You could have been a legit viewer... clicking to skip over segments of the video, presumably trying to find where you left off last time, or for some scene you remember, or the climax of the video... whatever.
Youtube does try to throttle the data speeds, when that first happened, youtube-dl stopped being useful and everyone upgraded their python versions and started using yt-dlp instead.
If you click to skip over, even clicking every minute, you're still not grabbing the whole thing, right? Whereas downloading is grabbing every second.
Depending on the player and how they cache it. Yes, if google monitor every byte to which client had downloaded, but that just seems like ultra micro managing, and have no idea how many players will it break. Youtube seems like one of those site, should allow people to download or make them a public utility on IPFS or something like that.
You are actually. Watch one second, the player buffers the next minute of video, then you skip ahead 1 minute. Process repeats.