30,000 Members strong - AAC tidbits

strantor

Joined Oct 3, 2010
6,782
I thought you were comparing your post #1 to the wikipedia article, which references several observations going back for years. Then you likened (something? assumed your post #1?) to someone who was awarded a PHD for stating the obvious and well documented. If you weren't the OP I would think you were dissing the OP.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,979
I use the option to keep spammers from knowing their safe. It works, I've caught quite a few in the middle of a run. Satisfying, that.
So, as a mod, what can you do to the spammers you catch?

As a spammer, what would I see if I got caught this way? The only thing I can think of is that I would see a message telling me some variant of my username not existing or my username being banned or that I don't have posting privileges.

While I can understand the satisfaction that can go along with knowing that that will happen, it just seems like there should be something more... tangible, involved. ;)
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,979
I thought you were comparing your post #1 to the wikipedia article, which references several observations going back for years. Then you likened (something? assumed your post #1?) to someone who was awarded a PHD for stating the obvious and well documented. If you weren't the OP I would think you were dissing the OP.
Oh, I see what you are saying.

No, I won't diss an OP from observing and commenting on the obvious (usually). I even pointed out in my original post that none of this was particularly unexpected (though perhaps how well the fall off appears to match an exponential distribution is of interest).

But I would deserve to be dissed if I went out and got a bunch of funding to come up with and publish these results. At the very least, I should be expected to come up with something new or confirm/refute something that is unsettled.

So I was likening the study to the PhD -- two things that purport to expand the human knowledge base in a meaningful way.

For instance, taking the exponential falloff that appears to exist here. If I thought that such information was useful to those that model the internet, for whatever reason, I would (should?) first be expected to perform an extensive literature search to see if such distributions had been observed and published previously. Depending on what I find, I may or may not have something worth pursuing. If theoretical models predict an exponential distribution but there don't seem to be an corroboration emperical data, then perhaps I can fill in that gap. If most of the theoretical models predict some other kind of distribution, then I have the potential to shake the tree a little bit if I can show that most real world situations are actually different. If I can find theoretical and emprirical results for things other than forums, then I have the opportunity to show that forums should be added to that group. If I can find empirical results but no theoretical analyses, then I can focus on filling in the analytical model. But if both the theoretical and emprical results are addressed in the literature, even if just in a single paper, then it becomes much harder for me to justify that my research is adding to the general knowledge base in a meaningful way. If I really want to pursue it, then I have to make sure that my work delves down to deeper issues than any of the prior work.

Now, it is possible (unlikely, but possible) that the 2005 study referenced represents the first time something like this was published in the literature. If so, then if it pointed that out and also pointed out how much of this is common knowledge that has just previously been disseminated by word-of-mouth, then I could be talked into buying off on the claim that it is contributing to the general body of knowledge -- at least as far as being something worth published, though probably not as anything close to deserving a PhD, which this study may or may not have been a part of.
 

loosewire

Joined Apr 25, 2008
1,686
I have watched the new member start ups for some time,each time you look

there is a new member signing up. This has been going on for years and your

numbers don't count off-topic,where guys are not racing to the top.
 

JoeJester

Joined Apr 26, 2005
4,390
WBahn,

Of the 30013 members listed in the membership list, 8184 (27%) has visited since 7/1/2011. I would classify the remaining as "lost interest".

In the last 90 days 3017 (10%) members visited the site.
 

Thread Starter

WBahn

Joined Mar 31, 2012
29,979
WBahn,

Of the 30013 members listed in the membership list, 8184 (27%) has visited since 7/1/2011. I would classify the remaining as "lost interest".

In the last 90 days 3017 (10%) members visited the site.
In general, I agree. There are certainly some people (who will have a considerable number of posts) that are come-and-go types who may indeed go for more than a year and then come back for a while and then go away again. I am like that on a couple of forums, one of which I used to be a moderator on. But the total number of such people is going to be small compared to the 8000+ you identified.

It is probably a very good guess that a very large fraction of the 5000+ members that have visited in the last year but not in the last 90 days also have lost interest. The number of people that, for whatever reason, have just taken a 90 day break is, of course, going to be larger than the 1 year come-and-goers, but I would still be surprised if more than a few percent of them ever post again.

Contaminating the data some what are the 121 InvisiMode guys, whose last visit is masked. Since this is a more advanced setting, these people are more likely to be (or at least have been) fairly active participants. In fact, 25 of them have 100+ posts and 11 are 1000+ posters. Not surprising given that apparently moderators like to use that mode in order to monitor spammers (and other things, I'm sure) more effectively.

On or after
01APR12: 3000
01MAY12: 2233
01JUN12: 1364
26JUN12: 586 (last seven days)
01JUL12: 291 (just two days)

One thing that would be interesting to do, and might be marginally novel, would be to plot the distribution of time-between-most-recent-two-posts and plot that, perhaps, against both number of posts and length of time as a member. But I think you've have to write some data mining app to gather the data. But something like that might be a worthwhile project assignment for an undergraduate course in web programming or data mining.

One thing that messes up the late
 

Georacer

Joined Nov 25, 2009
5,182
So, as a mod, what can you do to the spammers you catch?

As a spammer, what would I see if I got caught this way? The only thing I can think of is that I would see a message telling me some variant of my username not existing or my username being banned or that I don't have posting privileges.

While I can understand the satisfaction that can go along with knowing that that will happen, it just seems like there should be something more... tangible, involved. ;)
When you identify a spammer, there is a button in every post that calls the forum built in "decimator". It opens a pop up that allows you to select among multiple "delete" options. We usually select all of them.

When the member tries to login, he gets a message that states that he cannot do so, as he is banned for such and such reason.
 

Wendy

Joined Mar 24, 2008
23,415
Of course, they come back many times. But after a long registration process, we still have that little button. That can't be fun.
 

JoeJester

Joined Apr 26, 2005
4,390
WBahn,

The one small site that I extracted the membership data from their membership listing, I came up with the following catagories:

Registered only or 1 and done. This class of member was ordered to register in an attempt to get them into the conversation and when that failed, the next crop of people were ordered to register and post.

Lost Interest: Those whose average post count is less than one and hadn't posted 90 days.

Lurkers: Averaged less than 1 post in the last 90 days.

Casual posters: Averaged less than the overall average postings in the last 90 days.

Active posters: Averaged more than the overall average but less than the average plus twice the deviation in the last 90 days.

Hyperactive posters: Averaged more than the average plus twice the deviation in the last 90 days.

Even though this site has 100 times the members of the site I catagorized, I would suspect the numbers to be similiar ... with the obvious exception that no one can be "ordered" to register or post here.
 

Attachments

Thread Starter

WBahn

Joined Mar 31, 2012
29,979
Hmmm. Interesting.

I think I see some significant differences (for instance, our ona-and-done would appear to be higher compared to the other groups, even allowing for the different rules), but I'm not sure how much those differences are primarily reflective of the different policies.

There are a lot of interesing things to explore and think about, though almost all of them are absolutely meaningless and useless except to a small group of people (spammers and advertizers and ISP/network folks) and, for the people that do care about this stuff, I can't help but wonder how much information is read into such data that isn't really there.
 
Top