Howdy, Stranger!
It looks like you're new here. If you want to get involved, click one of these buttons!
Quick Links
Seems like the game has peaked on XFire - Part 2
This discussion has been closed.
It looks like you're new here. If you want to get involved, click one of these buttons!
Comments
Current Torstatus statistics show
US servers. After Combining. 5 servers total at current state.
1 almost dead Rp pvp
1 full pvp
1 full pve rp
2 full and 1 heavy pve
EU servers. After Combining. 4 Servers total at current state.
1 rp pve full
2 pve full
1 pvp full
Pacific Asia. After Combining. 1 Server total at current state.
1 pvx heavy.
It shows that SWTOR at current state could be ran with 10 server World Wide. Not only did they peak but the whole game can currently be ran with less server then they have just in the us atm.
I think your thread is a bit meaningless as the term server can mean so many things in an mmo. EA/Bioware upgrades the capacity and merged, so while yes some casual players left, as an actual player on one of these servers it has heald pretty steady numbers, through GW2 launch even (which I played). SWTOR looks solid at the moment, certainly EA/Bioware aren't panic at the present moment with bribes to players right now (which Blizz has done in WoW), and things are pretty steady with everyone looking forward to patch 1.4. That may hurt you to know, but SWTOR has a strong player base at the moment, and with free to play may re-attract the casual player base that left because they didn't like paying subs.
True, xFire displays trends as opposed to predicting them, but it does add new information, or at least independent information.
This is what makes the xfire numbers important as say opposed to the tor-status numbers. Both Tor-Status and, when it was still around, MMO-Junkies.net, are both tied to Bioware's own server load information which comes from here...http://www.swtor.com/server-status.
The official swtor server status information suffers from two main drawbacks. First, the data is quantized into light, medium, heavy, and full. This led to a lack of an ability to track player losses beneath light.
More importantly, those quantizations of full, heavy, medium, and light; are ranges, not set numbers, and are purposely undefined by Bioware. Futhermore, those ranges are not set, they are not permanent. They can and have been changed by Bioware, and sometimes the server caps are "rumored" to have been changed when xfire indicates they have not.
I don't if you remember, but MMO-Junkies.net, although it pulled its data from the same source as Tor-Status.net, displayed that parsed data in much greater detail, in realtime, and gave several more options (including an aggregate graphical representation of the data) then tor-status did. Back in February MMO-Junkies displayed what seemed likely an across the board manipulation of the Tor servers quantiztions so the same population displayed as being 20% higher then it was prior to the manipulation. The X-fire numbers bore out the observation. There was no sudden sharp decline in the nuber of players that Tor was losing. (The ease with which it was easy to see server threshhold manipulation is probably why MMO-Junkies.net was suddenly closed down. The site was set up by a third party hired by Bioware to compare and contrast Tor's numbers to WoW's numbers for the whole world to glory in Tor's success. When that scenario didn't play out, combined with the fact that MMO-Junkies made spotting server manipultion easy to spot, the site was uncermoniously pulled.)
When the servers were merged, there was a widley accepted unofficial rumor on the swtor forums that the destination servers had a much higher server cap then prior to mergers. (This was about the time that the mega-server technology thing was floated about, started by a brief blurb by Bioware in a few interviews.) Many claimed that the mergers weren't just a consolidation of the remainng playerbase, but the new super-servers actually represented an increase in player base as people were re-ubscribing to play o the new super servers. Well the xfire data, independent of Bioware's server reporting, proved that rumor to be patently false.
Also, althought there are much fewer servers today, those servers themselves can have their caps individually adjusted. The players self-polling surveys seem to indicate over and over that Bergen Colony, the last server that reported itself as heavy, had much lower population caps then the other servers. The smaller number of servers makes slight manipulations of the server caps a lot harder to spot on tor-status, but can still do nothing to alter the x-fire data.
In summary, x-fire adds something new to analyzing swtor's player numbers that otherwise unavailable. The xFire data offers an independent, unquantized sample set of which to spot trends in SWTOR's playerbase. X-fires games list is a vaulable check to itself as well. If there is an aberation in the way x-fire reports its player numbers, it affects and is reflected across a wide spectrum of games.
