Concerning Websites showing Server Population Graphs

[GalacticKegger]

The conclusions just aren't true though.

 

You can't tell exact populations with that system, but you can still see over all population trends with it.

 

And from the trends extrapolate that into over all subs, at least in correlation, if not absolute terms.

 

Its almost impossible for subs to be going up long term IF server status is going down long term (unless an MMO company is constantly changing the server status levels), because that would be on par with inventing a perpetual motion machine (which I'm sure game companies would want to do in both cases if it were possible, of course ).

 

So saying that because you can't tell exact numbers therefore you can't tell anything is just patently false.

BW nearly doubled the server count on launch day, yet somehow those server pop sites never caught it. So if you are really that intent on being right (for whatever reason) - and want to bank everything on bunk info - then by all means go right on ahead. The rest of us will simply ignore the numbers and play the game. Besides I think some folks are missing the part about BioWare constantly changing the server status thresholds. Or they are simply refusing to see past their own ... agendas.

 

Great post OP. I'm a co-founding Technical Director & IEEE Computer Society member - I agree completely with your post. I tried pointing this out (though with less gusto I'm afraid) and my thread died from hater abuse. http://www.swtor.com/community/showthread.php?t=318353. Hopefully yours will fare much better.

 

/bump

Edited February 23, 2012 by GalacticKegger

[TUXs]

Whatever OP!

 

You and ur voodoo economics remind me of a phrase my father used to say to me...he'd say "Son, hokey religions and ancient weapons are no substitute for a good blaster at your side"...err...I meant..."Son, hokey rationals and ancient mathematics are no substitute for a chance to be a **** on the interweb"...while still paying a sub fee...

 

 

Nice post! Thank you for taking the time to detail why the sites mean nothing.

[Trenter]

Please show me anything that says they doubled server caps. They never doubled server caps and I have proof of it in Biowares own words.

 

Here:

 

I want to take a second to correct the reports which speculated that a lack of congestion was a sign of slow participation. Shortly after launch, we doubled the efficiency of each server, allowing us to handle twice as many players and remove the waiting lines.

 

http://files.shareholder.com/downloads/ERTS/1670728299x0x539144/d895626f-e715-4f00-bd50-31bf095027fd/Q3_Script_-_20_2112_1251pm.pdf

[TUXs]

Here:

 

 

http://files.shareholder.com/downloads/ERTS/1670728299x0x539144/d895626f-e715-4f00-bd50-31bf095027fd/Q3_Script_-_20_2112_1251pm.pdf

 

Dude...nice post lol! I troll so hard, and it's the 1st I've seen that

[Drewser]

The first 5 weeks are perfectly valid, in fact in this case they really help with determining any current trends.

 

Just because something is atypical doesn't mean it's invalid, in fact often just the opposite.

 

 

 

You can't just narrow a question (or data set) down to ignore things that don't fit as you'd like, that's terrible data collection, statistical practice and science!

 

Wow! My college stats professor, if he is still alive, would like to have a chat with you.

 

Do you know what outliers are?

[GalacticKegger]

Here:

 

 

http://files.shareholder.com/downloads/ERTS/1670728299x0x539144/d895626f-e715-4f00-bd50-31bf095027fd/Q3_Script_-_20_2112_1251pm.pdf

May as well put the whole transcript up for everyone's perusing pleasure ...

http://files.shareholder.com/downloads/ERTS/1709088884x0x540332/dab48832-e802-4c78-9fe8-91a3bd6180a2/ERTS_Q3FY12_Transcript.pdf

[Drewser]

Please show me anything that says they doubled server caps. They never doubled server caps and I have proof of it in Biowares own words.

 

 

I'll ask again, if I show you the quote from EA regarding this - will you believe it?

[Goretzu]

Yes, the first five weeks vividly display what happens when you double the server caps.

 

If they doubled server caps it would mess it up, but frankly that's just conjecture, and you'd still be able to see it, especially if you knew exactly when and how much they'd changed things (edit - in fact it was just after launch, allegedly, so that would still allow a good 10+ weeks).

 

In fact if you knew both when and how much you'd be able to account for it.

 

 

 

 

BW nearly doubled the server count on launch day, yet somehow those server pop sites never caught it. So if you are really that intent on being right (for whatever reason) - and want to bank it all on bunk info - then by all means go right on ahead. Besides I think some folks are missing the part about BioWare constantly changing the server status thresholds. Or they are simply refusing to see past their own ... agendas.

