I’ve added a new graph to the bottom of the Second Life statistics page. It’s median user-concurrency for Second Life charted over the last 365 days.
The tall spike in the middle looks like a bit of an anomaly. I’ll investigate that in the data. Would a longer period be useful? Are there variant graphs that people are interested in seeing?
This entry was posted
on Sunday, 30th August, 2009.
You can follow any responses to this entry through the RSS 2.0 feed.
You can leave a response, or trackback from your own site.
Possibly related posts
What is Second Life user-concurrency?, Second Life user-concurrency falls to lowest levels in two years, Second Life concurrency finally stabilising, Concurrency graph tweaks, Graph tweaks
Leave a Reply
Commenters are to be civil, courteous and respectful to others, insofar as it is possible to do so. Beyond that, you're not required to agree with the opinions expressed by me or by others.
Think for yourselves! First time commenters will wind-up in the moderation queue and your comment won't appear right away. Ditto for anything that gets flagged by the anti-spam rules.
what’s the resolution? and is the zigzag due to people going to bed?
Hm, btw, if it is due to night/day, perhaps an interesting graph would be the difference between night day over the same time range as this graph. Or perhaps if there is enough resolution in the data if that data is avaiable, the rate and/or difference of logins vs logouts.
Interesting graph. Thanks, Tateru! I’d be interested in seeing a 2-year graph, if you’ve got it. I’m curious if the “hill” here is seasonal, because of summer vacations or whatnot.
Tigro: The zig-zagging seems to be the weekly oscillation of users — more people on the weekends, for example.
The resolution is one day. One data-point = one day. 356 points, 365 days. A two year or a lifetime graph shouldn’t be an issue to produce.
Hm, ok, I still think a graph showing the difference (or perhaps ratio, I’m not sure which would be better) between the concurrency in one point and the previous point would be interesting.
Btw, would it be possible to have an applet, or perhaps a server side script, to allow people to adjust range and scale on the graphs to zoom and pan as they see fit?
Is it possible to have a connect-the-dots weighted average? It would be easier on my old eyes, at least.
I take my previous comment back.
Can you do something to indicate what the vertical graph lines mean?
I’m left to guess those represent weeks? If so, what part of the week? Saturdays? Sundays? Mondays? Wednesdays?
I note on the stats page that it’s daily concurrency median. Packing extra text into the graphic is a bit tricky
Here’s some variations in range:
Six months:

http://taterunino.net/statcharts/median_conc_by_day-182-400.png
Two years:

http://taterunino.net/statcharts/median_conc_by_day-730-400.png
All time (that is, from April 2006, the earliest I have data):

http://taterunino.net/statcharts/median_conc_by_day-0-400.png
Okay, after a little more processing, I’ve got a chart of weekly median user-concurrency for Second Life, from May 2006. It looks like this:

That looks like a watershed curve to me. What prompted that reversal at the top I wonder.Can you smooth out enough noise to get an actual date for the peak here?
I’ve investigated the point where median concurrency peaks on the graphs. It takes place on 15 March, at the height of the in-world protests about Linden Lab’s 12 March announcement about Adult Content. From that point on, median user concurrency (daily and weekly) begins to taper off significantly in the following weeks.
Is there any noticeable effect from the still-rather-theoretical bot ban?
It’s got to be mixed in there somewhere, probably right near the end. The actual drop starts well before that, however.
Tateru, why “right near the end”? Linden Lab announced on their blog that they were going to go after Bots a couple of weeks after March, 15. I’d say that the first bots managers stopped using them in March.
I guess both reasons play a role in the concurrency drop:
- non-bots accounts that were mostly used in what now are Adult zones, don’t log in anymore because they didn’t (yet) verify themselves
- bots concurrency is lower because of the Bots policy
Ah, I didn’t recall they were quite so close together. Thanks for pointing that out.
any predictions about when the line will start to go up again?
That depends. There’s usually a small decline in Summer, but this looks nothing like that. A lot depends on the actual cause. I’ve contacted Linden Lab about the figures earlier in the week, and hope to hear more soon.
You won’t be hearing anything about it other than giving you non-related data that’s showing positive growth.
Second Life had met its peak of growth, deem it as the mark of its imminent Second Death. I know not everyone want to hear “end of the world is coming” craziness. I don’t like it myself but I won’t ever deny it.
I knew this is coming because I noticed last year season was fairly poor than a year before that season, and repeat that step to find pattern of annal growth chart. However, the worst news is, it’s getting worst faster than on a month I predicted. I expected that peak to be met on November 20th – 30th. So that data tells me it’s getting worst on a curve scale.
Another fun fact about L$ currency, every L$ has a USD value. Exchange at LindeX had dropped from average selling L$perUSD at 265 down to 260 this year. It may not seem like much but using total L$ volume, Linden Lab lost more than 220,000$ USD alone this year. Right now, rate for 259′s quantity are still larger than 260 stock pile. As I recall last May of 2009, 260′s stock quantity was at average 40 millions. Today it’s at 200K volume. Volume quantity are shifting down to lower rate as possible. If Linden Lab are smart enough… Today is a good day to lock the exchange rate to avoid bigger sink hole.
Sad time for SL, but you gotta love shortsighted Lindens. :/
how about two new graphics comparing this year to last year, one for concurrency and one for whatever would be the most significant measurement of the economy. The graphs would be somthing like the difference between the values for that day and the day 365 days before.
Perhaps also a third graph for new accounts as well.
Actually, I’ve got no data that I’d consider a significant economic measurement. Nothing I could point to and say this shows that the economy is larger/smaller/stronger/weaker. Well, except maybe exchange rate.
@Nacon? Lower values for the L$ exchange rate are better than higher ones. As demand for L$ increases, the number goes down. 260:1 means the Linden Dollar is stronger than 265:1. The lower that number gets, the more people are paying (and therefore demonstrably willing to pay) for Linden Dollars.
Correct. It does not mean it’s a positive thing for SL because only asset and fee that does not change are estates and uploads. While people pay for their fee on time, which converted to that equal value for every US dollar. Linden Lab can not provide unlimited dollar resource for every L$ volume there is. 6,213,401,795 L$ total in volume today. When most L$ volume which have not been paid or exchanged had its value raised as further rate goes down. So Linden Lab have to hold accountable for every L$ value to stay legal. Adding more L$ when people pay up to Linden Lab with a buying rate. Having more people throw their money in does not cause that rate to change. Not paying on time causes L$ in demand, order to pay off estates/lands…. are causing the rate to change. However, there are people trying exchange marketing by buying lows and selling highs. Linden Lab need to prevent those kind of abuse to protect their meaning of accountable asset. Otherwise it’ll become another Ponzi schemes.
I’m not going to bother pointing out what causing people unable to pay off their estate/land on time, because I can’t be any more clear about “that” issue when most people and linden lab no longer care to believe any of it. Not going to be my problem anymore. It’s best to let people investigate themselves to face a good cold-hard reality. …but don’t act surprised once you do. Please, just don’t.
I think it’s time that I bury my avatar in its digital grave in SL.