With a game such as swtor where its publisher uses fuzzy math to report subscribers on a good day (and a rage of 500k-1 million on a bad day.) having an independent set of data on swtor's playerbase is very useful. True, xfire users may not reflect an ideal sample of a game's playerbase, but xfire is being used to spot trends, not exact player counts. (As you pointed out, humans make predictions based on data samples) xfire is still the best there is, far superior to the dreck that nielsen attempts to sell to companies.
This was a lot longer then I thought it would be.
An example of the "increased server capacity" myth I mentioned in the wall of text in my last post.
This heavily rumored, unannounced server cap increase in not reflected in any way by the xfire data.
(And at this point, even tor-status shows a big dip in population since GW2 came out).
As to the rest of the Illysia's comments, it's probably better not to say anything if you hae nothing nice to say.
not even joking:
most interesting read I had on this forum
especially about mmo junkies...
can you mail me some info on them? having a hard time googling anything useful...
(I might want to write about that for another forum)
thanks in advance!
swtor pop has been stable the last few days... really surprising... still, it has lost 20% over the last few weeks
wow has dropped like a stone
gw2 now has 89.000h on xfire
SWTOR has 5.000h
gw2 has 18 times more time played
pictures (since I was stupid enough to exclude the name of the games)
1) list 2) WoW 3) SWTOR (looks good compared to WoW) 4) guild wars 2 (still climbing)
quote from early March:
GW2 now has: 81000h
WoW (which fell to 5th place): 22000h
SWTOR (17th place now): 5000h
-------------
WoW has lost 2/3 of time played
SWTOR 3/5
The main focus of this thread has always been about the number of people - for good reason as hours played can be a very poor indicator unless a game is "stable".
Once F2P was introduced things became "difficult" - although with the number of people playing SWTOR and using XFire still falling - the number went below 1,000 the other day - that suggests that EA are having a hard time promoting the current "unlimited free trial". Maybe there will be a bigger push when it goes F2P to level 50 later in the year.
And comparisons with other games are dangerous - they can be made but the number of unknowns is even more than when considering just a single game.
In the case of WoW for example a considerable number of people (1M+) are now annual WoW subscribers who got a copy of D3. So some D3 players will be WoW subscribers. Whether they will they stay WoW subscribers is another matter but they clearly decided the deal (essentially a $7.50 a month sub to WoW if they were planning to buy D3) was worthwhile.
GW2 will have attracted players from many games. The issue is where they will go when they have had "enough for now" - for now because they may go back to GW2 when their is an xpac. Mayb eto WoW if they have an annual pass. Maybe to the next F2P offering. And that is churn. And that is what SWTOR has toovercome. People playing through the stories and then moving on.
And at that point the XFire number will essentially reflect EA's marketing efforts. There will - presumably - be a dedicated hardcore and above that F2P people who try the game and then leave. Will the number drop even further? It is already lower than Heroes of Newerth for example.
True, xFire displays trends as opposed to predicting them, but it does add new information, or at least independent information.
This is what makes the xfire numbers important as say opposed to the tor-status numbers. Both Tor-Status and, when it was still around, MMO-Junkies.net, are both tied to Bioware's own server load information which comes from here...http://www.swtor.com/server-status.
The official swtor server status information suffers from two main drawbacks. First, the data is quantized into light, medium, heavy, and full. This led to a lack of an ability to track player losses beneath light.
More importantly, those quantizations of full, heavy, medium, and light; are ranges, not set numbers, and are purposely undefined by Bioware. Futhermore, those ranges are not set, they are not permanent. They can and have been changed by Bioware, and sometimes the server caps are "rumored" to have been changed when xfire indicates they have not.
I don't if you remember, but MMO-Junkies.net, although it pulled its data from the same source as Tor-Status.net, displayed that parsed data in much greater detail, in realtime, and gave several more options (including an aggregate graphical representation of the data) then tor-status did. Back in February MMO-Junkies displayed what seemed likely an across the board manipulation of the Tor servers quantiztions so the same population displayed as being 20% higher then it was prior to the manipulation. The X-fire numbers bore out the observation. There was no sudden sharp decline in the nuber of players that Tor was losing. (The ease with which it was easy to see server threshhold manipulation is probably why MMO-Junkies.net was suddenly closed down. The site was set up by a third party hired by Bioware to compare and contrast Tor's numbers to WoW's numbers for the whole world to glory in Tor's success. When that scenario didn't play out, combined with the fact that MMO-Junkies made spotting server manipultion easy to spot, the site was uncermoniously pulled.)
