On a video with 45,000 views I think the number of votes total doesn't matter as much as how many are up and down.

So if there was a similar video with the same views and likes but 0 downvotes you think it'd be significantly better? I'd guess the difference was noise.

I mean "maybe," right? It just seems like a sample-size problem rather than indicating any bias one way or another. And of course, as far as these things go, you only find out the quality after you watch some or all of it.

Yes, I agree. (I was just using the example given, which probably wasn't the best example for illustrating the value of public downvote counts.)

BTW, I should have known that if I questioned the value of public downvotes on a thread of downvoting enthusiasts that I would... get downvoted.

Take away the dislike count and you take away the total count too.

it's a video with low number of views though, 56 total votes is probably not a signal in that case.

56 samples on a stochastic process gives a pretty narrow normal curve...