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
Lizardbones is going a little further than trends - and that's OK and it is clever. It is dangerous and maybe the boat is being pushed out a little far but:
The assumption is that an XFire player is as likely to play SWTOR as EVE; I am OK with that. It won't be totally accurate because the games are not the same but that just goes into the error margin.
I haven't checked the exact numbers - which vary as we know depending on what day tou look at them but I'm going to assume it is a reasonable estimate. So - based on knowm EVE subs this gives an estimate of 4M - and we know that EA announced a maximum of 1.7M subs.
The big unknowm of course is the ratio of subs to active players. Lizardbones has dealt with a part of this looking at the average hours - again there will be a bit of varation due to days of the week etc. of course with an older game you may just get people who play on a Monday, some on a Tuesdayy however etc. We can't say though so ignore it: this is a peg in the ground.
Now as Eve, as an old game, may have some 'sleeper' accounts - folks who basically only play once a week for an hour or whatever. Even SWTOR will have some - people who subbed for 6 months but have now stopped playing. It is a possibility but again we don't know. A lot of people say that there are and there certainly it was the reason that Funcom gave when Microsoft ended collecting subs for AC years back and people had to actively resub. Ditto WAR in EU when EA took direct control numbers appeared to plummet. But we don'y jnow.
So if Eve is accurate etc. and the games are comparable then XFire would suggest 4M, in reality it is only 1.7M. A factor of 2.35 or so. Not brilliant if 100% correct but not way, way off. More interesting is the suggestion that SWTOR may be doing worse than XFire suggests.
I tend to play safe and assume it is doing better but they went with th etype of free trial with AR when the subs were under 300k. (300 k being what they announced so lower when they went with the trial).
Good job. We know there are roundings and errors and so forth but in a "is it in the ball park sense" fine.
Caveat: we are pushing th eboat out again when this is done ogf course.
I suspect you've missed the point of the 4 million players bit. I'm not saying that SWToR had 4 million players. I'm saying if you give any credence to XFire's numbers or XFire's trends, you end up with ridiculous results.
The response to this is that you can't take XFire numbers or trends at face value...you just have to 'see' the trends. Apparently this is supposed to happen without using any actual methods and especially without using any math. You're not supposed to relate XFire numbers to real world numbers, but at the same time you're supposed to make lots of estimates about real world values. Again, either without using any actual math, or if you do use some procedure to get that estimate, you sure as heck don't explain the procedure you're using. See that line representing XFire users playing SWToR? Well, just because that is a single line doesn't mean you can use the same methods of estimation at the beginning of the line as you would at the end of the line or the middle of the line for that matter. You're also supposed to remove the 5% of outliers in your head and that will somehow fix the numbers so they relate to reality. Which is something of a feat since we can't see the outliers. That, apparently, is how it's done.
As far as the method you've highlighted goes, we still end up with ridiculous values. It works in one spot if you make assumptions, but it breaks down when you move away from that spot in time. You end up having to change assumptions in other places to make it work. There is just no consistent or repeatable method to get XFire numbers to make sense. The numbers are nonsense.
I can not remember winning or losing a single debate on the internet.
The number of xfire-players is at 1426, which is an all time low.
You are probably looking at hours played, which I find less useful, primarily because more time played doesn't mean more players.
You are probably looking at hours played, which I find less useful, primarily because more time played doesn't mean more players.
If you're going to bother responding, at least make the effort to look at the same set of information I'm viewing. Metenso's chart, which he has posted every week for many weeks shows the number of players. The trend for SWToR in the past 4 weeks is up.
I can not remember winning or losing a single debate on the internet.
Hmm, so 1426 is greater than 2497...
I now do understand why we see things differently.
While lizardbones expertly demonstrates his lack of understanding of applied statistics and how firmly his head is stuck up SWTORS' behind, i am sure the OP of this thread is thoroughly thankful for the constant source of bumps that is lizardbones. It is great how we all contribute...
Anyway keep up the interesting work Metentos.
* Waves at Pushkina *
This is your post earlier in the thread.