When the servers were merged, there was a widley accepted unofficial rumor on the swtor forums that the destination servers had a much higher server cap then prior to mergers. (This was about the time that the mega-server technology thing was floated about, started by a brief blurb by Bioware in a few interviews.) Many claimed that the mergers weren't just a consolidation of the remainng playerbase, but the new super-servers actually represented an increase in player base as people were re-ubscribing to play o the new super servers. Well the xfire data, independent of Bioware's server reporting, proved that rumor to be patently false.
Also, althought there are much fewer servers today, those servers themselves can have their caps individually adjusted. The players self-polling surveys seem to indicate over and over that Bergen Colony, the last server that reported itself as heavy, had much lower population caps then the other servers. The smaller number of servers makes slight manipulations of the server caps a lot harder to spot on tor-status, but can still do nothing to alter the x-fire data.
In summary, x-fire adds something new to analyzing swtor's player numbers that otherwise unavailable. The xFire data offers an independent, unquantized sample set of which to spot trends in SWTOR's playerbase. X-fires games list is a vaulable check to itself as well. If there is an aberation in the way x-fire reports its player numbers, it affects and is reflected across a wide spectrum of games.
With a game such as swtor where its publisher uses fuzzy math to report subscribers on a good day (and a rage of 500k-1 million on a bad day.) having an independent set of data on swtor's playerbase is very useful. True, xfire users may not reflect an ideal sample of a game's playerbase, but xfire is being used to spot trends, not exact player counts. (As you pointed out, humans make predictions based on data samples) xfire is still the best there is, far superior to the dreck that nielsen attempts to sell to companies.
This was a lot longer then I thought it would be.
XFire provides more numbers than server status, this is certainly true. Now, for part two, how accurate are the numbers XFire is providing? To even know how accurate the numbers are, we need to know what kind of impact the overall population has on the numbers, when they are reported. Take a guess how many times anyone has done that in this thread. You also need to know the actual population numbers. At the very least you'd need the subscription numbers, but even that is questionable since XFire is reporting players and played hours, not subscribers.
What is my point here? Hang on a second. Ok, I'm back. The idea is that XFire is providing more numbers, making it more accurate. Except there is no accuracy associated with XFire. You have "More" players or "Less" players. It's a trend line. The trend line would be going up or down with the Server Stats as well. If XFire had any sort of accuracy associated with it, you could see if the trend was steeper or shallower...but there's no association between the angle of the line generated by XFire and the actual player population. 1 less player is the same as 1,000 fewer players and that's about as close as you're going to get. XFire adds lots of facts, but not any new information.
Regarding Nielson ratings, if XFire's numbers that we see were worth anything, they wouldn't be given away for free. XFire is selling the information that we cannot see. If the numbers XFire was reporting were worth anything to anyone, they'd be selling them.
I can not remember winning or losing a single debate on the internet.
Well, we agree xfire shows trend. We disagree as to the value of the extrapolations we can derive from those trends. I think they are rather telling when you combine the xfire data with outside information.
Tor:Status is not worthless, but as I said, can be manipulated. Also, because its data is sourced from individual servers instead of a total pool, you get a messy graph such as this:
http://www.torstatus.net/shards/us/trends
*edit* Since I wrote this torstatus expuged itself of all the "origin" servers, deceptively reducing the impact of my point.*
(That above link could be a thread unto itself)
Many gamers like to belittle x-fire's users as not being represanative of either their game, or all game's playerbases. I would counter that it is not perfect, but it is the best thing out there. 20 million accounts and 100-200 thousand daily concurrent users is nothing to snneeze at. (And it seems to be pretty accurate, trendwise at least.)
Investors seem to think so too. Xfire has been bought and sold several times. The latest itteration of xfire is backed by $50 million dollars of venture capital to seperate itself from Titan gaming.
I would contend that xfire's data is above and beyond the worth of the Nielson ratings. The xfire data may seem crude and simplistic to us, but that is only because xfire releases anough data to prove that they have a robust playerbase. The same goes for VGchartz. They have simple lists of game sales, but if you want the parsed, analyzed, deeply detailed statistics; you have to pay for the analytical tools and monetized data.
