Amazon reviews rating doesn't add up
amazon.comAnd the winner is: "Amazon calculates a product's star rating using machine-learned models instead of a simple average..."
https://www.amazon.com/gp/help/customer/display.html?nodeId=...
When you have a majority of one-star reviews that independently mention the fine bouquet of "stinky feet", let's pick 4.2. ¯\_(ツ)_/¯
On the plus side, it's no GPT-3 leaving such gems as:
"Looks like human skin, smells like gym shoes, this is a tragedy." and
"Smelled like stinky feet. They didn't taste like watermelon and had a nasty texture. I spent more time peeling them off the paper and couldn't enjoy them because I had to spit them out."
"My kids spit it out and so did I. If you still aren’t convinced and end up buying it anyway... you’ll regret it ... guaranteed... you’ve been warned."
Clearly some future food critics here. These are the absolute best negative reviews I think I've ever read. 4.9 stars!> Amazon calculates a product's star rating using machine-learned models instead of a simple average.
lmao! i thought you linked me to an April fools page for a second .-.
"63% of reviews are 5 stars" but when you filter to 5 star reviews, there's zero. small indie company. i'm guessing this is all just an excuse for Amazon to inflate their numbers
Ratings != Reviews, you can leave a rating without leaving a review. The math still doesn't add up, but it looks like there are actually 4 five star ratings. How that makes 63% of a total of 12 reviews is beyond me.
hmm, I wonder when they switched to machine learning though, because I remember some very mean technical analysis of their recommendation systems from 10+ years ago, where the problems were more the problems you would expect from naive averages and not taking into account outliers...
so maybe someone read that review and the recommended solutions and said screw that, we're going to do an ML solution, that'll show em.
Note: state of internet search being what it is not going to try to find that review of recommendation systems early Saturday morning before coffee.
anyone have an explanation for this? this isn't an issue of fake/bot reviews.
why does Amazon say the product is 4/5 stars even though no reviews are saying that... ?