Jump to content

Alternate Rating option Like/Dislike Button


Recommended Posts

Some research suggests it is difficult for humans to assign a value to an object. But it is easy when the choice is clear cut as in Like/Dislike.

My suggestion is we change the rating system to a Like/Dislike system instead of the 0-5 Star system that is currently in place. You can still display the ratting in terms of stars as follows:

if Likes != 0 then 
Stars = 5 * Likes / (Likes + Dislikes)
Stars = 0

one line in C

stars = likes?5*likes/(likes+dislikes):0;

How it might work...

Initially both Like and Dislike buttons are active, when one is pressed the users vote is submitted. The button pressed becomes inactive while the other remains active. If the other button is now pressed it becomes inactive and reactivates the original button, this also reverses the previous vote made.

You might also like to display a tool tip when hovering the mouse pointer over the stars, something along the lines of "This torrent has a rating of 4.7 Stars."

Link to comment
Share on other sites

I was really not a fan when Youtube, as an example, moved from 1-5 to +/-. I'm somewhat used to it now, but I'm glad they simply show a like/dislike ratio. What you're suggesting is a completely arbitrary quantification of an very broad set of responses. + can be anything from "Pretty good" to "I would kill my mother for this on DVD", and - can be anything from "This was kinda weak." to "I wish these people were dead."

If the spectrum of reponses to a particular item is distributed between "This is the worst thing ever" and "This is just OK", your method would average out to around "2.5 stars" which is essentially neutral, when in reality it should probably be around "2 stars" and clearly into negative territory. Additionally, if you only have + and -, you leave out a neutral option, and if something is boring, but inoffensive, you could be losing a significant volume of response which should drag the result towards the centre. I guess you could say you're sacrificing sample size and losing the stability that brings?

I'm on a site which was considering a similar change, around the same time Youtube made theirs and what seemed an acceptable compromise to me was to have something to the effect of --,-,/,+,++ and assign 1-5 to those values. For those who have trouble quantifying their response, they have two negative responses, a neutral, and two positive which would be assigned to 1-5 (which is essentially how I was quantifying my responses numerically anyway...) However having seen Youtube working for some time, I'm not sure they don't have it right, for a basic +/- methodology. (Which I still don't find ideal)

Another site to consider might be how Rotten Tomatoes does it. I find that in aggregate, the likelihood that a strong percentage (say an RT score of 90%) will find something good (good being >60%?) but that you will find it poor is small, because outliers like yourself are a part of the pool which is establishing that rating.

On the other hand, all ratings systems require some familiarity to truly gain the most benefit. 80% on one site is not 80% on another. IMDB for the longest time (still?) had an algorithm for rating which penalised small voting pool results to prevent "skewing". In the end I found that a film that was 7.5/10 but with only a few hundred votes might actually be equivalent to a blockbuster rated 8.5/10 in quality*. Niche films suffer on RT because they were never intended for a broad audience, and so you really ought to consider them at a handicap. (This "behind the scenes" skewing can be somewhat analogous to the "culture" of a site or collection of individuals.)

Also, never disable either, as clicking on it again should retract the decision. If I click + accidentally, I may not agree with - either, so you're forcing me to stick to one or the other because of a stray mouseclick. This is a pretty big peeve of mine.

(All this, and I have a post currently that says I don't even want to use the feature... heh.)

* Edit: Not just my own personal opinion, even, but when I did the math to determine the average, there were cases when it should have been 8.5-9.0 but I guess that'd screw up their Top 250?

Link to comment
Share on other sites

Very interesting. I have to say, I enjoyed reading your response. I only have one problem and that is with your assumption that everything will end up with a neutral ratting. You say niche films will suffer but who will be downloading them? Nobody is going to download a torrent for an independent film just to give it a bad ratting. Only the people who actually want to watch it will be ratting.

For example, I hate, well I can’t stand, "Harry Potter", so you'll never see me downloading it, hence my vote will never be counted exactly like it shouldn’t be. The star rating on torrents directly relates to torrent quality not content type^. Well at least that’s the way I see it. I'm not going to download some trashy soap opera either. But the people who want it will and they are the most qualified to rate it.

^ By quality I mean, apps complete and working or video watchable not blurry with terrible audio and the torrent offering what it promises i.e. is not a fake

If you want to abstain from voting that’s fine, but I totally disagree with a neutral button. A torrent quality wise should be either good enough to like or bad enough to dislike. Comparing the torrent rating to the film it contains on IMDB is like comparing apples to oranges. Personally, I think, if people are using the rating system to rate the content rather than the torrent we should just abandon the ratting system altogether.

Link to comment
Share on other sites

Ah, see, I didn't see anything about it referring to the torrent quality, rather than the content. I may be mistaken as to the goal of the rating, but in that case I think it's unclear in the client itself.

Now, I don't think everything will default to neutral, I was simply bringing up a flaw in +/- in that a widespread neutral response would go unrepresented, and could allow for a relatively smaller + or - response to dominate.

Example, say 100 +, 300 n who may not vote if there's no neutral option, and 10 -

Your formula would give that 4.5 stars, I think. 100/110*5

If + = 5, n = 3 & - = 1 that's 3.4

That's substantial lost data by removing the possibility of a neutral option. That's why I think +/- should be represented like Youtube or as a ratio. So it deemphasizes any numerical rating association and just presents likes vs. dislikes. In that case you know inherently that only those people who cared to answer are factored in, but if I saw something rated "3.5 stars, I would assume that was made up of people who also were neutral or weak+ or weak- about it.

You do bring up a good point though, that in a case like this, most people who will download niche or specialty torrents will likely not have done so because they stumbled onto it randomly, whereas a site like Rotten Tomatoes has people who specifically seek out films at random.

And my other post here is asking how to hide the ratings box completely from the delete dialogue :P

Link to comment
Share on other sites


This topic is now archived and is closed to further replies.

  • Create New...