As to your statement, "If the numbers XFire was reporting were worth anything to anyone, they'd be selling them.", well xfire most certainly does.
http://www.xfire.com/privacy_policy/
"Aggregate Information and Non-Identifying Information. We may share aggregated information that includes your Personal Information, Non-Identifying Information and Log Data with third parties for industry analysis and demographic profiling and to deliver targeted advertising about other products and services."
Xfire has several revenue streams. Among them are advertising, pais promotions through their communites program, but the most hush-hush and lucrative is selling the data mined from their highly desired market demographic.
There's a reason why investors throw money at X-Fire, and its not to better the gaming world by providing free quality of life features to gamers.
There is one valid criticism of x-fire that is rarely spoken about. (And you can see why xfire would never mention it).
Once you start to get into the 15+ titles down the list, the sample population is still quite a valid sample, BUT...
When you start to get into 1,000 players playing, the sample becomes small enough that even a little bit of manipulation can greatly alter the results of x-fire's sample data.
It would be easy for a company with even modest resources to increase their position on the chart substantially by running say a hundred or so x-fire accounts idiling during business hours.
(I'm not saying this happens, but it appears easy to do.)
Of course, x-fire may have in house procedures to prevent this (IP address checking, noticing a hundred accounts log on and log off everyday from the same geographical area) or at least notice and later discount this data.
Of course you could also go the complete cynical route and ask what is prevent a gaming company to not just pay xfire directly to bump their title higher on their lists. (There isn't really any sort of agencey regulating or reviewing x-fire's listed data results)
I just wanted to add this as potential valid criticisms of x-fire's data results besides the usual "a self-selected sample is not a valid random sample" that is usually the main criticism of xfire.
I'll agree with this, and maybe add a little more. Some games are just so far out of mainstream anyways that the users are possibly lucky they have a game installed, let alone other support software. Games targeting a very young population would fall into this category - especially if they have parental locks preventing loading new software + frequent formatting by their parents to clean up spyware from time to time.
Then there is the crowd who just mess around with a few games, but neither they nor anyone else considers them as gamers, so they never heard of XFire and don't read forums like this. I'm thinking of a lot of those ranked games 1000+ that you mentioned for these exceptions, after reviewing the titles.
Also, on the cynical route, it could happen, in theory (with a botnet?). It is largely known to happen with "forcing" a video on YouTube to go viral and appear on the front page. A company in California, US, for instance, uses a team of workers from all over to simply watch videos to up their view count, post them all over social networks, and then post comments to start engaging or even controversial and / or trolling conversation to keep users returning.
I don't support this site, but for educational purposes if anyone was interested:
http://www.sbwire.com/press-releases/buy-youtube-views/sbwire-154338.htm
It's really just one of many companies that do this for marketing. A quote on the site:
“What makes or breaks a video on YouTube is the number of views, rates and subscribers it has. No matter how interesting or ground-breaking a video is, it will not go viral without a little help, and our services are aimed at doing just that: helping businesses get up to 1 million views in the shortest possible time, taking all the hard work and investment out of having a video go viral”, stated the spokesperson of Buy YouTube views.""
There are software packages out there that claim to turn your computer into a youtube bot to spam-watch your own videos, but YouTube removed that capability (possibly because of widespread abuse) and considers geographical region now when deciding if a video is truly viral. There is a limit 200 or 300 something views where the counter stops and geographical information has to be verified before the count is authorized.
A main hitch to XFire vs Youtube is that YouTube views are permanent, a big jump in views will make a video go viral to get even more views, serving the purpose of getting the video to be a big hit for a couple days or so, then disappearing. But, continually maintaining a high XFire rating would be very time intensive .. and very expensive (over long periods especially), unless there is a social network that can make high ranked XFire games also go "viral" like a video. I suppose it comes down to how well XFire scutinizes incoming data sources.
Just jumping back to the viral videos on YouTube .. ever wonder how magically SWTOR had a day 1 video release (the slow motion lightsaber fight in a crowded city) that was instantly the #1 video on YouTube? Makes you wonder :-)
Want a nice understanding of life? Try Spirit Science: "The Human History"
http://www.youtube.com/watch?v=U8NNHmV3QPw&feature=plcp
Recognize the voice? Yep sounds like Penny Arcade's Extra Credits.
