YouTube removed 3.9M dislikes for White House | Hacker News Reader

YouTube removed 3.9M dislikes for White House

phzoe.com

65 pointssmsm4250 comments

The tone in this is... Less than stellar.

This is standard operating procedure, mainly dealing with deleted accounts, and reducing spam or bot actions on the site. 3.9M removed over the course of 500+ videos for a global account for one of the largest countries isn't all that surprising. They also remove likes and comments for spam reasons as well, which this "article" (if you can call it that) just so happened to ignore, I can't imagine why.

YouTube prunes accounts quite often, even streamers and boy bands will have their numbers rapidly changed to deal with spam or bots. This has been normal operating procedure for over a decade.

> The tone in this is... Less than stellar.

Agreed. I got down to the "So odd for the most popular resident ever!" and realized I got duped into reading some ideologue's rant against the other team and now I have to question the data provided and wonder what context got casually discarded.

This whole "Youtube deletes dislikes on White House videos" has been part of the alt right narrative for some weeks now, right next to "if masks work then why do we need to stay at home".

The data shows they deleted: 9005 Likes 3.9 Million dislikes 750,000 views

You're welcome to think that nearly twice as many diskikes were removed as remain is normal. I disagree. Thank you.

> is normal. I disagree.

Why though?

What's normal and can't abnormal things happen?

Personally I have no idea how likes vs dislikes are likely to be distributed in any given set of artificial operations. Why are you so sure about it?

Do you know how the sausage is made? I would assume only certain Youtube employees would know the actual truth. If you look at the About page for 81m.org, he writes about how he verifies with other large Youtube channels. Not that I disagree but the quickness of your dismissal is a bit unfair to the OP.

This is presented in a way that infers political partisanship, but without any evidence for or against that argument.

- No statistics for previous administration's like/dislike removal rate. Was it lower? Higher? We don't know.

- No statistics for YouTube like/dislike removal rates on other kinds of political or non-political videos. Maybe this is extraordinary behavior for YouTube, or maybe not. We don't know.

In other words, this is flame bait that people can read into whatever they want to confirm their biases. With a heavy dose of partisan editorializing ("Disgusting!").

At best, it shows a need for more principled study from someone else.

It's alt right propaganda. On the frontpage of HN.

If we wanted to be truly scientific about this and make for an interesting conversation, we would try to get to the bottom of this instead of shooting down the OP.

First mistake is assuming that any YT numbers are real. Second mistake is assuming that YT numbers are accurate. Third mistake is assuming that YT numbers represent what you think they represent.

YT numbers are definitely neither real nor accurate: https://lee-phillips.org/youtube/

I would assume YT stats often go through some sorts of spam detection ... thus this?

I see YT stats change up and down a lot. On videos with lots of likes and dislikes, and videos with almost none. I just assume it's spam / false positive detection / database synchronization... or whatever system at work.

Maintaining any kind of stats like this that matter / on a huge site has to be a 'try to keep it in the ballpark' at best scenario / endless adjustments.

It doesn't take much deep thinking to realize how these numbers are going to change for all sorts of legitimate reasons, in any direction. Maybe a number of times, tomorrow even?

Heck maybe if there was spam had nothing to do with the videos / source of the videos even? If I'm going test some up / down vote bottling ... probably going to pick a big target to see how far I can push the numbers.

9005 spam likes 3.9 million spam dislikes

Makes sense.

Dislikes removed, or bots banned (taking the votes with them)?

Youtubers value likes so the volume of like or dislike spam I suspect is pretty high.

I see likes and dislikes change over time even on videos with very few likes and views.

I just assume it is some sort of cleanup system.

I see a lot of people refuting this data, yet provide no counter evidence. Let's remove politics from this, is YouTube being deceptive? I am honestly asking.

I have noticed in the past controversial videos have their views/dislikes reset. Anecdotal, but I would like to see if there is truth to vote and view manipulation.

Hell look at sometimes how manipulated the trending pages are. I remember for months low view count Jimmy Kimmel clips were boosted to trending. Logged in or not. Seems like mainstream Hollywood stuff is far more likely to trend.

YouTube is silent about explaining why this happens. It can't all possibly be spam and bots.

It's similar to what Google does with other products like their Play Store.

Things like the removal of 1 star reviews [0] when an app rating goes down. They call this "review-bombing" protection.

And perhaps the easier option for this kind of official videos is disabling comments and like/dislike buttons for them.

[0] https://www.theverge.com/2021/1/28/22255245/google-deleting-...

Can you imagine how much of "scandal" this would have been if this happened during the last administration?

Maybe on twitter.

Otherwise I don't think anyone would care.

CNNs drop in ratings would beg to differ.

I have no idea what that has to do with an imagined scandal... over youtube down votes... like none.

