What wouldn't make sense is if you repeat this 1000 times and a car is never in your lane.
That means that something is wrong about how you are modeling the road and cars.
The claim that people are confused by is (asteroids with a 3% chance of hitting get the change revised to 0% more than 97% of the time).
There are parts of the road you can't see. If you were forced to estimate the chance you'd have to move for this particular purported car with just this information, you might throw up your hands and say P(they're telling the truth) * 1/4 until you can see it.
If a bunch of people like to lie, and you have nothing but one person's word, your move/no-move prediction for this car is likely to be wrong despite your model of the road-car-person system being perfectly right.
As more people say they saw the car, the estimate from these reports approaches 1/4, dominated by your understanding of the system. But you might have fewer early estimates that say you need to move per actual car and actual move, and many estimates for cars that don't exist.
I hope we're in a world where that makes sense and I'm actually right about this. But who knows!
When it gets a little closer, you can tell at least which half it's on, the left or the right. Now your estimate is either 0% or 50%.
Closer still and you tell which lane it's in, so now you're sure.