You are so right, Xfire numbers are at best useful to measure huge dips and ups for the same game, everything else is doubtful at best. When a game goes up or down 50% for more than a day something have likely happened but that is the most it is useful for.
On the other hand is EAs decision to make the game F2P rather speaking. They would not do so after 8 months without good reasons.
Certainly true to question the accuracy of the data but to suggest that it doesn't matter whether 1 less player less is the same as 1,000 fewer players is disingenious. The issue, as you say, was knowing how acurate the estimate was. There is accuracy but is it 1% or 50%? And all the info that we did get kept suggesting "not that bad".
As for Neilson vs. XFire I see that as B2P vs. F2P. XFire being funded by advertising. Not that Neilson tracks players outside the US of course and both EA and Activision have publically said the data is inaccurate.
And now that people can play SWTOR without being subscribers any attempt to estimate the number of subs is pretty meaningless as you would have to estimate how many players were playing on a trial.
At the end of the day the XFire data suggested a very steep decline from c. 2nd January - at a time when people and analysts were talking about SWTOR adding subs hand over fist. And that has come to pass.
Absolutely true about things getting ever harder once the numbers decline. The margins of error must increase.
Also a stable player population will react differently to the initial first flush of buyers and will change again when extended trials are introduced.
And over an extended period of time what happens with XFire will also become an issue.
The self-selection bit - doesn't matter; it isn't (see older posts) but even if it was that would simply increase the margin of error.
On the other hand is EAs decision to make the game F2P rather speaking. They would not do so after 8 months without good reasons.
Investors would not use XFire to determine whether or not to invest in EA or one of EA's games. They would use the decision to take SWToR F2P to make that decision. One set of information is useful, the other isn't.
Also, I don't think I really proved my point with my post. I think what I said was right, but it wasn't a very good response to the post I was responding to.
I can not remember winning or losing a single debate on the internet.
As for Neilson vs. XFire I see that as B2P vs. F2P. XFire being funded by advertising. Not that Neilson tracks players outside the US of course and both EA and Activision have publically said the data is inaccurate.
And now that people can play SWTOR without being subscribers any attempt to estimate the number of subs is pretty meaningless as you would have to estimate how many players were playing on a trial.
At the end of the day the XFire data suggested a very steep decline from c. 2nd January - at a time when people and analysts were talking about SWTOR adding subs hand over fist. And that has come to pass.
It's not disingenuous if it's true. There is no difference between 1 less XFire player in a game than 1,000 less XFire players in a game. Both mean "Fewer Players" and that's as close as it gets. The only reason we can allow even that level of relation between the two sets of data is because there's a historical correlation between declining XFire numbers and declining game players. It's an observation, with no margin of error because it can't be calculated. All that's been shown is "Less XFire Players" = "Less Game Players" and even that doesn't always hold true.
Nielson probably shouldn't even be brought up in these discussions. It doesn't have anything to do with XFire.
For your last paragraph...that's what XFire can tell us. Are there "Less" players or "More" players? There were indeed "Less" players. If there is no defined ratio between the number of XFire players the number of actual players, then the angle of the trend line is just as undefined though. If you replaced the information presented by XFire with numbers that differed by 1 every week, you'd get the same level of accuracy.
For instance:
Week 1: 1,000 XFire Users means ?? Game Players
Week 2: 870 XFire Users means ?? Game Players
Week 3: 630 XFire Users means ?? Game Players
Week 4: 220 XFire Users means ?? Game Players
Week 5: 100 XFire Users means ?? Game Players
and
Week 1: 10 XFire Users means ?? Game Players
Week 2: 9 XFire Users means ?? Game Players
Week 3: 8 XFire Users means ?? Game Players
Week 4: 7 XFire Users means ?? Game Players
Week 5: 6 XFire Users means ?? Game Players
give the same level of information because there's no ratio of XFire users to game players. A 50% drop in XFire users does not mean a 50% drop in game users, it just (probably) means there's a drop in game users.
The predictions of decline and doom would be meaningful if they weren't so common. Every game gets the same prediction. It's bound to eventually hit home.
