I know some members are keen on stats.
Did anyone check what the ‘before’ rate was for posts/day?
And did it go down with the pay system?
Or is it too soon to judge.
We can check via each post’s ID and timestamp (all times CST).
PID = 3000000, 02/15/03 05:43 PM
PID = 3250000, 04/07/03 03:00 PM
PID = 3500000, 06/01/03 00:22 AM
PID = 3750000, 07/24/03 08:31 PM
PID = 4000000, 09/18/03 07:15 AM
PID = 4250000, 11/19/03 11:12 AM
PID = 4500000, 01/29/04 08:34 PM
PID = 4670000, 03/22/04 07:37 AM (Close enough to the subscription activation time for me; someone more diligent than I can track down exactly which was the last post made before the switchover.)
PID = 4818863, 05/02/04 10:27 PM (ID and time of your OP)
Do a little math, and we get these numbers:
From 02/15/03 to 04/07/03: 4913 posts/day
From 04/07/03 to 06/01/03: 4596
From 06/01/03 to 07/24/03: 4643
From 07/24/03 to 09/18/03: 4509
From 09/18/03 to 11/19/03: 4022
From 11/19/03 to 01/29/04: 3502
From 01/29/04 to 03/22/04: 3241
From 03/22/04 to now (subscription period): 3577
The trend over the last year or so appears to be generally downward, and the time since subscription started was actually accompanied by an uptick. It’s too early to tell, however, whether this is a single-point anomaly or the start of a new trend.
This alexa graph shows a trend of going down in the last couple of weeks. This two-year graph puts it in perspective. (alexa only gets info from people who use the alexa toolbar so its def. biased information, but it gives an idea)
Earthling - That is so cool! I love it when you can get stats without having to use time travel to pick up data you didn’t think to collect earlier.
But it does confirm my suspicion that the posting rate went down ( 1/4 over a year) even when the delays went up.
alterego That chart tossed me for a bit, but it’s just non-intuitive, sort of upside down. The chart has high-ranking at the top, so a high posting rate would be at the bottom.
This just occurred to me…
From the membership data I posted in this thread, you can see that new registrations (red line) were declining through 2003 as well. However, the large spike in January comes from the slashdotted LOTR thread (so the rest of the year would have to be in decline compared to it), and there’s a general upward trend if we look back to 1999.
Seems to be a single-point anomaly, rather than a trend, because I wondered what the current rate was.
Last night at exactly 23 days since your last data point, we had post # 4897833.
That makes the current rate 3433 per day, or about 5% drop over the past month.
At that rate, the last post will appear on … well, I’m not that clear on the math, but Zeno’s Paradox kicks in and it could either be two years or infinity.
Guess what, your question prompted me to go look at posting data in greater detail and for a longer timespan; I think you’ll find this chart, which estimates daily post rates from May 1999 to the middle of this month, interesting. You’ll see that we are still on a general downward path that started in early 2003. However, it appears that there is a large-scale pattern of a spike followed by a long decline that spans about 15 months. If my observations hold water, you can interpret the chart as either meaning that we’re due for another spike soon, or that the increase from Jan. to Apr. this year was the new spike, and we’re now slowly sinking into oblivion. Of course, this may have as much accuracy as picking stocks purely from looking at charts – which is to say, not much – so time will tell.
(All this will be included in a much larger study that I promise to publish, but don’t hold your breath.)
Uh, I meant May 2000, which is what the chart shows. Sorry.
While I’m at it, I’ll add that even though the SDMB has existed on the web since 1999, the timestamps on posts made before April 28, 2000 are unreliable (i.e., a higher PostID does not always mean a later posting timestamp) and are thus discarded for this chart. I think 04/28/00 was when the board changed from UBB to vB, but since that was before my time, I’m not 100% sure about that.
You can probably trust those early timestamps for statistical purposes.
As I recall, they were caused by a different base hour setting, where the timestamps were considered Greenwich time instead of Chicago time or vice versa, and so the times are only off by about + or - 1/2 day.
No…it’s a lot worse than that. For example, PostID 130,000 is datestamped 2000-02-15, whereas PostID 135,000 shows 1999-11-02.
I think that “spikes” are caused by some outside news attracting attention, like the aforementioned Jan '03 slashdotted LOTR thread… So, it’s not like another one is “due”, unless someone finds some clever way of advertising the site somewhere that attracts lots of attention.
Very interesting info, though, thanks, y’all, I’ll attract the attention of the READER honchos to this.
Y’all have left out something that may have some bearing on the situation: A few months back, the amount of connections that we can make (our part of the pipeline) was restricted by the tech staff.
Demand on the system for the SDMB ate up all the access the Reader had for everything, including their in-house access and their other sites. Resources were reallocated and SDMB access was cut significantly as a result.
Think that might skew the numbers? We’re getting lots more complaints about people not being able to get in to the SDMB and overall system slowness, so it does not appear demand has decreased.
your humble TubaDiva
Administrator
Cool chart, Earthling. I’ve been curious about this myself. Thanks.
Interesting. Perhaps that explains the dip in activity the chart shows in Jan/Feb this year? Then the numbers climbed right back up, so maybe that would otherwise have been the spike this chart (supposedly) predicts? Then again, prognosticating is always a risky venture, so…
To follow up on Dex’s post, I added an indicator to the chart (update) to locate the “early 2003 spike.” Turns out it actually peaked in early December 2002, a full month before the LOTR thread was mentioned in slashdot. That thread was started in October 2002, though, and perhaps the increased traffic it generated was what caused slashdot to pay attention to it in the first place.
[highjack]I truly apologize for doing this…but, Earthling, I LOVE your signature line! Oh, and your mind. ;)[end highjack]