3 ms·

> "the number of views is also a better indicator of potential quality" Can't agree. Perhaps you've forgotten about clickbait thumbnails and video titles. "To

> "the number of views is also a better indicator of potential quality" Can't agree. Perhaps you've forgotten about clickbait thumbnails and video titles. "Top 10 Scariest Roller-coasters" with photo-shopped thumbnail of a coaster car mid-air between fake section of track. Literally millions of views from people wanting to see this incredible coaster that jumps the gap. What you find is an ordinary video of copied, grainy clips of ordinary coasters. Previously, such a video had a HUGE number of downvotes. Now those dislikes are hidden, the deception is permitted to flourish.

unclebucknasty 5y ago
I mean, all stats are game-able, so it's really relative. Dislikes seems among the easiest. So, again, I said views are (generally) a better indication than dislikes, not a perfect one. >Literally millions of views from people wanting to see this incredible coaster that jumps the gap. Don't know how much of a vid must be viewed before YouTube counts a view, but that's also where my suggestion to rate by percentage of the video viewed comes in. >Now those dislikes are hidden, the deception is permitted to flourish I think (hope) most people are trained to ignore those fake vids over time, but content that is outright misleading should really be flaggable and removed by moderators anyway vs hit with vague dislikes.
- exodust 5y ago
  Yes but people watch the whole video waiting for the thumbnail's promise that never arrives. Partial view stats won't help in that case. Link integrity has always been a thing. Fake thumbs and misleading video titles are a link integrity issue. By hiding downvotes, Youtube has awarded a gift to clickbait content creators. Don't forget clickbait thumbs and titles range from blatant lies to "somewhat exaggerated". View count won't indicate quality for those. Besides, a video released only 2 weeks ago might be high quality but low view count. 200 likes and zero dislikes for example on a new video from a low-key channel about a niche subject, is a very good indicator of quality for that new video. Now with that ratio hidden, all we see is "200 likes" but with possibly higher dislikes hidden from view.
  - unclebucknasty 5y ago
    Again, your entire position is premised on a reliably non-troll abused and reasonably uniform usage of the dislike button, two things we have reason to believe aren't the case. >people watch the whole video waiting for the thumbnail's promise that never arrives. Partial view stats won't help in that case. Neither signal is perfect, but it's easy to simply smash the dislike button for any or no reason. OTOH, actually watching the video takes effort and commitment, which is in itself a signal. Many people will give up fairly quickly. So, the chances that the majority of people are watching entire vids waiting for the thumbnail to happen is IMO substantially less than the chances that people just effortlessly smash dislike. And, I think people get trained pretty quickly to avoid the "impossible" thumbnail with massive landsharks eating trains or whatever. >video released only 2 weeks ago might be high quality but low view count Relative like-to-view ratio is probably a better indicator than factoring in the oft-abused dislike button. >200 likes" but with possibly higher dislikes hidden from view. Again, YouTube should offer a flag button for scammy or otherwise disagreeable content, then moderate accordingly. That's far better than a generic dislike button. But, we're probably going to agree to disagree here.

unclebucknasty 5y ago

I mean, all stats are game-able, so it's really relative. Dislikes seems among the easiest. So, again, I said views are (generally) a better indication than dislikes, not a perfect one. >Literally millions of views from people wanting to see this incredible coaster that jumps the gap. Don't know how much of a vid must be viewed before YouTube counts a view, but that's also where my suggestion to rate by percentage of the video viewed comes in. >Now those dislikes are hidden, the deception is permitted to flourish I think (hope) most people are trained to ignore those fake vids over time, but content that is outright misleading should really be flaggable and removed by moderators anyway vs hit with vague dislikes.

exodust 5y ago
Yes but people watch the whole video waiting for the thumbnail's promise that never arrives. Partial view stats won't help in that case. Link integrity has always been a thing. Fake thumbs and misleading video titles are a link integrity issue. By hiding downvotes, Youtube has awarded a gift to clickbait content creators. Don't forget clickbait thumbs and titles range from blatant lies to "somewhat exaggerated". View count won't indicate quality for those. Besides, a video released only 2 weeks ago might be high quality but low view count. 200 likes and zero dislikes for example on a new video from a low-key channel about a niche subject, is a very good indicator of quality for that new video. Now with that ratio hidden, all we see is "200 likes" but with possibly higher dislikes hidden from view.
- unclebucknasty 5y ago
  Again, your entire position is premised on a reliably non-troll abused and reasonably uniform usage of the dislike button, two things we have reason to believe aren't the case. >people watch the whole video waiting for the thumbnail's promise that never arrives. Partial view stats won't help in that case. Neither signal is perfect, but it's easy to simply smash the dislike button for any or no reason. OTOH, actually watching the video takes effort and commitment, which is in itself a signal. Many people will give up fairly quickly. So, the chances that the majority of people are watching entire vids waiting for the thumbnail to happen is IMO substantially less than the chances that people just effortlessly smash dislike. And, I think people get trained pretty quickly to avoid the "impossible" thumbnail with massive landsharks eating trains or whatever. >video released only 2 weeks ago might be high quality but low view count Relative like-to-view ratio is probably a better indicator than factoring in the oft-abused dislike button. >200 likes" but with possibly higher dislikes hidden from view. Again, YouTube should offer a flag button for scammy or otherwise disagreeable content, then moderate accordingly. That's far better than a generic dislike button. But, we're probably going to agree to disagree here.