The parent comment is saying that the 300% increase is tied to a 300% increase in amount of driveless car activity. In which case, of course incidents increased. If you add 3x as many normal cars on the road the traffic incidents would increase in the same way!
> If you add 3x as many normal cars on the road the traffic incidents would increase in the same way!
Are you sure? In a lot of cases you would get traffic jams. The rate of incidents in a traffic jam are likely different from non-traffic jam conditions.
A study from Australia says: "Results showed an approximately linear relationship between traffic volume and accident frequency at lower traffic volumes. In the highest traffic volumes, poisson and negative binomial models showed a significant quadratic explanatory term as accident frequency increases at a higher rate." [1]
What's worse, a big point in the article is that we don't know exactly what the denominator is on the incident rate, and we are not even really sure what the numerator is either. The companies running these protracted live-road experiments are not being forthcoming with their data on how many cars they are running, when, and exactly what the incident count is.
Quoting the article:
"The city is left in the dark about the exact number of driverless taxis operating in its streets, and the miles they’ve traveled. Data captured by state regulators, officials say, doesn’t capture the extent of the vehicles’ disruption and potential hazard on city streets."