 

Great post OP. I'm a co-founding Technical Director & IEEE Computer Society member - I agree completely with your post. I tried pointing this out (though with less gusto I'm afraid) and my thread died from hater abuse. http://www.swtor.com/community/showthread.php?t=318353. Hopefully yours will fare much better.

 

/bump

 

You're a co-founding member of something or other, but you don't understand that doubling the server count wouldn't necessarily show on server pop data like this (queues don't show either)?

Edited February 23, 2012 by Goretzu

[GalacticKegger]

I'll ask again, if I show you the quote from EA regarding this - will you believe it?

He/she/it has been spamming liar accusations in at least 5 other recent threads. Frankly I'm surpised the account hasn't been suspended yet.

[Goretzu]

Wow! My college stats professor, if he is still alive, would like to have a chat with you.

 

Do you know what outliers are?

 

5 weeks of launch data isn't an "outlier", though.

 

Although there could be outliers within it, of course (but that's nothing like the same thing).

 

 

There are certain flaws with earlier data (server numbers and queues), but that's not the same thing, nor does it make it worthless in the context being discussed here.

[Drewser]

5 weeks of launch data isn't an "outlier", though.

 

Although there could be outliers within it, of course (but that's nothing like the same thing).

 

 

There are certain flaws with earlier data (server numbers and queues), but that's not the same thing, nor does it make it worthless in the context being discussed here.

 

Not could be, there are.

 

Much like a retail company analyzing sales, you can't just include the holiday period as a 'normal' month. Well, you could but you would be foolish to do so.

 

It 100% applies to the context being discussed here. The first weeks of operation are inflated/inaccurate due to the reasons I listed a few times.

 

If you want to do an accurate trend analysis, you need to wait until that period has passed and start the comparison from there. Could you go earlier than Jan 24, sure? I wouldn't.

 

Around that date is going to be a good starting point for a base since it is past the first renewal sub date which generally has a big drop-off.

Edited February 23, 2012 by Drewser

[Kthx]

If they doubled server caps it would mess it up, but frankly that's just conjecture, and you'd still be able to see it, especially if you knew exactly when and how much they'd changed things (edit - in fact it was just after launch, allegedly, so that would still allow a good 10+ weeks).

 

In fact if you knew both when and how much you'd be able to account for it.

 

I'm not sure how this is conjecture: "Shortly after launch, we doubled the efficiency of each server, allowing us to handle twice as many players and remove the waiting lines."

 

We don't know exactly when after launch or over what period of time the capacity was doubled and we do not know what subsequent adjustments were made. But we do know that adjustments have occurred and that the adjustments were significant.

 

Something else: The server load definitions are discrete (Light = 1, Standard = 2, Heavy = 3, Very Heavy = 4, Full = 5). We have no idea if the categories scale linearly. In other words, it is entirely possible that the difference in population between heavy and standard is not the same as the difference between very heavy and heavy. The only thing we know with some certainty is that at any given point in time, Full > Very Heavy > Heavy > Standard > Light.

Edited February 23, 2012 by Kthx

[Gleneagle]

They'd have to be constantly changing for you NOT to be able to draw anything from it.

 

Any single point change would still stand out and then everything from there would be comparable.

 

And constantly changing server cap limits would be both very strange and frankly rather iffy.

 

Not strange or iffy at all: instead think reasonably modern and cutting edge.

 

Bioware is paying high dollar for Dell Corp. dynamic virtual servers: they increase or decrease according to a rule, most likely keying to customer demand.

[Goretzu]

Not could be, there are.

 

Much like a retail company analyzing sales, you can't just include the holiday period as a 'normal' month. Well, you could but you would be foolish to do so.

 

It 100% applies to the context being discussed here. The first weeks of operation are inflated/inaccurate due to the reasons I listed a few times.

 

If you want to do an accurate trend analysis, you need to wait until that period has passed and start the comparison from there. Could you go earlier than Jan 24, sure? I wouldn't.

 

Around that date is going to be a good starting point for a base since it is past the first renewal sub date which generally has a big drop-off.

 

 

 

Who is looking for "normal", there's no such thing as "normal" in this context.

 

 

There's just relative server use and extrapolating that to a correlation with over all playing and subs.

