Sliding and Tumbling Deceleration of a Motorcycle (Motorcycle Accident Reconstruction)


A quick aside...this is a draft of another section from the motorcycle accident reconstruction book that I'm writing with Lou Peck and William Neale. Lou, in particular, assisted with this section. Let me know what suggestions you have for making this more readable and useful. Thanks!


When performing speed calculations, reconstructionists typically assume that a sliding or tumbling motorcycle decelerates at a constant rate. This assumption is adequate for most accident reconstruction applications, though there will in reality be some variability in the deceleration rate along the slide distance, depending on which motorcycle components are engaging the road surface at any particular point in time. If the motorcycle slides across multiple surfaces, different deceleration rates may need to be assigned for each different surface. In practice, the reconstructionist would determine the slide distance based on the physical evidence, and then a range of deceleration rates would be selected from physical tests reported in the literature for similar type motorcycles sliding on similar surfaces. A number of these studies are reviewed in this articles.

Average Deceleration Rates for a Sliding Motorcycle

Day and Smith pioneered research on the topic of motorcycle sliding deceleration in 1984, analyzing the behavior of two downed motorcycles on various surfaces – a 1967 Honda CB305 and a 1973 Yamaha 550 Special. They towed the motorcycles using a rope (see the figure below) with an in-line force gauge, and documented the forces required to pull the motorcycles at 1 and 40 kph (25 mph). For pavement, Day and Smith found a sliding friction factor range of 0.45 to 0.58 during the 25 mph tests. For gravel, the friction factor was 0.68 to 0.79 and, for grassy earth, 0.79. Day and Smith noted that “during the testing, it was observed that projecting elements, such as the foot pegs and handle bars, would tend to plow into the soil at low speeds, creating momentary high drag factors.” In this testing, the motorcycle started in a capsized position on the ground, and so, deceleration from a fall was not a part of the friction factors reported by Day and Smith.


Lambourn [1991] explored how sliding deceleration rates experienced by motorcycles differed between tests where the vehicle was dragged at low speeds and tests where the motorcycle was dropped at a higher speed and allowed to slide to rest. For the higher speed tests, motorcycles were dropped from a low platform (already on their side) or allowed to fall to their side from an upright position. The photographs below depict the test setup for the upright drops and one of the tests from this setup. In his literature review, Lambourn noted that some studies had reported a speed-dependence on the sliding deceleration rate, with the rate decreasing with increasing speed. He also examined this issue of speed dependence of the deceleration rate in his testing.


Lambourn reported that the deceleration rates (friction factor) measured “in the low-speed drag tests gave a value close to the high-speed sliding value. The friction was affected by the road surface texture, the presence of prominent side projections, and the wearing away of these projections during the slide. Some speed dependence was noted in the upright-launch tests which appears to be due to the ‘digging-in’ of the machine as it falls to the road, rather than an effect of the sliding friction itself.” He also concluded that “the reason for there being a clear speed dependence in the results of Becke and of Ashton, but not in the tests reported here, is almost certainly due to the fact that in both their experimental methods the motorcycles were dropped a distance onto the road surface. This would subject the machines to a large decelerating impulse as they struck the ground, which would considerably increase the average deceleration in low speed tests but be relatively unimportant in high speed runs.” Lambourn concluded that the sliding deceleration rate of a motorcycle was dependent on the roughness characteristics of the road surface.

Other authors have also suggested that differences in test methodology – whether or not the motorcycle drops to the ground in the test and from what height – account for some of the variability in the deceleration rate for sliding motorcycles and the apparent speed dependence in the deceleration rates. For instance, Baxter [2017] stated: “One word of caution; read the test methodology as to how the friction values were obtained. In some drop tests, the motorcycle was pre-positioned laterally a few inches/centimeters above the road surface. Results using this method are generally lower than a motorcycle falling from vertical (normal position) onto its side on the road.” Hague [2004], Wood [2008], and Walsh [2009] also discussed the influence of the fall (capsize) on the deceleration rate. Hague stated: “The methodology of a motorcycle slide test can significantly affect the measured deceleration rate, apparently due to the speed lost in the initial ground impact. Those tests in which motorcycles were dropped from an increased height resulted in increased deceleration rates. During a road traffic accident the motorcycle will also lose speed upon ground impact and testing should therefore try to mimic this process. Bearing this in mind, the most appropriate tests for the majority of collisions would be those in which an upright motorcycle was allowed to capsize from a normal height.” This is different than the approach proposed by Wood and Walsh. They proposed splitting the speed calculations into separate phases for the capsize and the slide. Thus, within their method, the ideal test procedure for the sliding phase would not include a fall of the motorcycle. Walsh presented a method for incorporating the speed loss from a fall into the calculation of the deceleration rate from a test or the calculation of the motorcycle’s initial speed in a reconstruction.

