I've heard various ideas such as limiting it only after mods have expended their points (this will require implementing a cooldown to prevent a user from getting points again too soon). I want to hear your feedback, and I'll roll together something for the next major update of the site. Leave your comments
First, mod points now expire after eight hours. I'm willing to extend this to 12 or 16 hours after I'm sure comments will still reach +5 fairly regularly. With luck, we'll get to the point we can extend mod-points to last a full 24 hours which I suspect will end most of the complaining on them vanishing too soon.
I think mod points should ideally not be saved up for use on the moderator's favourite stories. Fast expiry doesn't eliminate that, but it does reduce it - which is good enough.
Any Soylent user should be capable of moderating comments on the majority of stories, so as soon as they get mod points they can dive right in and use up their points straight away. If they are too busy at that time, never mind, they'll get another chance in the future, and there are other people who can moderate today.
I know the expiry of points is tough psychologically, because it feels like you've wasted them. But actually the system can be (and I suspect already is) designed so that the overall amount of moderation happening is fairly constant, even on days when a higher proportion than normal of users failed to use their points.
So to people who complain about it, I say don't worry. I don't think anything bad happens if your mod points expire. The electrons are recycled :)
Also, I don't personally consider it of huge importance for comments to get to +5. I consider moderation a bit of a chore, so it would be easy just to go in with a +4 filter threshold and push most of them up to +5. But I try not to. I'm happy to leave a comment's score at +4 if it's good-but-not-that-good, even if it means I then have to open up more stories to finish my duty of using up those mod points. Should I reconsider?
dive right in and use up their points straight away
That advice may be misinterpreted by some people. For instance, I sometimes have mod points and open up a story, and read all the comments... and then don't make a single moderation. This is because I'm trying to avoid the 'early mod-up' effect. For instance, the first comments to a story may not end up being anywhere near the best comments once more time has passed. But moderators may feel the impulse to mod-up the best comment posted so far, even if it's not particularly insightful or useful. This leads to highly-rated comments that are lacklustre. (They are not necessarily bad, but perhaps not the best.)
In other words, I think the criteria for up-moderation should be 'this comment is genuinely good/useful', and not merely 'this is the best comment I've seen so far'. If a given comment thread ends up having no highly-rated comments, that's not necessarily wrong (maybe there were no worthy comments).
Modding in the way I suggest does mean that mod-points sometimes expire without being used. And it does mean that there are fewer highly-rated comments. But in principle it means that the visible comments are of higher quality. So I think it's worth it.
I agree with when you say "open up more stories to finish my duty of using up those mod points"... I think the general moderation task should be of the form: "Keep going back through stories until you find worthy comments". I.e.: don't just moderate the first few stories (that have few comments and few moderations), but instead keep going back in the post history looking for worthy comments until you run out of modpoints.
Yes, I think we are broadly in agreement about the approach to moderation, and that the wording I chose might be misinterpreted.
The other site allows mods to do this since it lasts for 3 days. Furthermore, once an article gone down a few spots, moderation and posting drop considerably, followed by mod counts.
You're doing the right thing by finding comments to moderate, but for the mod system to work, it really has to cater to the LCD, and the numbers don't suggest that most people do this. I need to write some code that dumps the moderator log, and gets exact numbers so I can rewrite the algo sanely.. Mod points were never intended to be as short shelflife as they are now, the system worked well at 24 hours right up until we got too many people not promptly using points. Until I rework the math with the new information in the comment DB, the mod algo breaks horridly if the expiration is too high.
Furthermore, once an article gone down a few spots, moderation and posting drop considerably, followed by mod counts.
Maybe we shouldn't get to choose which stories we are given the option of moderating? "You have moderator access. Would you be willing to moderate this story? [HeadlineOfSomeStory] Click here to view this story in moderator mode. ..."
I do like this idea, but I admit it's probably too much of a change from the status quo. People would object.
I need to write some code that dumps the moderator log, and gets exact numbers so I can rewrite the algo sanely.
Yes, well real data certainly trumps opinions - even my opinions :)
There's probably no such thing as a perfect moderation system - as much as we all may enjoy theorizing about one. So I guess the aim should be to spend as little time as possible getting a mod system that's good enough, based on the real world situation that the site finds itself in at the time. As such, maybe I have to rule out the idea that I've mentioned 4 paragraphs up.