First of all subs do nothing on game  gw1 is not sub  and few games are not sub and have updates and maint evry weak and more patch keep coming   whitough getting our money, we the comunitie of gamers are allowing company to steal our money for no reason   while  f2p game have the same routine as subs games  patch maint and uptades and expantion.

gw2 build the game for gamers  the way we really want it no subs ,a game that we can play and have fun a game e-sport whit competitive pvp   ,theres nor eason they gonna steal our money too for  ours to have fun  is simple to understand  -f2p game  bring more people   evrery year whit micro strans.aslong is not pay to win (wish is not the case whit gw2)

and the future is going to be f2p whit micro. adult build game for us (kids and adult) they know how economy is  is .. is easy for you kids  tahst your parent pay your sub and  then you come and talk sht abouth it  while company later made you pay for  expantion  evry few month  + subs lol shame..im grad gw2 will change alot of things .

(sorry for mi bad  spelling)

Originally posted by evilastro

Originally posted by komobo

Originally posted by evilastro

Haha nice work OP, I lol'ed.

As a scientist, I just laugh at people trying to use Xfire as an indicator of overall population.

and as a theoretical scientist, i just laugh at people who dismiss Xfire data as having no statistical relevance whatsoever.

Assuming that the users of Xfire, a completely optional and niche third party application, are representative of the total population of any game is silly, as a scientist you should know that.

If you had used a random population you can determine a correlation, but Xfire is not random, and the very nature of Xfire means that the data is biased and therefore irrelevant.

"It is biased in one respect, thus we cannot infer from it anything in any other respect", is basically what you are arguing. What you must understand is the fact, that bias in one respect does not imply bias in every respect. We know with 100 % certainty, that xfire users are indeed biased in terms of how many of them use xfire, exactly because they constitute the subset of pc gamers who use xfire. We know nothing else or whether this bias has any correlation to other aspects of xfire user's gaming behavior that may influence whatever unknown population parameter(s) you are trying to estimate.

To dismiss a biased sample purely on speculative grounds as to how the bias may manifest itself, is poor practice indeed. In science, we are seldom awarded the luxury of dealing with such niceties as 'completely randomized unbiased samples', which the basic predictive power of statistics is so dependent upon.
In an effort to answer such questions in a statistical sense, we require further sampling from different sources and then we perform the appropriate statistics i.e. multivariate statistics, correlation analysis etc. Now, you can argue that the xfire data is garbage since we don't have enough reputable information readily available and this is where i would have to agree with you, if we only had xfire data and for one game only.

However, i don’t believe that is entirely the case. Historically, Xfire data has beautifully reflected the population trends of many MMOs. These trends are often accompanied by 'anecdotal data' from other 3rd party sites measuring ie. server-load, game sales, activity on forums, activity in-game etc. confirming and in agreement with xfire data.
That is not to say that the points above proves the case for xfire in every respect, i certainly don't pretend that it does. And the degree of uncertainty is such that i'd avoid any attempt at a point estimate of ie. the number of swtor subscribers. As to the question of xfire's ability to reflect the general trends in population dynamics with some degree of accuracy, is in no way as easily dismissed as you make it out to be.

Originally posted by Kakkzooka

Originally posted by evilastro

Haha nice work OP, I lol'ed.

As a scientist, I just laugh at people trying to use Xfire as an indicator of overall population.

As a non-scientist, I'm shocked that you would view random sample data as a non-indicator of an overall population.

 

But what do I know? I'm just a lawyer. My girlfriend, however, is a doctor - and she disagrees with you.  She also has a license to perform autopsies on dead (so she says) people . . . legally.  And I have to sleep next to her. So, using that logic, I'm going to side with her.

 

No offense, of course.

If she ever gets a license to legally perform autopsies on live people... then I'd start to worry.