Donohoe [1991] reported sliding deceleration rates for a 1982 Kawasaki KZ1000 Police Special. Testing with this motorcycle, which was conducted at the Los Angeles Police Department’s Specialized Collision Investigation Detail, utilized a flatbed truck with a lift gate on the rear. The lift gate was positioned parallel to the roadway and approximately 6 inches above the road surface. The motorcycle, which was facing along the direction of travel, was dropped an upright position with its front tire on the lift gate and its rear tire on the roadway. The roadway was a residential, asphalt roadway adjacent to Dodger Stadium. The initial speed of the motorcycle was measured with radar. Donohoe reported 5 tests with sliding deceleration rates between 0.38 and 0.50.

In 1995, Raftery slid an unknown motorcycle wearing Suzuki Katana fairings from an initial speed of 85 kph (53 mph) and reported an average deceleration of 0.26g. Another test, seemingly from a similar speed, resulted in the same 0.26g. As a control test, Raftery took the same motorcycle, removed the fairings, and performed another test. The resulting coefficient of friction of was 0.33g. The coefficient of friction was calculated using the initial drop speed and the documented sliding distance. Raftery’s methodology involved suspending the motorcycle from a boom at the rear of a tow truck, driving the tow truck up to the test speed, and releasing the motorcycle from the boom. It appears from Raftery’s description of his tests that he suspended the motorcycle from the boom with the wheels rolling on the ground and when the motorcycle was released it was allowed to fall to the road surface.

Carter [1996] tested 8 different motorcycles (see the figure below) on 3 different surfaces (asphalt, dirt, and gravel) to determine their sliding deceleration rates from target speeds of 48 and 97 kph. All of the tests run on off-road surfaces utlized a target speed of 48 kph. The motorcycles that Carter tested included the following motorcycle types: standard, cruiser, sport, and touring. In all, Carter reported 50 tests. Carter reported that “some speed effects were observed, i.e., for higher speeds, the slide coefficient was lower (likely due to heat softening of structure contact points with the pavement).” Also, “for full fairing equipped motorcycles the slide coefficient was consistently lower than for non-fairing equipped motorcycles” and “deep gouges left by the motorcycle in the surfaces corresponded with higher slide coefficients.”


Carter attempted to improve on prior studies by developing a test rig that allowed for consistent positioning and release of the motorcycles. The motorcycles were positions front wheel forward and on their left or right sides. The motorcycles were released from a position with the lowest point on the side of the motorcycle approximately 5 centimeters above the ground. As Carter noted, “this release height was chosen to minimize the impact forces upon release, therefore restricting (to the extent possible) the tests only to energy dissipated during sliding.”

Medwell [1997] performed four motorcycle sliding tests using a fully-faired 1992 Kawasaki ZX-7 Ninja. In relationship to the test procedure he used, Medwell stated that “the tests were designed to approximate, as closely as possible, the motion of a motorcycle falling over from an upright position. The motorcycle was positioned upright on a fabricated platform mounted on the right side of a pickup truck…The height of the platform was adjusted so that its underside was as close as possible to the roadway surface. This test setup resulted in the motorcycle tire contact surface being approximately 90mm above the roadway. The motorcycle was held upright by an assistant riding in the bed of the pickup truck. The truck was accelerated to the test speed, then the motorcycle was released and allowed to fall over sideways onto the road surface.” In two of the tests, the Kawasaki initially slid along the pavement, but then traveled into a nearby area of grass, making them difficult to analyze. However, two of the tests were confined to the asphalt. Both had a release speed of approximately 80 kph (50 mph). The motorcycles slid for 69.5 and 86.3 meters (228 and 283 feet) before coming to rest. The calculated coefficients of friction were 0.36 and 0.29g. It is worth noting that the 0.36 value was obtained during the test involving the right side of the motorcycle, which is the exhaust side.