 

 

Game launch and holidays of course affect that...... but then that is the whole point.

 

 

 

 

 

 

 

 

However at least you are accepting that these sites do show over all trends and are not utterly worthless and meaningless (as the OP suggests), you just disagree as to the start point for their validity.

[Drewser]

Who is looking for "normal", there's no such thing as "normal" in this context.

 

There's just relative server use and extrapolating that to a correlation with over all playing and subs.

 

Game launch and holidays of course affect that...... but then that is the whole point.

 

They are worthless given the time frames we have available at this time, which was my point.

 

But given time, some reasonable assumptions could be made from the information there unless Bioware plays games with the server capacities.

Edited February 23, 2012 by Drewser

[iceperson]

That's nice Op however log into any server and guess what you will see? Less numbers on both sides and every week that number drops.

 

this isn't what i've experienced

[Goretzu]

I'm not sure how this is conjecture: "Shortly after launch, we doubled the efficiency of each server, allowing us to handle twice as many players and remove the waiting lines."

 

We don't know exactly when after launch or over what period of time the capacity was doubled and we do not know what subsequent adjustments were made. But we do know that adjustments have occurred and that the adjustments were significant.

 

We know it's after that original post date (assuming it's valid and that the poster doesn't have a time machine! )

 

 

Something else: The server load definitions are discrete (Light = 1, Standard = 2, Heavy = 3, Very Heavy = 4, Full = 5). We have no idea if the categories scale linearly. In other words, it is entirely possible that the difference in population between heavy and standard is not the same as the difference between very heavy and heavy. The only thing we know with some certainty is that at any given point in time, Full > Very Heavy > Heavy > Standard > Light.

 

Yep, but that only effects accurate numbers (something only Bioware will ever have), not the over all trend.

 

 

 

 

Not strange or iffy at all: instead think reasonably modern and cutting edge.

 

Bioware is paying high dollar for Dell Corp. dynamic virtual servers: they increase or decrease according to a rule, most likely keying to customer demand.

 

Why would they continually raise and lower server population caps?

 

Raising them to a level they are sure won't crash, yes. Perhaps even sequentially so, but continually doing so why and how would they do that? And to make any massive difference they've have to be doubling caps and doubling again...... so within a few raises you'd have server caps in the 1,000,000's!

[kevlarto]

Very good read. Logic is the way to go. People that quote those sites have obviously not read the disclaimer

 

logic is something that is missing from these forums most of the time..LOL Yes Good read..

I feel the same way about reviews and other peoples blogs, speaking from thier point of view.. they do not offer facts..

[Kunda]

They'd have to be constantly changing for you NOT to be able to draw anything from it.

 

Any single point change would still stand out and then everything from there would be comparable.

 

I think the issue the OP was trying to illustrate was that a site, like torstatus for example, may be using a fixed numerical value for servers listed as light/medium/heavy for all servers, despite their population caps and overall population being different. A server listed as light with a population of 500 is not equivalent to a server listed as light with a population of 750.

 

You can't really develop a trend if you don't know whether population is falling or the population cap is rising.

[Vydor_HC]

This exact same debate arose in Rift which had the same monitoring sites. The fans of the game made every attempt in the world to disprove how the stats from these sites were showing a decline in population. However, looking back at the results of the past year, the steady decline in the numbers of players was tracked pretty well by the sites. Sure there was no definitive numbers, but the trends were very apparent and spot on.

 

Now I am not saying this is happening here and I'm not arguing the accuracy of the numbers, but for me this is deja vu all over again. So I wouldn't be so quick to dismiss what might be gleaned from that information.

Edited February 23, 2012 by Vydor_HC

[GalacticKegger]

I'm not sure how this is conjecture: "Shortly after launch, we doubled the efficiency of each server, allowing us to handle twice as many players and remove the waiting lines."

 

We don't know exactly when after launch or over what period of time the capacity was doubled and we do not know what subsequent adjustments were made. But we do know that adjustments have occurred and that the adjustments were significant.

 

Something else: The server load definitions are discrete (Light = 1, Standard = 2, Heavy = 3, Very Heavy = 4, Full = 5). We have no idea if the categories scale linearly. In other words, it is entirely possible that the difference in population between heavy and standard is not the same as the difference between very heavy and heavy. The only thing we know with some certainty is that at any given point in time, Full > Very Heavy > Heavy > Standard > Light.

