« Hours of the Day (by Zone Type) | Main | On the internet everyone knows you're not actually an elf. »
September 02, 2006
Hours of the Day (by Weekday / Weekend)
We the split our data by period of the week - weekdays versus weekends. Note that our census bot performs the snapshot independently for Alliance and Horde, so the "average server population" listed in the graphs below is roughly half of the true server population.
There are several things about the resulting graph that surprised us. First, we were expecting a higher peak on the weekends. The data showed, however, that peak server populations on weekends was the same as that on weekdays. Second, the time of the peak was also the same. Instead, the difference was that server populations ramped up earlier and tapered off slower on the weekends.
Most of the zone types followed the overall pattern as shown in the graph above. The interesting exception was the high-end instance zones. In the graph below, we see that peak populations in these zones is higher on weekends than weekdays. The ramp-up also start a few hours earlier, although the tapering off is relatively similar.
Server Sample: RP (High), PvE (High), PvE (High), PvP (High), PvP (High)
Server Sample: RP (High), PvE (High), PvE (High), PvP (High), PvP (High)
Sampling Period: Month of January
Sampling Resolution: ~12 minutes
Parsing Method: The sample unit is each unique character in each hour of the day.
Data Filter: None
Sample Size: 179,003 characters
Posted at September 2, 2006 07:24 PM
Trackback Pings
TrackBack URL for this entry:
http://blogs.parc.com/cgi-bin/mt-tb.cgi/87
Comments
Hi guys,
This is a little off topic but; What software do you guys use to make your graphs and diagrams? It sure as hell aint MS Visio. That I can tell...
ed
Posted by: ed at September 13, 2006 07:44 AM
Hi Ed,
We use SPSS (a powerful statistical package) for Windows, v.13 (the graphs were much worse in previous versions, only use this one if graphing is what interests you the most).
Posted by: Nicolas Ducheneaut at September 13, 2006 11:46 AM
This is really great stuff. Have you been able to gather any data on choice of profession (smithing, enchanting, etc.), or do you have an idea of how one might go about doing that?
Posted by: Isaac Knowles at September 13, 2006 12:24 PM
Hello all,
And thanks a lot for the wonderful job. ^_^
I have a comment concerning the interpretation of peak population. You write : "First, we were expecting a higher peak on the weekends. The data showed, however, that peak server populations on weekends was the same as that on weekdays."
What if the peak was the same because of a ceiling effect? Could it be that the actual population maximum on a given server was 3000 (I don't know if this information is available elsewhere)? This would explain that, weekdays or weekends, the population never goes above 1500 (as you record half of it).
Since all servers the census was made on are high pop, could this be the case?
Sorry if that half formed idea is not very well explained, english is obviously not my native language (and brain is muddy too, need coffee). ;-)
Keep up the good work !
Best regards,
Skybird, mage 60 Kirin Tor-EU
Posted by: Skybird at October 13, 2006 06:13 AM
I think it would be interesting to restrict the data to players that are on for stretches of time. When do players that play for n={1,2,3,4,5,6} hours in a row play? How is that different at 60 versus before 60? You might excuse an occasional absence or two from the sequential data as survey error/alt swapping
Posted by: Sservis at November 29, 2006 05:23 PM