If XFire is only measuring 'trends', why do you keep bringing up a single number*? I was under the impression that single numbers didn't matter. Make up your mind. Either XFire's job is to display trends and the individual numbers don't matter, or it measures trends and the individual numbers do matter.
If XFire only displays trends, then it has exactly 3 output values; "Less", "The Same" or "More". In that case, displaying the individual numbers makes no sense. The % of the rise or fall is irrelevant. The % of rise or fall in one game's numbers is not relevant to the % of rise or fall in another game's numbers. In this case, it is not possible that anyone has calculated or predicted anything using XFire's numbers because the numbers are known to be bad and should be ignored.
If XFire measures trends, then the individual numbers do matter, the percentage of rise and fall in the numbers does matter and it's wrong. These values cannot be reconciled with reality.
* The number isn't on Metenso's chart yet. When it's added to the chart with all the other numbers, I'll take a look at it.
I can not remember winning or losing a single debate on the internet.
1417 for Tuesday AND the game is ranked 14th overall... #15 is Aion.
Seriously? XFire peaked on XFire about 5 years ago, who the hell still uses it?
1417 SWTOR players yesterday.
An honest review of SW:TOR 6/10 (Danny Wojcicki)
guys it's not hard.
everybody does not use xfire.
simple solution.
look at other guys and use those as standards.
if swtor > other games = swtor doing good
if swtor < other games = swtor doing poorly.
this isn't difficult.
if swtor has 1,000,000 players yipie;
if on average every other game has 20,000,000 players....
what if on average other games have 20 players...
called standards. simple.
I did comment the xfire figures, those are raw data directly from the source we are discussing, no invention of mine.
If you need to resort to lying and pick text out of context to feel good, so be it. I'm done discussing with you.
You have certainly proven your point...
Correct. End of discussion. Xfire is a bad sample no matter how you slice it. Not random, no way to measure margin of error, nothing tangible whatsoever.
Well, back when it was ranked the forth most played game, did you then also think nobody used Xfire? When it's down to 14th spot, now you insinuate that "nobody" uses it? Poor excuses don't disguise the fact that TOR is a sinking ship.
The "4 millions" was brought up by YOU, don't you dare put me responsible for YOUR predicted figures, that's a lie. I quoted and commented on YOUR post, repeated below. Originally posted by lizardbones That would mean SWToR started with something like 4 million players.
And, as everyone that followed this thread has seen, I have never presented any calculations, predictions or absolute numbers about subscribers. You on the other hand...
I did comment the xfire figures, those are raw data directly from the source we are discussing, no invention of mine.
If you need to resort to lying and pick text out of context to feel good, so be it. I'm done discussing with you.
You have certainly proven your point...
The 4 millions was brought up by me. I'm reasonably sure that there's nobody in this thread who would attribute that to you. If what you posted was misleading, that's not my problem. I at least did you the courtesy of copying your entire post.
If you're going to use any individual XFire numbers as a reference point, then the numbers at some point have to make some sort of sense compared to reality. That includes using a percentage of change from one point in time to another. The numbers do not make any sort of sense and cannot be reconciled with reality with any degree of accuracy whatsoever. Using the numbers gets you crazy values like SWToR starting with 4 million subscribers.
An accurate reflect of what XFire says, would be a chart with one word on it. That word would be "Less", "No Change" or "More". You could even go so far as to have a chart of each week's players, and each week your data point would go up 1, go down 1 or not change position. That would display the trend with all the accuracy that XFire can bring. It would display the trend on a weekly basis without actually trying to measure it, which XFire cannot do.
I can not remember winning or losing a single debate on the internet.
I didn't think it was any more accurate then than I think it is now. It was nice to see the game doing well somewhere, but that didn't validate XFire's numbers.
I can not remember winning or losing a single debate on the internet.
You are either looking at some faulty data or are reading them wrong. The chart you are refering to is saying the same thing I am saying: http://i50.tinypic.com/2pzgu8p.png
The number of xfire players 4 weeks previous to the number for yesterday is 2624 (I might have missed by +/- 2 days).
A drop from 2624 xfire players to 1426 xfire players in 4 weeks is a extremely steep decent.
