Discussed hundreds of times, Matt has stated that they are looking for alternative ways to display score.
If i understand correctly though, you just say they keep the % score, just add a +/- counter?
If i understand correctly though, you just say they keep the % score, just add a +/- counter?
throatslasher wrote:
I've always found the % value on guides confusing. What I would absolutely adore on guides is an additional section for the score where it shows the # of upvotes and the # of downvotes. My karma guide has 41 votes and 80%. I honestly have no freaking idea how many upvotes and downvotes I have, which I feel would be useful information. I don't think it's a 80/20 split like the 80% would suggest. Any thoughts about my idea? Would it be damaging in any unforseen way?
It's the Bayesian average.
We had #of upvotes/downvotes. I think we removed them because people were getting mad, or something.
I'd really like to have a more accurate score display too.
JEFFY40HANDS wrote:
But yes this idea has been discussed more than Pam Anderson's tits in a plastic surgeons office.
I personally believe a upvote/downvote system would be much more accurate. The current % system is damn confusing seriously :/
+rep me and you'll get good karma, wich means you won't be reincarnated into a rock! Isn't that awesome?
JEFFY40HANDS wrote:
But yes this idea has been discussed more than Pam Anderson's tits in a plastic surgeons office.
You are right that it is not going to be exactly a 20/80 split, due to the formula mentioned by Lugignaf. And Anastasios is right too, we have been discussing some alternative score displays. We're just not decided on what is best, yet.
There are some issues with the current display. Due to that formula, new guides look like they have a bad score, which sucks. Guides that have a solid number of votes are also still not showing their "true" score. Authors tend to get SUPER hung up on the score and become extremely demotivated and agitated if it drops by just 1%.
Additionally, people view score as a statement of "this is the best way to play this champion", when that isn't really what score represents on MOBAFire. Instead it represents "this is the guide that the most people enjoyed reading and found useful". For example a guide with a better build and little to no explanation could potentially end up ranked lower than a guide with an average build but a well written guide, because that is what the readers appreciated more.
All if this can be a little confusing to new users, and frustrating for authors. We have a number of ideas to help combat this and are already working through some of them.
We've implemented the Elo verification system, and are in the middle of implementing a series of new filters for browsing, that will help drive users to guides that are more relevant to what they want. This will help improve the reader experience and should provide authors with more interested readers. These features will be finding their way into other sections of the site as well, such as the champion pages.
A lot of the issue also revolves around the way scores are displayed. We need a rank in the back-end, but we don't necessarily have to use that as the display (which we do now). We could use the ranked score quietly and instead display information that is more useful and enjoyable for both authors and readers. This is what sites like Reddit do - their ranking scores are hidden, you just see the "true" score and know that more is going on to rank content. This updated display could be what you said, just the up vote and total vote counts, or it could be a 5 star/bar display like we're currently using on DOTAFire and SmiteFire, which users on those sites seem to like better.
We're open to other suggestions, as well as feedback on some of the above options :)
There are some issues with the current display. Due to that formula, new guides look like they have a bad score, which sucks. Guides that have a solid number of votes are also still not showing their "true" score. Authors tend to get SUPER hung up on the score and become extremely demotivated and agitated if it drops by just 1%.
Additionally, people view score as a statement of "this is the best way to play this champion", when that isn't really what score represents on MOBAFire. Instead it represents "this is the guide that the most people enjoyed reading and found useful". For example a guide with a better build and little to no explanation could potentially end up ranked lower than a guide with an average build but a well written guide, because that is what the readers appreciated more.
All if this can be a little confusing to new users, and frustrating for authors. We have a number of ideas to help combat this and are already working through some of them.
We've implemented the Elo verification system, and are in the middle of implementing a series of new filters for browsing, that will help drive users to guides that are more relevant to what they want. This will help improve the reader experience and should provide authors with more interested readers. These features will be finding their way into other sections of the site as well, such as the champion pages.
A lot of the issue also revolves around the way scores are displayed. We need a rank in the back-end, but we don't necessarily have to use that as the display (which we do now). We could use the ranked score quietly and instead display information that is more useful and enjoyable for both authors and readers. This is what sites like Reddit do - their ranking scores are hidden, you just see the "true" score and know that more is going on to rank content. This updated display could be what you said, just the up vote and total vote counts, or it could be a 5 star/bar display like we're currently using on DOTAFire and SmiteFire, which users on those sites seem to like better.
We're open to other suggestions, as well as feedback on some of the above options :)
http://www.mobafire.com/league-of-legends/karma-guide
Shows the numbers there.
You have:
35/41 Upvotes
6/41 Downvotes
Shows the numbers there.
You have:
35/41 Upvotes
6/41 Downvotes
That actually makes me feel way better than 80%. I like that 86% of people liked my guide. Dunno. We couldn't rank guides like that, because 1 upvote 0 downvote would be 100%, but I don't nkow. Maybe you could rank guides that have over 20 votes on this formula-
upvotes / total votes = %
upvotes / total votes = %
We've considered modifying our ranking system to abandon the formula once you reach a threshold of votes. In the end it wouldn't matter too much, because right now all guides are in the same boat. What matters more is the perception of the score. 1 upvote 0 downvotes = just over 50%. That is fine for ranking but is terrible for perception. Hence why we are considering separating score display from ranking.
You need to log in before commenting.
<Member>