I dunno, defaults are a tricky thing. Ideally these Nice / Good / Great levels should be dynamically scaled to the community norms, e.g. based on the median and standard deviation of the number of likes a typical post gets on that community.
Problem is, that’s a shifting target. What about in 2015 when the community is 6 times larger and the average post gets 3x the number of likes? Do we grandfather in all the old badge awards at the lower levels somehow?
I would prefer to err on the side of safer defaults and let communities scale these down to whatever level they need, rather than having them be too low at launch of a community and take things away from people.
Yep, one of the problems right now is that you can’t scale them at all: The badges are ‘system badges’, so you can’t edit the name, or query, or anything, except for disabling them outright. And the names aren’t that forgiving to adding a new one below them - What’s good, but lower than “Nice” and “Good”?
BBS definitely is fine with the current 10 likes threshold - they’re getting 5-6 Nice Post badges a day!
But the exponential dropoff in the higher levels is way sharper there - 8703, 545, 15.
I wouldn’t focus too much on Sitepoint’s, I believe the main cause for our lack of stats in this area is our community doesn’t like a lot of posts. I primarily see Staff and TL3’s liking posts than the rest of the community (not sure if our stats line up with that information, but I believe it to be true at this stage).
I’m not sure what to do about it though either. Trying to get people to use the feature has been an uphill battle for us. We still get several posts of “Thanks”, or “That was good”, or “I liked this article”. All of which fit nicely with simply liking the article or the original post it was in response to.
@HAWK, maybe we should have a brain-storming on how to get people to use the “like” feature. I’m not sure what more we can do to get them to use it…
I’d ban or at least change their realnames to something nasty if my users started giving me a hard time about badges or make new insulting ones to make up for the loss so I might be the wrong person to comment on this, but FWIW in my opinion the badge system needs to be either highly configurable (as it sort of is AFAIK?) or completely disabled by default.
How I see it, the trust levels in general need to be much more refined and should at least coexist with predefined, standalone user groups as an assistive technology to automate moderation.
I’d say that if we’re counting, why regard posts or topics that have not been liked at all? Well then we would have the problem with types of participation - some people prefer not to like posts especially if they’ve already replied to it and congratulated the person directly.
If that’s not confusing enough, then we might have categories of posts with different levels of importance; second hand category might benefit a lot from trust levels, while general chit chat does not, it would be completely disabled for “friday night drunk category” and very important with “dentistry pro tips” -category.
So back on topic, if the thresholds are a not up to date, they should be fixed and if the users are a problem, well, I’d say there’s no technology that can help you with that other than moderating with an iron fist…
edit: just logged in with uber admin and noticed that there are trust level groups and the possibility to add groups for static management etc. so
I wouldn’t think that the volume of likes would be related to the likes column being displayed. The number of likes a post of comparable quality across gets should be roughly proportional to how many people read it, and therefore how many people are on the forum to read it.
BBS has lots of participants who know about liking, enough so that they reach 10 likes on a post 6 times a day. Meta doesn’t have enough volume to hit 10 likes except on extremely rare occasions, and neither do several other forums (clearly, based on the numbers).
Maybe a solution would be to let the admin plug a number in for the badges…? Or perhaps a site setting for one of three sets of numbers.