Last year a friend of mine (MMO friend only) demonstrated with as much precision as possible that XFire was an accurate correlator of intraMMO populations:  both in terms of those logged in and those subscribed. 

The methodology was simple:  A script pulled all XFire numbers over a period of 3 years for the top 20 MMOs that were out there.

Data was then assembled from separate posts that measured in-game populations from within the game. 

Data was also gathered from companies that trade publicly and released subscription numbers, or from where subscription numbers could be directly calculated.

Ratios were calculated.  Simple regression, least squares, was applied. 

The result?  R2 was .88. 

So you can hazard guesses until the cows come home as to why it is correlated.  But anyone with a statistics background who ALSO has a bit of common sense could likely figure out that SHOCKER...Xfire populations are a representative sample of a game's playerbase in so far as likelihood to play and subscribe is concerned.  It does not matter if you get more hardcore gamers, it just so happens, fortuitously, that they tend to quit games as much as your less hardcore non x-fire gamer does.

Or at least, good enough to sneak out a .88 R2.

Cheers,

RexTerra aka Loicenick aka Leaky the Leviathan aka Litigator

Because until those numbers of yours are actually revealed it is all hear say.

Without the numbers to verify the results X-fire will always be considered a non valid tool.

edit- frodo is spot on there.  There is too little to verify it with.

The mods said there can be only one and this isn't it.

Xfire numbers do not give any kind of good sampling regarding player population.  If that just happens to be your agenda. 

Xfire numbers do give a good sampling of player population.  If that just happens to be your agenda. 

If a game you like has good X fire numbers, its legit.  If a game you like has bad X fire numbers, it is not valid.  If a game you hate has good X fire numbers, it is not valid.  If a game you hate has bad X fire numbers, that is an excellent sample of its population. 

That seems to be what is done here.  Entirely. 

Originally posted by FrodoFragins
I think XFire has value.  But your argument is pointless.  Too many games don't reveal active player numbers.  You can't analyze non existent data and then claim a level of confidence in XFire.  XFire is good at getting an idea of how the active number of players for a game changes.  It gives ZERO info on subscribers who may pay but are taking a break.

Wrong.

Again, active subscriber data was available for a number of games:

a)  Funcom (AoC  and now the Secret World, where they gave very precise numbers)

b) Activision and EA games

b) Games with complete server parses provided by the company (Asherons Call, Dark Age of Camelot, EVE online)

c) Games with robust search functions to parse shard/server info: (Too numerous to mention)

Again the key is this:  if these numbers were so pie in the sky, then why did they correlate so well?  If they were not representative samples, why would the line of fit be so tight?  Magic 8-ball guessing?  At some point luck stops being an answer.

There is already a sticky for this in this very forum.

Please keep the discussion there.