Ya ... server status thresholds are not fixed by any means, and BW moves them around often. Their displayed values (not intrisic values) represent a relationship between fluctuating status thresholds and player count. Since we do not know the server status values, and we have no way to measure player count, the server population status site's tables & charts & plots & graphs are essentially based on a general equation that might look something like this:

 

[unknown variable ranges (Light ↔ Full threshold settings)] - [an unknown quantity (actual online population)] = floobie dust.

 

The constantly variable server status threshold settings won't even allow for accurate trend monitoring.

Edited February 23, 2012 by GalacticKegger

[Felizcat]

Couple of issues:

 

1) The numbers offered by mmojunkies and torstatus are the best info us users have, short of somebody moving from planet to planet and writing down numbers. We have to live with them and interpret them to the best of our abilities.

 

2) Bioware adapting the actual numbers of how many players online constitutes heavy load makes sense for them to fine-tune capacity on their end. If the actual server hardware was running at 30% capacity while their capacity indicator said "FULL" should certainly allow them to adapt. Or the other way around, their physical server running glowing red while the load indicator says STANDARD would need some adjustment, too.

 

But, it should be clear to them after a very short time what the real indicators are and become stable afterwards. Meaning, if they start HEAVY load at 1500 people online and adjust that up to 2000 after 2 days and again to 2400 after a week, it won't change much thereafter, unless better hardware is introduced. Which is unlikely during the first year of operation, EA wants to have some return-on-invest before they refresh.

 

3) We also face some competing measurements in regards to use and capacity: We players spend loads and loads of time in the first couple of weeks and now that number comes down. And that's what the server status (and a hand count) indicates: total playtime is coming down ( the sum of hours played over all players). What has not changed and may actually be slowly rising is the number of subscribers. I am sure there are plenty of people who played 60+ hours/week in December who now only play 20+ hours, and they are still subscribed. and they will be subscribed when they only play 10 or less hours per week. Bioware is perfectly fine with that.

 

But there are still a good number of newcomers who signed up yesterday, ready to put in their 60+ hours/week. We just don't notice them anymore, since they are spread out over 250+ servers.

[Drewser]

This exact same debate arose in Rift which had the same monitoring sites. The fans of the game made every attempt in the world to disprove how the stats from these sites were showing a decline in population. However, looking back at the results of the past year, the steady decline in the numbers of players was tracked pretty well by the sites. Sure there was no definitive numbers, but the trends were very apparent and spot on.

 

Now I am not saying this is happening here and I'm not arguing the accuracy of the numbers, but for me this is deja vu all over again. So I wouldn't be so quick to dismiss what might be gleaned from that information.

 

That sums it up. When there is six months to a year of data, it becomes more reliable.

 

The population could tank or it could stay exactly where it is today over the next year. Hell, it could even go up if Bioware fixes some things and releases some interesting content.

[Vydor_HC]

That sums it up. When there is six months to a year of data, it becomes more reliable.

 

The population could tank or it could stay exactly where it is today over the next year. Hell, it could even go up if Bioware fixes some things and releases some interesting content.

 

That's not the point I am making. What I said was after the data was collected, it meted out that people that were noticing a decline were right all along. The skeptics, who used the same arguments I'm seeing used here, were wrong.

 

I am not proclaiming as fact that the numbers here are right, but what I am saying is, don't be too hasty to dismiss them, as we have a good example to go by.

Edited February 23, 2012 by Vydor_HC

[Felizcat]

This exact same debate arose in Rift which had the same monitoring sites. The fans of the game made every attempt in the world to disprove how the stats from these sites were showing a decline in population. However, looking back at the results of the past year, the steady decline in the numbers of players was tracked pretty well by the sites. Sure there was no definitive numbers, but the trends were very apparent and spot on.

Not trying to pick on your post in particular, it just reinforces the point I want to make: "Server-Population" isn't very well defined. What does it mean in reality?

- Subscribers?

- People with toons on that server?

- Hours played?

 

I recently created a toon on "The Shadowlands" and was taken back how populated that server is. When asked why, I was told, the "Red vs Blue" team has their home there. That certainly will skew the number of people with toons on that server, just in case there will be an event.

 

And as I said in an earlier post, the total number of hours played on a single server is what should concern us as a community, looking for a group, raid or guild.