Artillery is not an exact science, if you believe the crap of surgical hit, you are very naive. Sure there are progress but you can still have 100m of error rate.

This wasn't 100m of inaccuracy. Russia shot artillery into Kyiv, into civilian residential areas. They shot into an area with a shopping mall, then lied about it being "nearly empty".

What is the sample?

If you'll refer to my initial comment, I was expressing some doubt about this. I'm guessing the Donbass and Crimea were not included.