bone bending moments for forward and backward rotating vault landings. Optimising the landing strategy to minimise each of the GRF reduced the peak vertical and horizontal GRF by 9% for the backward rotating vault and by 8% and 48% for the forward rotating vault, compared to a matching simulation. However, most internal loading measures (bone bending moments, joint reaction forces and muscle forces) increased compared to the matching simulation. Optimising the landing strategy to minimise the peak bone bending moments resulted in reduced internal loading measures, and in most cases reduced GRF. Bone bending moments were reduced by 27% during the forward rotating vault and by 2% during the backward rotating vault landings when compared to the matching simulations. It is possible for a gymnast to modify their landing strategy in order to minimise internal forces and lower GRF. However, using a reduction in GRF, due to a change in landing strategy, as a basis for a reduction in injury potential in vaulting movements may not be appropriate since internal loading can increase.     

The effect of cost function on optimum technique of the undersomersault on parallel bars

The undersomersault, or felge, to handstand on parallel bars has become an important skill in Men's Artistic Gymnastics as it forms the basis of many complex variations. To receive no deductions from the judges, the undersomersault must be performed without demonstrating the use of strength to achieve the final handstand position. Two male gymnasts each performed nine undersomersaults from handstand to handstand while data were recorded using an automatic motion capture system. The highest and lowest scoring trials of each gymnast, as determined by four international judges, were chosen for further analysis. Three optimization criteria were used to generate undersomersault technique during the swing phase of the skill using a computer simulation model: minimization of peak joint torques, minimization of horizontal velocity before release, and maximization of angular momentum. The techniques used by both gymnasts could be explained using the second optimization criterion which facilitated further skill development. The first optimization criterion generated a technique advocated for beginners where strength might be expected to be a limiting factor. The third optimization criterion resulted in a different type of undersomersault movement of greater difficulty according to the FIG Code of Points.     

Kinematics estimation of straddled movements on high bar from a limited number of skin markers using a chain model

To reduce the effects of skin movement artefacts and apparent joint dislocations in the kinematics of whole body movement derived from marker locations, global optimisation procedures with a chain model have been developed. These procedures can also be used to reduce the number of markers when self-occlusions are hard to avoid. This paper assesses the kinematics precision of three marker sets: 16, 11 and 7 markers, for movements on high bar with straddled piked posture. A three-dimensional person-specific chain model was defined with 9 parameters and 12 degrees of freedom and an iterative procedure optimised the gymnast posture for each frame of the three marker sets. The time histories of joint angles obtained from the reduced marker sets were compared with those from the 16 marker set by means of a root mean square difference measure. Occlusions of medial markers fixed on the lower limb occurred when the legs were together and the pelvis markers disappeared primarily during the piked posture. Despite these occlusions, reconstruction was possible with 16, 11 and 7 markers. The time histories of joint angles were similar; the main differences were for the thigh mediolateral rotation and the knee flexion because the knee was close to full extension. When five markers were removed, the average angles difference was about 3 degrees . This difference increased to 9 degrees for the seven marker set. It is concluded that kinematics of sports movement can be reconstructed using a chain model and a global optimisation procedure for a reduced number of markers.     

The influence of simulation model complexity on the estimation of internal loading in gymnastics landings

Evaluating landing technique using a computer simulation model of a gymnast and landing mat could be a useful tool when attempting to assess injury risk. The aims of this study were: (1) to investigate whether a subject-specific torque-driven or a subject-specific muscle-driven model of a gymnast is better at matching experimental ground reaction forces and kinematics during gymnastics landings, (2) to calculate their respective simulation run times and (3) to determine what level of model complexity is required to assess injury risk. A subject-specific planar seven-link wobbling mass model of a gymnast and a multi-layer model of a landing mat were developed for this study. Subject-specific strength parameters were determined which defined the maximum voluntary torque/angle/angular velocity relationship about each joint. This relationship was also used to produce subject-specific 'lumped' muscle models for each joint. Kinetic and kinematic data were obtained during landings from backward and forward rotating gymnastics vaults. Both torque-driven and muscle-driven models were capable of producing simulated landings that matched the actual performances (with overall percentage differences between 10.1% and 18.2%). The torque-driven model underestimated the internal loading on joints and bones, resulting in joint reaction forces that were less than 50% of those calculated using the muscle-driven model. Simulation time increased from approximately 3 min (torque driven) to more than 10 min (muscle driven) as model complexity increased. The selection of a simulation model for assessing injury risk must consider the need for determining realistic internal forces as the priority despite increases in simulation run time.     

An approach for developing an experimentally based model for simulating flight-phase dynamics

 This paper presents an approach for developing an experimentally validated dynamic multisegment model to simulate human flight-phase dynamics and multijoint control. Modeling and experimental techniques were integrated to systematically examine the contribution of multiple error sources to the accuracy of the model and to determine the complexity of a model that adequately emulates the dynamic behavior at the total-body and multijoint levels during flight. The accuracy of the model and of the experimental data was assessed using an inverse dynamics simulation of flight-phase motion for two representative cases: (i) a physical model released from a bar and (ii) a gymnast performing a layout dismount from a bar. Multijoint models with varying numbers of segments were assessed in order to determine the complexity of the model that adequately simulates the flight-phase task. A five-segment model was found to adequately simulate the layout dismount performed by the gymnast. The error introduced during modeling and digitizing contributed to an apparent violation of the conservation law manifested as large external forces acting on the nonactuated joints. These results demonstrate the need to reduce sources of error prior to testing hypotheses regarding feedforward and feedback components of the multijoint control system. The proposed approach for quantifying sources of error provides a crucial step that is required in the development of experimentally based dynamic models designed to examine and test hypotheses regarding multijoint control logic. Received: 5 April 2001 / Accepted in revised form: 25 April 2002 Acknowledgements. This work was funded by Intel, a dissertation awarded from the International Society of Biomechanics, the Internationale Federation de Gymnastique, the International Olympic Committee, and Pfizer. We express our special thanks to Kathleen Costa and Witaya Mathiyakom for their assistance in data collection. Correspondence to: P. S. Requejo (e-mail: requejo@rcf.usc.edu)