Bartlett [2007] reported motorcycle drop tests from Motorcycle Crash Reconstruction classes conducted at the Institute for Police Technology and Management (IPTM) from 1987 to 2006. All of these tests were conducted on asphalt or concrete, but the surfaces varied from class-to-class. The drop techniques also varied from class-to-class. Bartlett observed: “The results are a chaotic mix of sliding and tumbling, not unlike real motorcycle crashes.” Bartlett’s dataset initially consisted of 237 drop tests using 107 different motorcycles. Twenty tests were discarded because the reported drop speed, slide distance, and deceleration rate were inconsistent with each other. Additional tests were excluded in which the motorcycle was dropped from a pickup bed or in which the motorcycle slid off of the road surface onto the off-road terrain. The final dataset included 162 tests with 99 different motorcycles.

Bartlett reported that the deceleration rates trended slightly higher with increasing speed and that the overall average deceleration rate for all of the tests was 0.521 ± 0.140g. Bartlett also combined his dataset with other available datasets [McNally, 2006 and Lambourn, 1991, for instance]. This resulted in 386 tests for which Bartlett reported deceleration rates of 0.480 ± 0.134g.

In 2003, McNally and Bartlett slid a fully-faired Suzuki Katana at IPTM’s Special Problems and analyzed the results via frame-by-frame video and field data (known initial speed and measured slide distance). Video analysis yielded a coefficient of friction of 0.42g while the sliding distance and known initial speed yielded a result of 0.39g. Bartlett has also reported nine additional tests performed using fully-faired motorcycles during IPTM classes over the years [2007]. The individual results were not detailed in the paper, but combined with the data from Raftery, Medwell, and McNally, the total set of 14 tests had an average coefficient of friction of 0.37g with a standard deviation of 0.08g.

In 2004, Hague compiled data from prior studies where motorcycles were allowed to capsize and then slide to rest. Hauge concluded that “the analysis shows that a more accurate estimation of deceleration rate can be made if the motorcycles are split into two different categories, based on the presence of fairing, crash bars and/or panniers.” He noted that, while “one might expect partially faired motorcycles to have lower deceleration rates than unfaired machines…the two categories exhibit similar deceleration rates. Perhaps also initially surprising is that fully faired machines equipped with panniers gave similar results to the partially/unfaired motorcycles.” In relationship to crash bar equipped motorcycles, the only available tests were those conducted by Lambourn [1991], in which the deceleration rates varied between 0.25 and 0.35. Hague noted that, “As expected, fully faired and crash bar equipped motorcycles decelerated at relatively low rates. The crash bar equipped results are probably artificially low because they were all dropped from a very low height. If they had capsized from an upright position, they would have lost additional speed on striking the ground which would increase the average deceleration rate, more so at lower speeds. Although deceleration rates as low as 0.2 have been suggested for crash bar equipped machines there appears to be no published data to support such low values.” Hague reported an average deceleration rate for partially faired and unfaired motorcycles of 0.39 and for fully faired motorcycles of 0.27.

Peck worked with several members of CA2RS (California Association of Accident Reconstruction Specialists) to perform 14 sliding tests using modern GPS data acquisition technology, which captured the motorcycle speed at 10 Hz (ten times per second). In 2014, Peck reported this data, which included two tests with fully-faired motorcycles, a 1989 Suzuki GSX-R750 and a 1991 Suzuki GSX600F. These tests yielded sliding deceleration rates of 0.42 and 0.47g, respectively. 

Missing from the literature cited so far is documentation of the sliding deceleration rate for motorcycles equipped with frame sliders, a common sport bike modification. Frame sliders, usually comprised of a plastic composite, are mounted to the sides of motorcycles to mitigate damages during a fall. Over the course of two years, Peck collected data from track crashes at track days and racing events at New Hampshire Motor Speedway and New Jersey Motorsports Park. All analyzed crashes involved motorcycles equipped with a QSTARZ GPS data acquisition system (5 or 10 Hz) [2014]. In total, data from 15 crashes were collected and analyzed. All 15 crashes involved faired motorcycles equipped with plastic composite frame sliders. The average coefficient of friction was 0.45g with a standard deviation of 0.09g. These numbers are more consistent with data from non-faired motorcycles, indicating that frame sliders actually increase the sliding deceleration rate for sport motorcycles. In addition, these crashes involved the potential interaction between the motorcycle and the operator that could occur in the real-world.


