Determining liability in an auto collision has often been viewedas more of an art than a science. Therefore, it shouldn't be toosurprising that a record number of disputes are being sent toArbitration Forums for resolution. If you empower the claimsadjuster with the flexibility to define the standard to which theycompare a driver's performance, then it is not particularlyshocking that the resulting analysis supports their initial take onliability. Because no standard is more valid than the next, you areleft with a stalemate of sorts, regardless of how hard youargue.

Although there may be some doubt about an individual liabilityassessment, there is no doubt about its impact on the bottom line.Not only does it affect the amount an insurer pays on the claim,but it also keeps the company from properly assessing the risk andcollecting the appropriate premium. Auto collisions areparticularly well suited for scientific analysis. After all, thevehicles involved are governed by the laws of physics described byNewton many centuries ago. Advances in technology not only make ascientific approach feasible, but cost-effective as well.

A collision occurs because two vehicles attempt to occupy thesame space at the same time. As long as both cars reach impact atthe same time, we have a common point on each vehicle's timeline.This allows us to work backwards and calculate where they were atany point prior to impact, assuming that we know what each wasdoing. Our objective is to evaluate what the driver either couldhave—or should have—done in response to a perceived hazard. Thisrequires taking certain factors into consideration, includingacceleration rates, initial speeds, distances, perception/responsetimes, and so on. With these variables, we can calculate the totalstopping distance of the approaching vehicle. If it is at leastthis distance from impact, then the collision is avoidable. Theadjuster will be able to collect some of these items fromstatements and others from police reports. During the course of theaccident investigation, published research will also provide areference and assist in making the most accurate determinationpossible.

The Monte Carlo Analysis

One may wonder about the reliability of these values. That iswhere the Monte Carlo analysis comes in.

The Monte Carlo analysis is a technique by which the collisionis evaluated multiple times using a mathematical model. Each inputvariable is replaced with a uniform probability distribution with areasonable range of values identified by the investigator or anormal probability distribution from research with a defined meanand standard deviation.

For instance, the approaching vehicle might be operatingsomewhere between 35 and 45 miles per hour. A random value withinthis range is selected for the series of calculations thatdetermines whether or not the collision occurs. This process isrepeated, for all of the variables, 15,000 to 100,000 times untilthe model accounts for almost every possible combination of values.The uncertainty of these values is addressed by the fact that weuse every reasonable value in our calculations. Count how manytimes the collision occurs out of the sample population and itprovides us a percentage representing the probability of acollision.

Because the perception/response times are calculated fromcompiled research, you are actually comparing this driver'sperformance to the remainder of the population. Not only do youknow how the average driver would have responded, but you also knowhow any and all drivers would have responded.

Percentiles and Liability

How is this valuable? Let's go back and look at a pathintrusion. If the Monte Carlo model suggests that 95 percent of thedrivers, when presented with this hazard, would have been involvedin the collision, then it is very difficult to argue that thedriver of the approaching vehicle should be penalized because he orshe was not in the top 5 percent of all drivers. Conversely, if themodel suggests that 80 percent of drivers would have successfullyavoided a collision, you can effectively demonstrate that theapproaching driver's performance was well below average and asignificant contributing factor, even though they were proceedingwith the statutory right of way.

Now take it a step further. How many times do you have anadmission that a vehicle was traveling marginally above the postedlimit? Sure it is an ordinance violation but did it make adifference? The model allows you to examine the impact ofindividual variables. For instance, the approaching driver admitsto traveling 45 mph in a 35-mph zone. Run the model at 35 miles perhour and we may find that the probability of a collision was 45percent. Now run the model at 45 miles per hour, and it increasesto 55 percent. Even though it is only a difference of 10 miles perhour, we can say that the unlawful speed increased the probabilityof a collision by 10 percent and made it more likely than not (>51 percent).

Gap acceptance is defined as the percentage of drivers thatwould have pulled into the roadway with a vehicle approaching atthat distance. Take the speed argument and apply it to the gapacceptance evaluation. By traveling faster than the flow oftraffic, the intruder would have perceived that he or she had moretime that was truly available, making it more likely that theywould attempt to cross.

Hopefully this helps illustrate why a math-based approach canoffer a competitive advantage in liability negotiations. In theremaining segment, let's look at a hypothetical case.

Path Intrusion from a Side Street

The insured is an elderly driver with her husband in thepassenger seat of their full size sedan. She is stopped on a sidestreet at the sign and pulls westbound to cross the roadway. She isstruck in the left rear half of the vehicle by a 20-year-old maledriving a sports coupe northbound in the left lane. During thecourse of the investigation, the young driver admits to traveling35 or 40 miles per hour in a 25-mph zone, and claims that theinsured pulled out so suddenly that he didn't have any time to hitthe brakes.

The police then cite the insured for failure to yield the rightof way. The claimant driver is alleging injuries that resulted notonly in treatment but also in lost wages. He retained an attorneywho has filed suit for policy limits of $100,000.

After an initial evaluation of the case with a right-of-wayviolation, the claims adjuster felt that they could accept 80percent and try to negotiate 20 percent on the approaching driverfor unlawful speed.

The Monte Carlo collision simulations for this loss showed thatthere was a 58-percent probability that this collisioncould have been avoidedby a vehicle operating between 25and 30 miles per hour. For a vehicle approaching between 35 and 40miles per hour, as admitted by the claimant driver, the probabilityof avoiding this collision dropped from 58 to 45 percent, making acollision 13 percent more likely. The youngdriver's decision to speed made a collision more likely than not.The driver's lack of response resulted in an impact over twice thespeed of an average driver presented with the samecircumstances.

The Monte Carlo analysis provides strong evidence that the youngdriver's unlawful speed was the proximate cause of this collisionand its severity. The claims adjuster is now in a position to arguefor more contribution on the approaching driver, despite the factthat his or her insured was cited for failure to yield.

Next time, we'll look at a scenario involving a left turn withapproaching traffic.