Lot's of people. Xfire has built in Twitter functions and Facebook Chat that can be used while playing. It also allows for live video streaming, video capture and screen shot captures that can be uploaded directly to your Twitter and or Facebook accounts. On average there are 2.5 million people on it during the day. It has 21 million registered users world wide. Just because you've never heard of it...doesn't mean people do not use it.
Also, XFMobile is one of the hottest selling apps on iTunes store. So there you go.
DOH! I was reading data points as weeks, not days. Which makes not a bit of sense.
>>There has been no upward trend in SWToR XFire numbers.<<
My bad.
** edit **
That is a steep descent. That only matters if the numbers themselves are valid. If the numbers can't be reconciled with real subscriptions (i.e. you can't take 1426 XFire users and get the number of subscribers with a measurable error rate) then the numbers themselves are bad. Which means the angle of the trendline is bad too - you can't get a 'steep' or 'shallow' descent without using the numbers. The direction of the line could be perfectly valid (up or down) - but the angle itself shouldn't be measured because there's no way to measure it accurately.
I can not remember winning or losing a single debate on the internet.
Lizardbones: In your above example about 4M I haven't missed the point.
If, based on the calcs (not checked but assume they are fine) I don't see an estimate of 4M as really, really bad.
If based on
- based on XFire players,
- data smoothing for days of the week
- assuming that EVE players and SWTOR platers are comparable
- assuming that an old subscriber game and a new subscriber game
can be compared then to be off by a factor of 2.35 or whatever it is is pretty good. Not perfect but a peg in the ground. If the example had come up with something really silly then it would be different.
If the ration is accurate - with whatever caveats you gave - then if there are now 15% of the max players then you would say:
15% of 1.7M is 255k but we "may" based on these EVE calcs be over stating and it could be 100k (roughly).
Now this stuff is really pushing the boat into deep water so we say: trends only - may suggest 255k but really messy, lots we don't know and with all these free trials the player number has shot up to 1500 (or whatever). Subs are going up!!! Or maybe not.
Calculating actual sub #'s from the XFire data is obviously not going to give us very accurate info and will involve quite a bit of educated guessing and application of similar patterns....
BUT... even if we get something as inaccurate as a +/- 50% margin of error... you can still see (along with all the other indicators we have) where TOR is headed.
And? if the game closed tomorrow you would be affected how? SWG would be relaunched, GW2 would be released early with honorable mention just for you, You'd finally finish that novel you've been working on?
Nope, you would all just go haunt one of the other games you dont play. Better yet what would you goofballs do if the numbers went up on XFirez .........oh noes, dont even say that.
I'm just interested in the trends of all of the MMOs listed. I'm sure it will see a rise in active players with the next big patch and the later patch when they add rated BG's.
The numbers for Eve have the same problems as the numbers for SWToR. The sample method is bad and not enough information is given to reconcile the numbers with reality. Each distinct set of numbers needs a new calculation or a new assumption to make them work. If Eve's XFire numbers fall 10%, their subscriptions don't fall 10%. IF Eve's XFire numbers rise 10%, their subscriptions don't rise 10%.
The only thing you add by trying to create a relationship between Eve's numbers and SWToR's numbers is more wrong numbers. Of course, it's not any more wrong than anything else here, but still, it's just wrong.
Skip numbers altogether. XFire is not a measurement tool. The numbers are "Rising" from week to week or "Falling" from week to week. Don't even use terms like "Steep" or "Shallow". Then you would be using XFire with all the accuracy it allows.
I can not remember winning or losing a single debate on the internet.
There is one statement that is pretty spot on for XFire. When XFire numbers trend down, so do the subscriptions of the game. If you don't take things any farther than that, then you're probably OK.
For instance, if you take the graph of SWToR, give the Y Axis a range of -100 to 100 and give the first X value a 0. Each week, if the XFire numbers go up, increase the X value by 1. Each week, if the XFire numbers go down, decrease the X value by 1. Then you get to see the overall trend. You eliminate all of the bad measurements because you're no longer measuring. You're just displaying the trends. Supposedly this is the important bit, but for some reason saying the trends are the only important part seems to be immediately followed by some assumptions that lead to a predicted subscription number.
I can not remember winning or losing a single debate on the internet.