There is speed loss that occurs when a motorcycle capsizes. Thus, when the capsizing phase is included in a test, the deceleration rate will be higher than if that phase is not included. In theory, the deceleration rates from tests including the capsizing phase could be adjusted to eliminate the speed loss from the capsize. Then, when reconstructing a real-world crash, the reconstructionist could calculate the speed loss from the sliding and capsizing phases separately. This is the approach pursued by Wood [2008] and Walsh [2009]. While theoretically appealing, this solution seems impractical to carry out fully due to the voluminous number of tests in the literature and the widely varying, and sometimes not documented, test procedures. A more practical solution is for the reconstructionists to parse the data out by surface type and motorcycle type and then to apply it without adjustment for the capsizing phase. In this approach the analyst would not include a separate speed loss calculation for the capsizing phase and any tests not including this phase would represent conservative values in the spread of deceleration rates.


1.     Bartlett, Wade, et al, “Motorcycle Slide-to-Stop Tests: IPTM Data through 2006,” Accident Investigation Quarterly, Spring 2007.

2.     Baxter, Albert T., Motorcycle Crash Investigation, Institute of Police Technology and Management, 2017.

3.     Carter, T., Enderle, B., Gambardella, C., and Trester, R., “Measurement of Motorcycle Slide Coefficients,” SAE Technical Paper 961017, 1996, doi:10.4271/961017.

4.     Day, T. and Smith, J., “Friction Factors for Motorcycles Sliding on Various Surfaces,” SAE Technical Paper 840250, 1984, doi:10.4271/840250.

5.     Donohue, M.D., “Motorcycle Skidding and Sideways Sliding Tests,” Accident Reconstruction Journal, Vol. 3, No. 4, 1991.

6.     Hague, David, “Calculation of Speed from Motorcycle Slide Marks,” Impact: The Journal of the Institute of Traffic Accident Investigators, Spring 2004.

7.     Lambourn, R., “The Calculation of Motorcycle Speeds from Sliding Distances,” SAE Technical Paper 910125, 1991, doi:10.4271/910125.

8.     McNally, B., Bartlett, W., “Motorcycle Sliding Coefficient of Friction Tests,” Presentation at IPTM Special Problems in Accident Reconstruction, 2003.

9.     Medwell, C., McCarthy, J., and Shanahan, M., "Motorcycle Slide to Stop Tests," SAE Technical Paper 970963, 1997, doi:10.4271/970963.

10.  Peck, L., Focha, W., Gloekler, T., “Motorcycle Sliding Friction for Accident Investigation,” Proceedings of the 10th International Motorcycle Conference, Institute for Motorcycle Safety, Essen, Germany, pp. 62-67, 2014.

11.  Raftery, B., “Determination of the Drag Factor of a Fairing Equipped Motorcycle,” SAE Technical Paper 950197, 1995, doi:10.4271/950197.

12.  Searle, John A., “The Trajectories of Pedestrians, Motorcycles, Motorcyclists, etc., Following a Road Accident,” 831622, Society of Automotive Engineers, 1983.

13.  Walsh, D.G., Wood, D.P., Alliot, R., Glynn, C., Simms, C.K., “Motorcycle Capsize Mechanisms and Confidence Limits for Motorcycle Capsize Speeds from Slide/Bounce Distance,” 18th EVU Conference, Hinckley, UK, 2009.

14. Wood, D.P., Alliot, R., Glynn, C., Simms, C.K., Walsh, D.G., “Confidence Limits for Motorcycle Speed from Slide Distance,” Proc. IMechE Vol. 222, Part D: J. Automobile Engineering, 2008.

About Nathan

Nathan is an accident reconstruction expert at Kineticorp. He is dedicated to mastering his craft, and for the past 20 years, he has dedicated himself to research and writing as a means of developing authentic expertise that provides real value to juries. Nathan developed this article as part of the research that will ultimately make it into his book on motorcycle accident reconstruction.