WAR (Warhammer Online)

Originally posted by Azrile

Originally posted by templarga

It did not predict AOC. To think it did, is foolish and ignorant. Age of Conan was hurting long before X-fire predicted anything. To say so is inductive logic.

X-fire has not predicted anything! A prediction would be that this game would fail because X-fire says so. There is no predictive aspect to X-fire, only subscriptive. Learn the terms.

Second, I do not play WAR - I canceled in the first 30 days and to think I do, would only indicate I have a bias. I have no bias except I hate to see people make a mockery of statistics.

I played Age of Conan, WAR and other games. Right now I am in between games and could care less if a game fails or succeeds.
 

But by trying to show I have a bias, you clearly indicate that you have a bias and that is, you think anyone arguing against you are your precious X-fire are fanbois and could care less about the data we mention.

 

 follow my math

June 30th    WOW was known to have 4.5M players in NA and EU.    On Xfire, WOW was showing at 10 times the number of players as AOC.

Given that information, what would you guess the subs of AOC to be on June 30th?

If you do the math, you should have guessed 450k subscribers.

A few weeks later.. Funcom reported they had 420k subscribers on June 30th.

Xfire accurately predicted the  subscriptions before they were released by Funcom.

On Sept 17ish..   Warhammer ws showing about the same number as AOC was showing June30th. (and again 1/10 of WOW).  You would predict Warhammer to be about 450k right?   That is when they released the 500k number.

The math you do by using Xfire has ALWAYS been proven accurate within 10-20% any time a developer releases official number.  There are a few games that you can't use it on.  EQ2 and UO both basically block Xfire from being run, and EVE sort of 'promotes' xfire.   But games like LotRO, AOC, War and WOW have been almost perfectly predictive of any number that has been released.  There is no reason for an Xfire user to be biased towards or against one of those 4 games.

With the same mechanics I predicted 2 days ago, WAR must have an equivalent of about 350K subs at the moment. About 1/12 th of Wow population right now.

In fact - just like in AOC - I even said it was probably just a little less because older MMO's are not being played on a daily basis by their subs (see EVE) and -of course- you have that % of Macintosh players that are not counted for in XFire. So just like AoC had "about" 450 K, in reality it was at 420 K subs at that same moment.

It is all proven btw: the servers were full back then at the end of September (750K accounts), now everyone can see for themselves they are more empty with a lot of low/low servers.

So 350K is more than half of 750K (or the last famous 800K used by EA). So all the servers show this and STILL some people don't believe the predictions of Xfire.

How  much evidence do you guys want?

And for those programs that block XFire, the program doesn't work of course, but for all the rest it is working, because several thousand users are more than enough to have a "consistent !!!!!" sample.

You all forget the data have a consistancy level you can't find in a 'one shot' poll either.

So the ONLY exceptions are for the games blocking Xfire and in those games where the sample is just too small (I would say everything under 1 K XFire users),  But just look at the above examples and you see it all.

Originally posted by daeandor

Originally posted by LondonMagus

Originally posted by Abrahmm

Originally posted by daeandor

Xfire, Xfire, such a bunch of controversy it brings up here.

 

Here's the facts:

1.  Xfire currently has about 7.5k WAR players.

2.  7.5k is statistically significant for populations into the millions with a greater than 99% confidence level and less than 2% confidence interval.  See statistical significance and sample size calculator for easy understanding.

Take that information however you like.  The bottom line is that in evaluating trends in a game's usage, Xfire is statistically valid.  It's no different than the Nielsen Ratings for TV viewership, which is an extremely highly used statistical tool for the TV stations and cable providers.  Thier data is significantly more refined due to millions of dollars being poured into the study of their system, and about 80 years of Nielsen data, but the basic method is similar.

Excellent post. It amazes me that so many people are too stupid to understand this simple concept.

The size of  a sample is irrelevent in terms of calculating it's statistical relevance unless you can prove that it is representative of the population you are trying to analyse.

Xfire can be easily demonstrated to have heavy demographic biases not representative of the general gaming community, so it is not suitable  for this type of statistical prediction regardless of sample size.

If people working on the Nielson Ratings & Gallop Polls used such sloppy logic they would all be fired.

It amazes me that so many people are too stupid to understand simple statistical concepts.

Actually, that's part of my point.  If those are accepted standards, then why can't Xfire be considered representative sample?

Try reading what I said & thinking about what you are saying a few more times & you might get the point.

Contrary to what several Xfire devotees have tried to claim this is not about pretending Warhammer is actually doing really well in spite of what Xfire numbers predict.

I like the game & wish it were doing better, but it is still early days yet & all new launches seem to get roasted nowadays. I admit it has a few problems, but don't agree they are insurmountable.

This argument is about the correct way to manage statistics, not whether Xfire seems to vaguely point in the right direction at least some of the time. 

Anecdotal evidence about occasions when Xfire numbers have followed the real trend are irrelevant & do not magically make it a perfect barometer for predicting future activity. 

Whilst there are mountains of statistical texts to back up the arguments against the reliability of Xfire data, all it's advocates seem to have are blag & bluster.

Unless you are an expert in statistics, you are not qualified to say weather x-fire is valid or not. How random the population is, and how significant that randomness is, us stat noobs can only verify from historical data.

