Tracking visitors at the client level deflates the actual number of visitors. On the other hand, server-side tracking provides a more accurate number with the tradeoff of not knowing for sure if the client is a human behind a browser.
But even if ad-blockers were not used, people still disable JS, JS files fail to load, JS takes longer to load than the user stays on the site and many other potential issues.
My comment was not directly related to yours, I agree that DNS analytics are probably worse, I was just wondering if theoretically is it possible to produce high-accuracy analytics when everything can be spoofed/cached, etc.