Badge Discussion: Like Ratios + Spending Likes

Uh, this already exists? You get badges for liked posts.

For 1.6 we can do a repeat visitor badge as well as sequential days award, perhaps sequential days + reading count, to encourage regular visits.


Once people are promoted to trust level 4, they receive triple the amount of likes a basic user would, and wouldn’t it be just slightly unfair to those who are trust level 4 as they would have to like triple the amount of posts?

I’m a trust level four on the Hopscotch Forums,, and am very excited and curious about this!


I hit the allowance ceiling once from odd circumstances (catching up lost reading time over one week, in my edge case.) That and I think the default number of given likes allowed in a set period was lower at the time.

Getting a badge for that would make me smile. And when I needed it the most; *exhale* A “hit likes ceiling” badge would say “okay, stop clicking the hearts or you gonna keel from reverse operant conditioning headaches again.”

The Reader badge is a start. These user-contributed badge queries are worth noting and checking out. In your circumstances and needs, you may want to check out the offered extended Reader badge queries.


May it’s OT but I have it already. Not sequential (too strict) but visits + readings + posts. I used unlock badges and levels


RIght. I meant something like

1 Like

The thing about my class-based communities is that many of the normal measures like “been here a long time” don’t apply. I need more measures like “did some stuff this week”. I think I can work from examples like “30 likes accumulated over the span of 15 days”, have them set to get revoked daily and run a cron script to record who had what badges every Sunday at midnight to stick into the gradebook. Or something like that.

OOh. Some of those look helpful. Is there a way that I can get your SQL?

I was thinking earlier today that it would be cool if there were an API that would let people see/download/import badges from other sites. Permissions would be tricky, though. I suppose there could be an ok-to-export-without-admin-permissions checkbox or something.

I’ll go re-read the cool badge queries again. Thanks for the reminder. What I’d like to do is take all of those and make a more cohesive set of explanations, and then perhaps tie it in to my badge creator script.

1 Like

Naming is hard but I’m using the following names for now:

  1. Admired = have a high like ratio
  2. Generous = used up all your likes
  3. GivesBack = have a high liked / like ratio

Per my original comment, I believe the like mechanism is a golden hammer at best:

Any thoughts on measuring sentiment instead, or, as well as?

Machine Learning APIs: Text Analytics for Sentiment, Key Phrase Extraction, Language Detection and Topic Detection


Rather than counts or ratios, perhaps award badges based on positive contribution?

(This comment scored 82%)


@eviltrout mentioned some of these will need minimum thresholds to work, too.

1 Like

Okay I’ve implemented the badges in a way that I think makes sense. The ratios are based on me looking at records here on meta that I felt deserved the badges. I am more than willing to change them following discussion.

(Badges should be awarded overnight)


  • (count of posts with at least 1 like) / (total post count) > 0.75
  • requires a minimum of 5 posts

Gives Back

  • (likes given) / (total posts with at least 1 like) > 5.0


  • uses up their likes in a day



I would up the minimum on that first badge a lot. Maybe 100 to start. Perhaps you should require trust level 2 as well…

We need to seriously think about this because these badges can only be awarded once, and if they are awarded for trivial stuff, the badges will be meaningless.


Also we need to think through the progression from bronze to gold


Wouldn’t that mean 100% of your posts have a like? That seems a bit difficult to achieve…

Edit: Ohhh, nm, I see, you meant the minimum number of posts from 5 to 100. got it.

1 Like

Agreed, but can they instead be revoked weekly or monthly?

If you gave us the SQL we might run them against our own user bases.

Open Source FTW!

Some Numbers on my site:

  • Admired: 67 Members (0.44% of last 15 days active users)

  • Gives Back: 5 (0.03%)

  • Generous: 0

My point is the requirements could be shifted if we simply ensure the target audience for the badge starts with people who have achieved tl2 or even tl1.

It would be lame to have people get “admired” after their first two days on the site.

Or as @sam noted just have bronze, silver, and gold versions of these. (Well, not for reaching max likes in a day, the other two) also note the number of likes per day has increased several times, so people might have reached the like limit earlier, when it was lower…


@eviltrout, I feel that the Generous badge is biased against long term users. If a first day user joins and looks around a bit, they are very likely to reach there like limit. On the other hand, a TL4 user would need to like almost every post made on a moderately busy site to hit their limit.

For reference, what are the default rate limits for likes? I tried searching on Github, but I have no idea where to look.

Edit: Just saw @codinghorror’s post. If the goal is to hit TL2+ users, I’m not sure the generous badge is a good idea. It would encourage users who want the badge to just open every post they can and like everything.

That’s fine. Their name is “signed” on every post they like, and the likee gets notified as well. So it is a very public action. People aren’t that generous with likes because they know it is akin to cosigning a post, as in “I, Jane User, also support the words printed here.”

Generous is not any kind of risk for gaming IMO

1 Like

OK. Even if it isn’t at risk for gaming I still feel it is biased against those users with higher trust.