I can not remember winning or losing a single debate on the internet.
If we had the EA data we would be able to say exactly what the difference between 1 and 1,000 was. Just because we don't doesn't make them the same. By that logic we may as well abandon counting because 1 is the same as 1,000 and so the same as 1,000,000. Hence many = fewer!
From the outset this thread was never about "doom and gloom". Indeed it rebuffed the usual "doom and gloom" approach by refusing to focus on hours played - always high in the first few daysbefore falling back resulting in "the game is failing" talk.
And what the tracking of the XFire player numbers suggested very, very quickly was that SWTOR was tanking; heading for something approaching oblivion. It was not behaving like WAR, or AoC or Rift etc. Those and other games, I agree with you, see a big drop of maybe 50% after the 30-days is over but then the number levels off. Hence WAR still had 300k subs 6 months out. And in the next 6 months you get a further albeit slower decline.
For SWTOR to have done as well as EA would have had to announce subs of 900k+ at their last conference call.
Estimates were made from the XFire numbers - more than once. With margins of error and the assumption that XFire was understating the numbers - so add on a few hundred thousand because "surely things can't be this bad". And they were.
SWTOR has not experience a normal mmo drop off. It has behaved like a single player game with 80 or 150 or whatever hours of gameplay. And when it is done it is done.
The only thing that has held true with XFire is a correlation between a drop in XFire players and a drop in actual players. The predictions and estimates did not come with a margin of error. A margin of error is a calculated value and you can't calculate the margin of error when all you have is one set of data and nothing to compare it to.
People did estimate from XFire numbers, and they were wrong. They were off by hundreds of thousands of players. To be fair, they couldn't possibly be right unless they were really good at guessing. XFire doesn't show subscribers, it shows players. The ratio of subscribers to players is unknown, as well as the ratio of XFire players to actual players. The ratio between XFire players and actual players changes at unknown rates. There are too many unknowns for anyone to make a good estimate of players or subscribers in any game. Unless, of course, the developer releases subscribers and player numbers.
If we allow that one instance of something different happening with SWToR validates XFire, we've achieved the following:
Instead of having 1 degree of difference between values in a range like "Less", "No Change" and "More", now we can have 2 degrees of difference between values in a range like "A Lot Less", "Less", "No Change", "More" and "A Lot More".
To be honest, I'm fine with XFire itself, but people here want it to mean something that it doesn't mean. They keep comparing it to political polls or scientific studies. They keep throwing out words like "random sampling", "self selected sample" and they keep talking about the size of the sample being taken. XFire is worse than not knowing because it just increases the amount of misinformation.
Here's some stuff that doesn't happen with XFire.
Nobody compares Weeks 1 through 8 of an MMORPG's life with Weeks 1 through 8 of another MMORPG's life using XFire's numbers.
Nobody compares the XFire population numbers to the game population numbers, showing how the differences in the overall population affect the game's populations.
Nobody notes how all the game's XFire numbers rise and fall together, based on XFire's overall population instead of the game's populations.
I doubt anyone will really be able to tell me about the problems with looking at the above three things and getting something meaningful out of it. It's probably too much to expect, but the clue is that XFire doesn't tell you enough about the numbers they are showing you for the numbers to be meaningful.
People keep referencing XFire's numbers, and they're not even doing a decent job of looking at the numbers themselves. Never mind the horrible job being done relating XFire's numbers to actual players or subscribers.
I can not remember winning or losing a single debate on the internet.
SWTOR now under Aion
barely above EVE online (a "niche" game, according to the SWTOR forums in January ^^)
not in top 20 any more
SWTOR has lost 60% of its players in the last 4 weeks on xfire
yes, I know... xfire players hate swtor, but buy it still, then leave the game to make it look bad...
I would think they leave the game because they don't want to play any longer. Leaving the game to make it look bad seems like kind of a reach.
I can not remember winning or losing a single debate on the internet.
It was obvious sarcasm ...
That's what happens when a competitor releases a game with a smaller budget and 10x the depth of your own game. Oh yeah and it doesn't charge a sub. It also happens when you lay off a lot of your devs and set the rest to getting the game ready for F2P.
I'm very curious what the response to SWTOR would have been if it had been B2P.