Greasy Fork

Hello,

Your script is very good, but the total ratings and the star ratings are not always the same as Amazon...

Why ? Do you have any idea what could be wrong please ?

Thank you for your answer.

Google Chrome Browser with Tampermonkey

Two factors:

• Rounding error.
  We calculate the number of reviews simply by taking the total number of reviews and the respective percentage of each star rating. Those are given by Amazon as whole integers. So without laborous loading subsites in the back, we cannot go more exact as the info is simply not provided by Amazon. I don't know about the Amazon API, but using it would trigger too many connections during normal surfing, I guess. I'll keep it simple here (oh and I'm not good at JavaScript, that's another reason).
  Eventually this means (depending on their algorithm) if they say it's 11%, the real figure should be between 10.5% and 11.5%.
  The bigger the total number of ratings is, the bigger the deviation will be speaking in absolute reviews.
• Amazon is constantly messing around with reviews.
  • Throwing spam out
  • Deleting fake reviews
  • Not counting reviews that ended with a dispute but still show them
  • Asynchronous update of countings and calculations (which makes sense. Otherwise one new review would trigger the whole calculation / DB writing / site generating chain which is not only heavy load on a scale like Amazon but also often unnecessary)

This script is more made for giving you an impression if a 4-star product has just 3 bad ratings or 300 because we often do not look at the total amount or don't have a good feeling for how much 4% of 217 is.

Hmm it seems I cannot edit my post or trigger spam protection by posting the amazon link.

Just wanted to give an example with numbers. On amazon.com.be/dp/B004MMG34Y/

Given:
4534 total ratings
11% 3-star ratings (we don't get any more exact value)
We calculate:
11% of 4534 = 499 (of course we round again here, but that will give max error of 1)

On the review subpage Amazon says it's 515 3-star ratings.

Looking at the rounding, numbers should be between 10.5% and 11.5% which is between 476 and 521. Both calculated and given by Amazon numbers are in that range.

In general that means we have a derivation of 0.5%. On larger numbers that may be a surprising absolute amount. In the case of 4534 total reviews it means plus or minus 23 reviews.

Okay, I understand and thanks a lot for all these information :-)

On the other hand, the absolute review numbers don't appear anymore on the reviews page with the new version 1.2.2.

Some links as examples :
https://www.amazon.fr/dp/B07W5D4FTT
https://www.amazon.fr/product-reviews/B07W5D4FTT/ref=acr_dp_hist_4?ie=UTF8&filterByStar=four_star&reviewerType=all_reviews#reviews-filter-bar
https://www.amazon.com.be/dp/B004MMG34Y
https://www.amazon.com.be/product-reviews/B004MMG34Y/ref=acr_dp_hist_2?filterByStar=two_star&reviewerType=all_reviews

Fixed script insertion for reviews page. Thanks for the notice.

Including absolute review numbers in the popup/tooltip is a good idea and in the past I already included that function. Then amazon changed something and now it became quite complicated to do this. The popup is not simply displayed but it is generated at runtime, which means I need to have an observer running to insert the code when possible. It is doable, but I am not very experienced and need some motivation and time for that... Another complicating factor is product variations. They change big parts of the page without reloading the page. So you need another observer to trigger the other observer sigh

Hello,

All right, I understand.

Indeed, when websites are working very well, they are updated with often bugs and many of us don't understand why to break what was working very well before :-(
Otherwise the script works perfectly now and I hope it will help many people.

I thank you very much for all your work and for your superb script, it's very nice to you ;-)

You're welcome.