Does Warcraft Realm Give Incorect Ratios?

Guest · Post by **Guest** » Sun Feb 27, 2005 10:21 am

So, The horde OUTnumber the alliance on Stormreaver US?

http://forums.worldofwarcraft.com/threa ... #post24180

according to your site, yes they do.

I have asked for over 1 month for some help with these uploads from 10pm to 2am.

No takers?

Post by **Rollie** » Sun Feb 27, 2005 12:15 pm

There could be more Horde characters, yes, but if you look at the Player Activity graph for Stormreaver, you'll see that on average (for the given data) there are usually more Alliance players online than Horde.

Guest · Post by **Guest** » Mon Feb 28, 2005 7:07 am

Rollie wrote:There could be more Horde characters, yes, but if you look at the Player Activity graph for Stormreaver, you'll see that on average (for the given data) there are usually more Alliance players online than Horde.

At no point during the day are there more horde online than alliance. Yet what is displayed:

Stormreaver
Alliance:9,091 (46%)
Horde: 10,625 (54%)
1 : 1.2
19,716

Are the largest A:H population imbalances occuring during the daytime hours? Is that an anomoly? Is that and the lack of data fouling the ratios even further?

For example:
8am 617 Alliance - 6 samples taken
8am 281 Horde - 5 samples taken
A:H Ratio = 2.2 : 1

1pm 1003 Alliance - 5 samples taken
1pm 631 Horde - 8 samples taken
A:H Ratio = 1.6 : 1

Then during 'peak' playtime (after school/work) the population starts to normalize with Alliance continuing to lead. But, lack of alliance data = many alliance players we have never counted in the census , yes?

8pm 1773 Alliance 1 sample taken
8pm 1415 Horde 7 samples taken
A:H Ratio = 1.3 : 1

For exampe players that play from the west coast that log in around 10:30pm and play to 1am have never been counted on alliance side. I think that is a significant number of characters.

Yet on the front page Horde is shown leading, not alliance....

Guest · Post by **Guest** » Mon Feb 28, 2005 8:43 am

Hmmm, you need to consider how these numbers are collected etc.

Now the person doing census on horde may be very diligent and does not delete his or her local character database that person will have a very large portion of the total number of characters resubmitted over and over even though they may be older than the 30 day window. If the person doing alliance census uses purge continually to only have fresh data then that side will only show fairly new characters. Many people collect local data for their own curiosity and then submit that data so not purging is of course in no way wrong, neither is purging. When Blizzard introduces the date() function Rollie will be able to give each character a datestamp that will hugely improve the quality of data collected but until then we will need to apply common sense to the numbers we see.

In your case the concurrent numbers are propably much more indicative of the true relation of accounts playing.

Other things to consider is that at times servers are down, this results in a large number of new characters created on other servers provided they are not already full and have long queues.

//ALanthus

Guest · Post by **Guest** » Mon Feb 28, 2005 8:50 am

Hmmm, also you have a number there with only 1 sample, just a few samples isn't really enough due to the huge difference between weekday and weekend numbers and the fact that sometimes the timestamp is just plain wrong since Blizzard isn't synchronizing time on the servers. (shouldn't be a problem on the EU servers though, content may be an issue but the company hosting there at least has the capability of doing professional hosting which among other things include synchronized system time.)

//Alanthus