You said no one would care other than twitter. However cnn made it's money the past four years sensationalizing everything about the previous administration, so that is evidence that people other than twitter care.. CNN's former viewers.

Whatever CNN is, I think it is more than and has existed longer than "made it's money the past four years sensationalizing everything about the previous administration".

Let alone how that ties to its ratings.

You're out in the weeds with these characterizations and connections.

CNNs ratings have dropped precipitously after Trump left office, after rising quite a bit during his time in office. Several accusations against their hosts, such as Maddow, have been met with responses of 'this was exaggeration' in courts of law (an argument judges have accepted).

With that in mind, and keeping in mind CNN is a pretty mainstream news source, it's pretty obvious that a large segment of CNNs viewers did care about such scandals during the previous administration, which means quite a bit of non-tweeters did.

I gotta tell you every post is like a couple opinions and then these connections and assumptions that just don't make sense in any way.

You are comparing breaking treaties, attacking allies, sympathizing with our long time enemies, rolling back regulation, mishandling pandemic to youtube removal of votes?

If this is the biggest thing the right media can sensationalize, I say the current administration is doing pretty damn good.

There is nothing wrong with rolling back regulations.

Current admin jokes about nuking Americans.

But I guess making peace with enemies is worse yeah

Not when enemies are continuing hostile actions toward us.

It would make the rounds on conservative media[1].

[1]: https://www.foxnews.com/tech/trump-complains-twitter-purged-...

This is a scandal regardless of administration. That's why it's on the front page on Hacker News and being written about by major media publications.

you're supposed to just write in all caps "why is the media silent about this" in classic unsubstantiated fashion

I should add, "this would be seen as a much bigger deal and harped on extensively".

Do you think that's hypocritical? Trump and his most visible supporters made provocation and mockery a hallmark of their style, so of course people react more strongly to whatever they do. It would be unrealistic to expect otherwise.

No, I'm talking about sensationalized media coverage. How scandalous something is, is largely a function of how much the news talks about it.

What makes you think that it wasn't also happening during the last administration?

No one normal cares about youtube video rankings. That's why there wasn't any website like 81m.org site tracking white house youtube videos and hiring bots to downvote them.

It just silly. Regular people who are interested in politics care about actual laws passed, relationships with other nations etc, not some yotube, twitter, fb wars.

The big question that is missing to be answered is what it looks like generally on youtube, especially with highprofile youtube channels with a lot of people who dislike them, it is rather rare where more than half of all viewers dislike a video.

A incident like this also happened in the past and there youtube said that they remove malicious or spammy votes, though did not seem to comment on the concrete case. https://www.politifact.com/article/2021/jan/25/did-youtube-r...

Is there a less biased source doing similar research?

Now all the people have to re-dislike.

This guy also has a bunch of posts on the same site denying climate change and rehashing conspiracy theories about fraudulent votes for Biden. I wouldn't take anything he says very seriously.

I'm a little shocked that this is getting so much traction here. The site (81m-dot-org) seems to be a right-wing "the election was stolen from Trump" site. Their "about" page says:

It is true that Biden won the most votes cast for a U.S. presidential candidate, amassing about 81 million votes. Such a popular president should have some truly epic YouTube numbers. Let's keep track. I'm sure we won't see anything weird happening, such as dislikes being thrown out in the wee hours of the morning.

When you're dealing with "voting" on the internet (like Reddit, YouTube, or HN do), there's going to be vote manipulation. Some of it you might be able to catch early on. Some of it you'll end up fixing with batch jobs that identify "huh, there seems to be a few hundred users using this IP address and all watching this video and hating it at exactly the same time."

This blog linked here has such posts as "Something Rotten in Georgia" where she questions the accuracy of Georgia election results (https://phzoe.com/2021/01/06/something-rotten-in-georgia/) and "81 Million Ballots" where she says that "there is ample reasons to suspect that foul play bumped [the number of votes for Biden] by several million" (https://phzoe.com/2021/03/12/81-million-ballots/).

Are we really upvoting right-wing propaganda now?

Exactly same thoughts. Not only right-wing propaganda but straight up conspiracy nonsense. How is this getting traction on HN!

Not surprised. Youtube has been known to sell likes/dislikes for years. "Trending" videos for Pop Music will be minutes old yet have thousands of likes and views. The views and like/disklike ratio are about as useful as comments or Twitter followers.

>>Youtube has been known to sell likes/dislikes for years

How do they actually do that? First I hear about it.

I don't know it's actually youtube that does it. It's actually known that you can just buy an influence by purchasing access to bot accounts[1]. This happens on YT, Twitter, FB and other places, and it isn't just likes/dislikes but also comments.

[1] https://www.youtube.com/watch?v=1PGm8LslEb4 ha he has more videos about that

Known by whom lol the flatearthers?

Reply on news.ycombinator.com