Mobafire League of Legends Build Guides Mobafire League of Legends Build Guides
's Forum Avatar

The Rating System...

Creator: Drunken Monk May 11, 2012 9:34pm
13 posts - page 1 of 2
1 2
Drunken Monk
<Member>
Drunken Monk's Forum Avatar
Show more awards
Posts:
11
Joined:
Aug 5th, 2011
Permalink | Quote | PM | +Rep | Commend May 11, 2012 9:34pm | Report
... Is fine the way it is, but I dont understand why mobafire opts to show a percentage rating as opposed to simply showing the number of likes/dislikes like solomid.net for example. That percentage makes it hard for new guides to get off the ground. If you have a new guide that has 18 upvotes and 2 downvotes, thats a great ratio to have for a new guide. But how many people are going to be willing to try out a guide with an 80% rating when theres guides for the same champion that are 90% rated. It just seems like a very silly and superficial way to rate a guide as opposed to just showing the raw numbers to people. And the weird thing is that you guys USED to show likes and dislikes but now you just show the number of votes a guide has. Why?

I also dont understand why guide makers dont have the option to keep on comment to vote after the first 20 votes. That would combat so many troll votes.
A Chubby Baby
<Veteran>
A Chubby Baby's Forum Avatar
Show more awards
Posts:
3936
Joined:
Feb 14th, 2012
Permalink | Quote | PM | +Rep | Commend May 12, 2012 1:01am | Report
for the percentage calculation, check out jhoijhoi's guide for "how voting works", i personally really agree with you for the "keep on comment to vote after 20 votes thing"
Lilpwninprincess
<Member>
Lilpwninprincess's Forum Avatar
Show more awards
Posts:
213
Joined:
Jan 29th, 2012
Permalink | Quote | PM | +Rep | Commend May 12, 2012 1:38am | Report
I really wish we could keep the comment to vote thing on. It helps avoid the troll votes that just down vote and don't leave comments :<.
Lugignaf
<Moderator>
Lugignaf's Forum Avatar
Show more awards
Posts:
10967
Joined:
Feb 8th, 2011
Permalink | Quote | PM | +Rep | Commend May 12, 2012 12:10pm | Report
Everyone that thinks keeping the C2V feature on after 20 votes: We had that on for quite a while. Unfortunately, that led to massively over-inflated guide scores with no way in hell to bring them down.

While this is fine for actually good guides, if it's bad and it go top the top, like so many did, it would never come down. This was shown by a mejais-Morgana guide. While it sounds good in theory, in practice, it never worked out.

There are also a lot of other "good" top guides that do not deserve their place because of how bad they are from this old system.

Trust me, I've seen both sides of it.

Also, having the C2V on makes everyone scared to DV.

Let's say I went on a mass DV spree of the bad Maoakai guides that were over 20 votes. Let's say they're AD/DPS Maokai, just for example's sake. Now, with C2V on, people look at me thinking, "he just wants to inflate his own score" and everyone comes to DV me when all I was doing was weeding out the bad guides. If I don't vote down on them, they may get to the top and we'll see DPS-kai running around.

TL;DR not having the C2V on after 20 votes is good for everyone. We tried it the other way, it didn't work


Thank you to jhoijhoi for the sig, and all the dividers in the guide.

PsiGuard
<Community Manager>
PsiGuard's Forum Avatar
Show more awards
Posts:
7973
Joined:
Jun 26th, 2011
Permalink | Quote | PM | +Rep | Commend May 12, 2012 12:24pm | Report

It helps avoid the troll votes that just down vote and don't leave comments :<.

Actually, a lot of those aren't troll votes. Everyone just likes to think that they are because there couldn't possibly be something wrong with their guide. Matt has made it clear several times that everyone has a right to one vote per guide on the site, for whatever reason.

This is a topic that has been discussed a lot and I think Lugi explained it fairly well.
Thanks to TinyStar for the sig!
Drunken Monk
<Member>
Drunken Monk's Forum Avatar
Show more awards
Posts:
11
Joined:
Aug 5th, 2011
Permalink | Quote | PM | +Rep | Commend May 12, 2012 1:00pm | Report
I regret posting that last sentence because things got massively off track.