Originally posted by LondonMagus

Contrary to what several Xfire devotees have tried to claim this is not about pretending Warhammer is actually doing really well in spite of what Xfire numbers predict.

I like the game & wish it were doing better, but it is still early days yet & all new launches seem to get roasted nowadays. I admit it has a few problems, but don't agree they are insurmountable.

This argument is about the correct way to manage statistics, not whether Xfire seems to vaguely point in the right direction at least some of the time. 

Anecdotal evidence about occasions when Xfire numbers have followed the real trend are irrelevant & do not magically make it a perfect barometer for predicting future activity. 

Whilst there are mountains of statistical texts to back up the arguments against the reliability of Xfire data, all it's advocates seem to have are blag & bluster.

Read my post four posts earlier. This is ALL about stats and no denying this.
But you are ignoring always the same thing: your audiance in XFIRE IS already defined.

They are PC game players, on the internet, that INSTALLED Warhammer or AOC or any game you want to follow and it is a significant sample of a few thousand players (MORE than enough for the measurement), PLAYING games.

Meaning XFire - for ITS purpose - shows which games are being  played in the western world. This IS your sample.

What "other" elements do you want? ...Numbers of players.... Not what their hair color is, not what their religion is. Xfire is not there to predict the next president and HOW some social groups may view Obama.

XFire HAS already that predefined target group. The difference is you WANT to "protect" a certain game by posting every post above the last post and attacking a very good tool that actually protects players.

If it wasn't for Xfire, companies would still let you believe they had TONS of players who are no longer there on your servers. FunCon could STILL say they had zillions of players.

Be glad and don't be short sighted. The evidence is in front of your nose and for the sake of defending a game (you proved that in your last post), you sound much like that Iraqi fellow "no problems at all we are doing fine", "our servers are full", "we have 800 K players" and there is no problem.

Great. But reality is checking on your doorstep and each day you're becoming more and more greener in outfit.

I can understand your support for a particular game. I very well can understand you love it and play it. I can even understand you hate me for not liking certain aspects of your game, but please don't look like an Iraqi propaganda machine.

Originally posted by strategy

Originally posted by LondonMagus

Contrary to what several Xfire devotees have tried to claim this is not about pretending Warhammer is actually doing really well in spite of what Xfire numbers predict.

I like the game & wish it were doing better, but it is still early days yet & all new launches seem to get roasted nowadays. I admit it has a few problems, but don't agree they are insurmountable.

This argument is about the correct way to manage statistics, not whether Xfire seems to vaguely point in the right direction at least some of the time. 

Anecdotal evidence about occasions when Xfire numbers have followed the real trend are irrelevant & do not magically make it a perfect barometer for predicting future activity. 

Whilst there are mountains of statistical texts to back up the arguments against the reliability of Xfire data, all it's advocates seem to have are blag & bluster.

Read my post four posts earlier. This is ALL about stats and no denying this.
But you are ignoring always the same thing: your audiance in XFIRE IS already defined.

They are PC game players, on the internet, that INSTALLED Warhammer or AOC or any game you want to follow and it is a significant sample of a few thousand players (MORE than enough for the measurement), PLAYING games.

Meaning XFire - for ITS purpose - shows which games are being  played in the western world. This IS your sample.

What "other" elements do you want? ...Numbers of players.... Not what their hair color is, not what their religion is. Xfire is not there to predict the next president and HOW some social groups may view Obama.

XFire HAS already that predefined target group. The difference is you WANT to "protect" a certain game by posting every post above the last post and attacking a very good tool that actually protects players.

If it wasn't for Xfire, companies would still let you believe they had TONS of players who are no longer there on your servers. FunCon could STILL say they had zillions of players.

Be glad and don't be short sighted. The evidence is in front of your nose and for the sake of defending a game (you proved that in your last post), you sound much like that Iraqi fellow "no problems at all we are doing fine", "our servers are full", "we have 800 K players" and there is no problem.

Great. But reality is checking on your doorstep and each day you're becoming more and more greener in outfit.

I can understand your support for a particular game. I very well can understand you love it and play it. I can even understand you hate me for not liking certain aspects of your game, but please don't look like an Iraqi propaganda machine.

As I said blag & bluster, rather than answer the valid statistical points raised both on this thread & the other 'Xfire Controversy' one, you just attempt to discredit my argument by accusing me of purely trying to protect the game. 

I particularly liked the 'Iraqi propoganda machine' insult, most amusing but don't give up your day job.

Most of my comments on this subject have been purely in relation to statistical theory & whether I support the game is irelevant as that is how survey analysis works whether you like it or not.

As laughable as most of your insults are, your support of Xfire & evident dislike of Warhammer could be argued as to discredit you objectivity also.

The simple fact is though that Xfire is a biased sample for many reasons not just demographically & that its results are unreliable for the type of predictions that its fanatics constanty make. This doesn't mean that in hindsight it always has to be completly wrong, it just can't reliably predict the future.

I look forward to your next tirade of pointless abuse rather than statistical facts. 

Originally posted by LondonMagus

Actually, that's part of my point.  If those are accepted standards, then why can't Xfire be considered representative sample?

Try reading what I said & thinking about what you are saying a few more times & you might get the point.

Try not being an arrogant ass.  Go read up on the arguments against Neilsen and you will find them exactly the same as what people are talking about here.  The bottom line is that Neilsen is still considered acceptable by the industry.  Why should mmos be any different?