When Brian Simmons showed a room of transportation safety professionals a DriveCam® video clip of a driver crashing into a vehicle that had run a stop sign and was travelling through the junction, many in the audience shook their heads and clucked in disapproval. The driver should have braked sooner. The driver should have braked harder. The driver should have seen it coming. The driver should have reacted more quickly.
That disapproval, however, was precisely the wrong response. Leaning on the laws of physics and data from the DriveCam clip, Simmons, a safety director for Waste Management Inc., showed through a series of calculations that there was, in fact, little the driver could have done to avoid the collision.
It was an eye-opening exercise that Simmons has gone through thousands of times as a law enforcement officer who has investigated numerous accidents and been an expert witness in more than 20,000 cases involving vehicle collisions.
Simmons shared his techniques during a talk he gave this spring as part of the Annual Lytx User Group Conference in San Diego. Here’s a recap of the advice he gave during this informative presentation on effectively using DriveCam video analytics:
Don’t give in to hindsight bias. When the video lands in our inbox, most of us already know the outcome. This knowledge biases our judgment, leading us to jump to conclusions about what happened. Even when we do watch the video for the first time, our surface level analysis of video will often lead to the wrong answer.
Be objective. This is tough, Simmons acknowledged, especially if you or your company has a stake in the conclusion of your analysis. Two common errors tend to creep in at this stage: the tendency to automatically blame the driver, and the tendency to automatically side with the driver because of a personal connection. Both are likely to interfere with your ability to find out what really happened. This article talks about the role of emotion in effective coaching.
Don’t rely on the driver for the full picture. In Simmons’ experience, the least reliable sources of information are the drivers involved in the incident. “Most drivers are dealing with self-preservation” or they don’t always recall events fully or objectively, Simmons said. Those who have DriveCam will have a tremendous advantage in that they receive an accurate snapshot of the hard physical data, such as speed, brake timing, brake pressure, and so on, Simmons said.
Perfection is the enemy of good. Safety analysts sometimes spend hours on a 12-second clip trying to figure out what the driver did incorrectly, poring through data and matching that up with best practices or company policy. “When you measure against perfection, your employees will never measure up,” Simmons said. “It’s not about what was ideal, or what your best driver would do. It’s about whether your driver’s reactions were reasonable.”
Never think about the liability. If you do, you’ll be looking for exculpatory evidence, which should not be the point of the exercise, Simmons argued. “Everything you do should be done on the merits of the physical evidence,” he said. Again, the goal of incident analysis is to find out what happened and whether the driver’s actions were reasonable.
In discussing incident analysis using physical analysis, Simmons walked through a few basic calculations. Simmons said he often starts with converting known speeds (miles per hour) to velocity (feet per second) using this equation:
Velocity = Speed x 1.466
Next, he determines whether the event was a simple event or a complex event. In simple events, people rely on conditioned behaviors. An example of a simple event is moving your foot to the brakes once you detect active brake lights on the vehicle you are following. Most collision events, however involve complex responses to unanticipated events. Reacting to complex events typically takes more than three times as long as simple events. Here’s rule of thumb:
Simple reactions = 0.5 seconds
Complex reactions = 1.6 seconds
Multiply the velocity (feet per second) by the appropriate reaction time to figure out how far the vehicle will travel before the driver can process the situation and prepare a response. The delay between the time a situation presents itself and when the driver reacts can increase by as much as two seconds if the driver happened to be doing a routine sweep of his or her mirrors.
Simmons cautioned that collision events are often complex, involving numerous factors. Covering all the factors, and the physics that play into those factors, was beyond the scope of the talk. He summed up his talk with two pieces of advice: rely on the physical sciences for incident analysis, and be as objective as possible when seeking the truth.