What is meta? ×
Meta Stack Exchange is where users like you discuss bugs, features, and support issues that affect the software powering all 149 Stack Exchange communities.

Often a higher rep means a user has been around longer. As a user gains more rep it's easier to implicitly gain even more from the trail they have left behind because of the abundance of it.

It might be interesting to show the usefulness (or "worth") of a user in addition to amount of points, through some other means.

For example old timer Joe does a decent job,

Joe Blow (1 year) @ 6000 rep.

100 up votes / 50 down votes = 2:1 ratio

but new user Bob provides more useful info

Bob Smith (4 months) @ 1000 rep.

10 up votes / 2 down votes = 5:1 ratio

This is simplified and doesn't calculate correctly by SO rep assignment rules, but you get the gist.

Personally I would love to have the ability to filter on perceived worth. It would be very useful to increase your chances at seeing more experience or worthwhile answers and such.

Note: I'm not being biased because I have low rep scores, because likely I have comparatively low ratios too. ;)

Update: There are other ways and combinations of metrics and events to determine a user's worth so focus on an ethereal worth metric that is theoretically perfect, and not on my simple example that obviously doesn't suffice to fit the bill.

Turned this into a feature request which obviously would need lots of discussion.

share|improve this question
An interesting idea. The challenge would be the algo for the worthiness metric. –  o.k.w Mar 12 '10 at 3:21
I think there's a germ of an idea there, it would just have to be fleshed out. –  Lance Roberts Mar 12 '10 at 5:24

3 Answers 3

See http://hewgill.com/~greg/stackoverflow/users/ for some statistics on the 10k+ rep users. This shows the accept rate (percentage of given answers that have been accepted), along with average upvotes and downvotes per answer. For each user this is also broken down by tag.

It would be possible to run this for other users, but I haven't taken the time to do so yet.

share|improve this answer
Nice........... –  John K Mar 12 '10 at 3:26
@jdk: The big thing to notice here is that the high(ish) rep users all generate many more upvotes than downvotes. Not no downvotes because even good technical opinions will find a few dissenters. –  dmckee Mar 12 '10 at 4:46

A lot of times people will give downvotes because they don't like the programming language someone used, or some ancillary comment, etc. This doesn't necessarily correlate with their usefulness, so I don't think that would be a good measure to use.

share|improve this answer
+1 - Especially on Meta! –  Adam Davis Mar 12 '10 at 1:54
Through time you would think that type of edge case behaviour would be averaged out in the proposed new metric. I mean the same problems apply to rep as it stands now. –  John K Mar 12 '10 at 1:56

Reputation isn't supposed to convey the average worth of a user's contributions, it's meant to measure their overall contribution to SO. Yes, those who consistently provide good content for a year will have a higher reputation, and are considered more 'trusted' by the site (allowing them to do more of the community functions).

Are you asking for a new metric to replace reputation, or to stand beside reputation?

share|improve this answer
To stand beside. –  John K Mar 12 '10 at 1:54
It also has the (IMO, undesirable) effect of discouraging new users from entering the fray. It doesn't take many instances of seeing a lesser answer by a higher rep user be accepted and upvoted above yours before most people lose interest in the reputation system. –  hemp May 25 '10 at 21:15
@hemp - "seeing a lesser answer by a higher rep user be accepted and upvoted above yours" - This is very subjective. But when you see this behavior, please point it out on meta here and have people evaluate the answers. Often we find that the higher voted answer is better in several ways, but when both answers are right, it's really subjective anyway. If you ever see a wrong answer upvoted above a right answer, definitely point that out. –  Adam Davis May 25 '10 at 23:49
What would you say is the proper etiquette for requesting a review of a marked answer? –  hemp May 27 '10 at 0:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .