Almost all of Waymo's data is with a safety driver and it shows that the safety drivers were able to prevent 29 collisions that would have occurred had the system been operating without them (there were 18 actual collisions in 6.1 millions miles). I admit it does seem likely that a heavily monitored Waymo employee is better than the average driver.The way accidents per mile is calculated is fraught with issues. If you think Waymo is in more accidents than average driver then examine your sources. For example are statistics based on serious accidents, accidents reported to the police, accidents where driver is at fault? May find it hard to correlate with the Waymo data.
You might be surprised how sophisticated the 30 billion miles of simulated driving is.
The number I've heard is 1 collision per 150k miles and Waymo is worse than that. I also see that there were 8 collisions that resulted in air bag deployments in 6.1 million miles which is worse than Tesla's numbers (though I think Waymo is including the other vehicles involved in the collision). I also wonder if Chandler, AZ might have a lower accident rate than average? Of course getting rear ended is much better than many other types of collisions and that should factor in to safety calculations. I agree that it's hard to correlate.
I'm not a big believer in simulations for predicting severe accident rate.
Anyway, I do think it's plausible that Waymo is safer than the average human but it doesn't seem conclusive to me.