111
Just Bezos Things (lemmy.ml)
submitted 9 months ago by to c/ManufacturingConsent@lemmy.ml
6 comments fedilink hide all child comments
 

Yeah sure, this movie with 1.3 rating and 5 (out of 100) metascore, with only 50k ratings, is the #1 movie this week. The algorithm is really transparent and trustworthy, clearly there's no manipulation going on.

(for those who don't know, IMDb is owned and operated by Amazon)

An image showing "Top 10 on IMDb this week", with number 1 being "Melania", with a rating of 1.3 stars
you are viewing a single comment's thread
view the rest of the comments
[–] [S] 18 points 9 months ago (2 children)

#3 has way more ratings and way higher ones too:

I guess it could be lower because it's a series or something, but #10 is a movie, has a higher rating, and has more ratings too:

My only other idea of what "popularity" could mean is the page view count, but then it should probably be called something other than "top 10 on IMDb" IMHO.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 6 points 9 months ago

    If I were tasked with implementing an algorithm tasked with being "justifiable" but also putting it at the top, I'd simply refuse.

    If I was a complicit piece of shit, though, I'd probably use engagement, scaled by recency... with whatever recency curve puts it at number 1.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 1 point 9 months ago

    Yeah but the list is “this week” not “this month” for the second and tenth ones or “all time” for the third one. So even if you got 91k reviews this month (#10 came out on jan 16th) that’s, let’s check it, an average of 30k a week. And most reviews would have come in the first week anyway.

    So this really cemented it even more for me, it’s literally just number of “actions” on any given listing in a week. Especially since the top two are really low rated.

  • permalink
  • fedilink
  • source
  • parent