There’s a gap between outright fraud and analysis not making a completely accurate comparison. In this topic it’s tricky to make sure you’re comparing representative trips: for example, one study earlier this year found that Tesla's reported safety improvement disappeared after accounting for disparities in the age of the drivers, the type of driving, age of the vehicle, etc. – teenagers making bad choices cause a lot of crashes but the kind of people who buy a high-end car tend to be much older and are less likely to drive like that in any vehicle. Waymo is definitely saying things which look like they’re trying to compare apples to apples but it’s not hard to trip over data like this even if your intentions are fully honest. The fact that they’re publishing data and encouraging independent review is a good sign.

Laughing in Elon Musk.