>And the quality of posts? The post kill-rate (removal) actually dropped -hovering below 2%. This is less than half of the number incurred when registration was in place.
edit: the above refers to an automatic spam-killing filter, and nothing else. From what I saw, at least.
-------
Some numbers:
Behind registration: ~4.5%, out of an estimated 125 posts/day * 31 days ≈ 174 spam posts over 31 days = 5.6 spams per day
No registration: An anemic estimate of 1% of spam killed, out of an anemic estimate of 2,000 posts/day over the rest of the graph (2000 * (108-31) * 0.01) ≈ 1540 spam posts over 77 days = 20 spams per day
Removing registration lowers spam? Not really. Comparatively? Yes, because you have more non-spammers, and registration is not a barrier to entry for a spammer. But it doesn't reduce spam, and non-spam quantity is not equivalent to quality.
I'd be interested in a moderator-killed post number, weighted against the users/moderators ratio. I'd bet it skyrocketed - anonymity is fertile ground to flamebait.
edit: the above refers to an automatic spam-killing filter, and nothing else. From what I saw, at least.
-------
Some numbers:
Behind registration: ~4.5%, out of an estimated 125 posts/day * 31 days ≈ 174 spam posts over 31 days = 5.6 spams per day
No registration: An anemic estimate of 1% of spam killed, out of an anemic estimate of 2,000 posts/day over the rest of the graph (2000 * (108-31) * 0.01) ≈ 1540 spam posts over 77 days = 20 spams per day
Removing registration lowers spam? Not really. Comparatively? Yes, because you have more non-spammers, and registration is not a barrier to entry for a spammer. But it doesn't reduce spam, and non-spam quantity is not equivalent to quality.
I'd be interested in a moderator-killed post number, weighted against the users/moderators ratio. I'd bet it skyrocketed - anonymity is fertile ground to flamebait.