The main point of this thread was a percentage rating vs just showing likes/dislikes. If nothing else, Id like some kind of explaination as to how a percentage rating is beneficial, because to me, it just seems to hurt guides that are starting out. Especially guides that try something new and I honestly believe that to be the rating systems fault.

An 18 like / 2 dislike ratio is going to be more attractive to a reader then a 80% rating when the top rated guide is 90% rated. The percentage rating leads the reader to believe that theres a massive difference in quality between an 80% rated guide and a 90% rated guide when thats really not the case. The 80% rated guide just has less votes and a couple downvotes drastically impact the rating of a new guide. Thats makes it hard for a new guide to take off. Unless youre Jhoi or Duff and you get 50 votes on your guides first day.
Lilpwninprincess
<Member>
Lilpwninprincess's Forum Avatar
Show more awards
Posts:
213
Joined:
Jan 29th, 2012
Permalink | Quote | PM | +Rep | Commend May 12, 2012 3:28pm | Report

I regret posting that last sentence because things got massively off track.

The main point of this thread was a percentage rating vs just showing likes/dislikes. If nothing else, Id like some kind of explaination as to how a percentage rating is beneficial, because to me, it just seems to hurt guides that are starting out. Especially guides that try something new and I honestly believe that to be the rating systems fault.

An 18 like / 2 dislike ratio is going to be more attractive to a reader then a 80% rating when the top rated guide is 90% rated. The percentage rating leads the reader to believe that theres a massive difference in quality between an 80% rated guide and a 90% rated guide when thats really not the case. The 80% rated guide just has less votes and a couple downvotes drastically impact the rating of a new guide. Thats makes it hard for a new guide to take off. Unless youre Jhoi or Duff and you get 50 votes on your guides first day.


I agree that It'd be more appealing to see 18 Likes / 2 Dislikes instead of %. Especially for guides starting off. One dislike whether its genuine or a troll vote just drops the % quickly. I get what your saying and I like the idea honestly. It'd seem more appealing.

- also @ Lugi; that sucks that it didn't work out D:, didn't know it's been tried already I've seen so many people post about wanting to keep it on haha glad to know the reasoning behind why it's better not to now :D
Lugignaf
<Moderator>
Lugignaf's Forum Avatar
Show more awards
Posts:
10967
Joined:
Feb 8th, 2011
Permalink | Quote | PM | +Rep | Commend May 12, 2012 4:52pm | Report

An 18 like / 2 dislike ratio is going to be more attractive to a reader then a 80% rating when the top rated guide is 90% rated. The percentage rating leads the reader to believe that theres a massive difference in quality between an 80% rated guide and a 90% rated guide when thats really not the case. The 80% rated guide just has less votes and a couple downvotes drastically impact the rating of a new guide. Thats makes it hard for a new guide to take off. Unless youre Jhoi or Duff and you get 50 votes on your guides first day.

Completely forgot to comment on this.

The thing that the guides are scored on is the bayesian average, not the ratio of likes to dislikes.

I'm not going to claim to know how it works but you can google it and probably figure out how it works.


Thank you to jhoijhoi for the sig, and all the dividers in the guide.

jhoijhoi
<MOBAFire Mother>
jhoijhoi's Forum Avatar
Show more awards
Posts:
14001
Joined:
Mar 20th, 2011
Permalink | Quote | PM | +Rep | Commend May 12, 2012 5:30pm | Report
I actually agree with the OP. 18/2 is much better than 80%. It'd also show the really "bad" top rated guides with 1500/500 - you can EASILY see that 1500 likes is good, but a guide with 500 DISLIKES must be bad.
GrandmasterD
<Member>
GrandmasterD's Forum Avatar
Show more awards
Posts:
7950
Joined:
Sep 26th, 2011
Permalink | Quote | PM | +Rep | Commend May 12, 2012 5:48pm | Report
It's just that the % part of the rating makes no sense. It implies that 80% of the votes are upvotes which is not true since it uses the Bayesian average system. Not sure which constants mobafire uses though.
1 2

You need to log in before commenting.