Change the reputation limit algorithm
Disclaimer: This idea would favour me personally. That's not my reason for proposing it, but it's no doubt the reason I started thinking about reputation limits in the first place. I believe it would be genuinely useful for the community though.
Summary: Limit reputation growth by question, not time. I suggest a maximum of 100 points per question through voting.
I believe that the current "200 reputation points per day through voting" limit is arbitrary and does not accomplish its goal. I'm assuming that the goal is to make reputation *reasonably* consistent in some form, so extremely popular answers don't skew things
massively. (To take an example from my own account, I've received 46 upvotes for expressing my view on whether to use < or <= in loop conditions. It's a considered view, but not really *that* useful!)
Surely a person posting 10 good answers a day is just as valuable to the site as a person posting 1 answer per day for 10 days. The
current model favours those who can answer at least a few questions every day - it's not terribly hard to build up 200 points in a day,
even without "working" particularly hard. I see no reason to favour that particular class of user - other than to get users back to the
site as regularly as possible. This "regularity" goal doesn't contribute to the quality of information as I see it - an occasional, hard-working user can contribute just as usefully as a regular, light-usage-per-day one.
However, limits are useful - see my example earlier. No one question/answer should skew someone's reputation out of proportion
to its usefulness. I suspect that once an answer has received more than 10 upvotes, more upvotes are not a particularly good indicator of linear usefulness. Badges are available for numbers above that,
so why not just cap a question's reputation that way? A maximum of 100 points per question, with no time limit.
we tested this and it makes almost no (literally!) difference to reputation. The daily rep cap is extremely powerful; we’d have to remove that before doing a per-post (question or answer) cap.
I'm all for doing both. Some time limiting is good, but limiting reputation by question limiting particularly important. There are some users who have earned 3k+ reputation based on just a single question. This really doesn't represent those of us that slog out the "30-70 views" questions very well at all.
The real problem is that you can still take negatives when you're at the daily cap. Take this question http://stackoverflow.com/questions/495722/help-me-sell-linq-training-to-my-boss/495754#495754. I posted this when I was at the cap and I knew it would be controversail. In the morning I saw I was down 10 rep because it got 5 downvotes... but it also got 20 upvotes. Silly system.
I agree that capping reputation by question is more sensible than capping reputation by time. In a similar idea I posted here, http://stackoverflow.uservoice.com/pages/general/suggestions/39746-idea-for-a-better-way-of-allocating-rep, I also think the value (ot cap) for a given question should be determined by the users rather than set at an arbitrary fixed value.
I agree, this is supposed to be bot prevention but it obviously isn't working too well if people are hitting this wall easily without being bots. I agree that time-scale is the wrong mechanism to use for this.
i asked the same here: http://stackoverflow.uservoice.com/pages/general/suggestions/66869-200-reputation-limit
You could raise the cap (to 400, say) but do something like a progressive tax where you earn fewer and fewer points for things as you near it. Make the veterans earn it.
Jon Skeet commented
It's a pretty annoying cap when you basically reach it by breakfast time and then get occasional downvotes during the day ;) (More seriously, I still think it's a very *artificial* cap - I don't see why time should be a relevant factor in reputation.)
Oh, and the 200 rep per day cap isn't a very firm cap. You can gain additional rep (at least between re-calc/normalization - they may adjust it then) by accepting an answer or having an answer accepted.
You could always do both.