It seems like a YouTube bug, that they are performing view tracking on the client, when they own the whole server backend and could just as well track them server side (which wouldn't be blockable in the first place)

▲

nonameiguess 5 days ago | parent | next [-]

It seems like server-side would suffer from issues due to buffering lookahead and autoplay. A client can request a video that is skipped within seconds, but if buffering causes it to request five minutes worth, the server only sees five minutes were requested, whereas the client can clearly tell how much of that was actually watched.

▲

paxys 5 days ago | parent | prev [-]

Server-side tracking would be the easiest thing in the world to spoof.

▲

swiftcoder 5 days ago | parent [-]

How is it any easier to spoof than client-side tracking?

In the server-side case I can certainly increase views by fetching the video multiple times, but in the client side case I can hit the analytics endpoint directly just as easily

	▲	paxys 5 days ago \| parent [-]
		Bot detection systems all work client side. What browser are you using, what cookies do you have, how is your mouse moving, how much time are you taking between clicks, how many captchas have you solved. All this and more is collected and passed to the server, which can then determine if the view was valid or not. Plus you can do this multiple times during a watch session rather than just in the beginning when the video is requested. If the adblocker is blocking this data then the server has nothing to go on.