How YouTube is Advancing the Science of Accident Reconstruction


YouTube is a treasure trove of real-world crash data that is advancing the science of accident reconstruction. Take the videos posted by Ken Snyder (aka RNickeymouse), for instance. On the weekends, you may find Snyder among those perched at Edwards Corner with a camera. This 180-degree curve is the final curve on a section of the Mulholland Highway motorcyclists often call the Snake (the photograph below shows this curve). The Mulholland Highway winds its way through the Santa Monica Mountains for about 32 miles, beginning at its intersection with Mulholland Drive in Calabasas and continues to the Pacific Coast Highway just west of Malibu. Each weekend, motorcyclists and spectators flock to Edwards Corner where there is plenty of space to park and plenty of action to keep you mesmerized.


Snyder captures video of riders crashing on this curve and then posts the footage to his YouTube channel. To date, he has posted 105 crashes, including the one below. Snyder’s channel boasts nearly three hundred thousand subscribers and tens of millions of views. Apart from the income this surely generates for Snyder, his videos are the kind of real-world data that accident reconstructionists can and should eat up. In the past, if accident reconstructionists wanted to study crashes, they had two options. They could use the physical evidence to piece together crashes that had already occurred or they could stage a crash test. Accident reconstructionists have dreamed up all kinds of crash test scenarios, including tests with motorcycles. They have pushed motorcycles into carsand thrust them into guardrails. They’ve shoved them out of the back of moving trucks and drug coverall-clad engineers across the ground, all in the name of being able to tell their clients how a particular crash occurred.


There are limits, though, to the scenarios you can stage when a crash test dummy is driving the motorcycle. Try getting a dummy to lean the right direction in a curve. For crashes like the ones that Snyder captures, reconstructionists were left to infer the cause after the fact, examining the tire marks and scrapes on the roadway and the damage to the motorcycle or to the rider’s clothing and helmet. With the advent of YouTube and the insatiable craving we humans have for dramatic footage, this situation has changed. For the accident reconstructionist who has the ambition to pick through Snyder’s videos, there is much to be learned that simply is not accessible through crash tests or through the physical evidence that is often left behind after a crash. These videos yield unique data because they give accident reconstructionists access to parts of a crash sequence that leave no physical evidence – the rider actions that precede the crash, for instance, or the motion of the rider as he is thrown through the air.

To see this, consider one of the Snyder’s crash videos below – a high-side fall involving a rider operating a Honda CBR1000RR traveling through Edwards Corner towards Malibu. The rider in this crash lost control of his motorcycle when his rear wheel began sliding to the outside of the curve, depositing a tire mark on the roadway in the process. From the audio, it was evident that, prior to this loss of control, the rider was applying throttle to increase his speed – a detail that likely would have been inaccessible via the physical evidence. The motorcycle and the rider then began leaning to the left and the rider responded by steering the front wheel to the right and decreasing the throttle input. The throttle release led to an increase in the available traction which, in turn, led to the motorcycle righting itself. The rider reacted by steering the front tire back to the left. Initially, this steering was ineffective and the motorcycle began falling towards its left side. As this occurred, the rider’s butt left the seat.

Eventually, the steer input caused the rear wheel of the motorcycle to kick out to the right. The rear tire of the motorcycle deposited another tire mark on the roadway and the tire forces threw the motorcycle into a rightward lean. The motorcyclist’s groin and left thigh impacted the fuel tank. Eventually, the motorcycle tires lost contact with the ground and the rider was projected upward and out ahead of the motorcycle. The motorcycle landed on its right side, and then, the rider landed in a seated posture facing opposite his initial direction of travel, having been launched 54 feet in the air. The velocity of the rider, the distance he was thrown, and his trajectory and motion would likely have been inaccessible without the video. The motorcycle slid to rest on its right side. As the rider traveled to rest, his torso and head rotated toward the ground and slammed onto the asphalt. He then tumbled to rest. While this sequence deposited a couple of tire marks and scrapes on the roadway and surely damaged the right side of the motorcycle and the rider’s clothing, much of the subtlety of the motion – of both the bike and the rider – would be lost to history without this video.

Beyond qualitative analysis of the motion that the video enables, accident reconstructionist also need to put hard numbers to a sequence like this. How fast was the motorcyclist traveling when the rider lost control? At what rate did the motorcycle decelerate during the loss of control? At what rate did the motorcycle and rider decelerate once they landed on the ground? How far was the motorcyclist thrown? This is the science of accident reconstruction and it’s the kind of analysis that requires us to determine the times and distances over which events in the video occurred. The timing comes straight out of the videos – for the video above, Ken Snyder provided his raw footage which was shot at 60 frames per second. To get the distances for quantitative analysis, a visit to Edwards Corner allowed us to document features of the roadway and the surrounding area that could act as landmarks for analysis of the video.

Mapping the distances between landmarks was accomplished using Faro laser scanner, with which we measured millions of points at the scene to define the geometry of the curve. This device also photographs the site and maps the colors of these photographs onto the measurement points. The result is a realistic-looking and accurate cloud of points that allowed us to generate a computer model of Edward’s Corner, from which we could take measurements. The image below shows the data we collected with the Faro laser scanner.


Using distances measured from this mapping of Edwards Corner, we determined that the motorcycle in the video was traveling 43 mph as it approached the exit of the curve. Our mapping of the curve also revealed its geometric characteristics, which in turn enabled us to determine what would be required of a rider traversing Edwards Corner at this speed. For instance, a rider traveling through the curve towards Malibu would encounter a left-hand curve with an upslope. During the first half of the curve, the upslope is 3.2 degrees and during the second half 4.9 degrees. There is a total elevation gain of 27.6 feet over a 390.8 foot distance around the curve. At the westbound entry to the curve, the cross-slope is negligible. It increases to a maximum cross-slope of 5.7 degrees (with the inside of the curve being lower than the outside). The cross-slope then begins to decrease again, reaching a value of 1.9 degrees at the exit to the curve. The radius of the curve in the center of the westbound lane is approximately 82 feet. This is a tight curve and a rider traveling through the curve at 40 mph would need to lean to the left about 45 degrees to the left relative to the road [Rose, 2014Carter, 2015]. This amount of lean will test the psychological limits of many riders, even of many experienced riders. Watanabe & Yoshida [1973] found that the maximum lean angles utilized by novice riders were typically in the range of 15 to 25 degrees and those used by experienced riders were in the range of 34 to 40 degrees.

So far, we have analyzed only four of Snyder’s videos and a few rollover crashes. There are still 101 crash videos on Snyder’s channel that we could analyze, and I’m sure there are thousands of others out there on YouTube. Just as Snyder’s videos reveal features of these crashes that would not be accessible without the video, other crash videos on YouTube will reveal aspects of crashes that precede the physical evidence – the timing and manner of driver responses for instance. We have only scratched the surface with the videos we have analyzed. What accident reconstructionists can learn from videos of real-world crashes is only limited by the determination that we can muster to mine, vet, and analyze this data.


  1. Rose, N., Carter, N., Pentecost, D., “Analysis of Motorcycle and Rider Limits on a Curve,” Collision: The International Compendium for Crash Research, Volume 9, Issue 1, Spring 2014.
  2. Carter, N., Rose, N., and Pentecost, D., “Validation of Equations for Motorcycle and Rider Lean on a Curve,” SAE Int. J. Trans. Safety 3(2):126-135, 2015, doi:10.4271/2015-01-1422.
  3. Watanabe & Yoshida, “Motorcycle Handling and Performance for Obstacle Avoidance,” International Congress on Auto Safety, 1973.