Concerning Websites showing Server Population Graphs

[Kaelshi]

I've been seeing people reference websites like torstatus and mmo junkies when attempting to argue server population points.

 

I thought I would make an independent post - without an argument on whether pop is actually rising or falling, in an attempt to explain them.

 

I have been associated with websites in the past that monitored server population(not currently with any of the ones I'll be mentioning and I am compelled to not post the ones I worked with in the past)

 

I am actually an economist by trade with a background in accounting - many of these websites consult with us so they can better understand the information they receive as well as to better educate those who visit their websites.

 

I will try to be as simple as possible.

 

Here is a link to MMO Junkise Server Status website

http://www.mmojunkies.net/index.phpapp=statistics&module=servers&section=list&game=swtor

 

It appears to show a drop in population. The reality is you can not determine actual population numbers from this chart. They explain why at the bottom of the page

 

Note: Values in y-axis DOES NOT represent player count but instead aggregated server-status. For more information about interpreting them and explanation on how they're acquired, please visit our FAQ page.

 

What This Means - the y axis of the chart is the axis that is on the left and rising - on their charts it rises in increments of 100(so 100...200..300 etc.). This is NOT a count of the number of people logged in to SWTOR. this is a numerical representation of the aggregate server status. This means MMO Junkie has assigned a number value to each server status(example light = 1, medium = 2, heavy = 3) They track what each status of each server is, combined them, and come up with a total number. They then present this info on a graph, over time.

 

I will not go into the faq sections of each website which deal with this but I will explain, hopefully succinctly, why we can not rely on this information alone to determine server population.

 

Stephen Reid, a senior community manager had this to say about server population back in December:

 

"We’re monitoring all of our servers around the clock, and we’re raising server population caps where required. However, to ensure our long-term goals of server stability and healthy populations, we do not want to raise population caps too rapidly. We want incoming players to populate lower population servers. At the same time, higher population servers will not be ‘locked’ because we want to allow people to join a server to play with their friends if they wish to.”

 

The key to this is the raising of server limits. There are currently several threads on this forum that have people discussing where the population cut-off for very light/light/medium/heavy/very heavy are. It seems many people have seen the cut-off for the same status at different numbers - so one person says there was 500 ppl on and it was medium while another says there was 500 but it was still light on their server etc.

 

This is important when it comes to the graphs for this reason:

 

Server A - at launch

light 0-500

medium 501-1000

heavy 1001+

 

It is a popular server with long queues so Bioware determines it needs to increase the server cap. We don't know how it effects it at this point but the assumption is that the above chart would then change to look something like this:

 

light 0-750

medium 751-1250

heavy 1251+

 

With the addition of new players lets say the population levels out around 1100 players at peak. The MMO Junkie chart in December would have shown that as heavy, assinging the number value of 3 to that server. Now it would show as medium, assigning a number value of 2. This would mean the chart for this server would drop by a fairly large amount(since if at off-peak they have say 5-600 people, beforei t would have been medium-2, and now it's light -1)

 

The chart drop would actually correspond in this scenario with a population RISE.

 

Conclusion

 

We do not know what sub #/population # are right now. There are to many variables. Without a doubt some servers will increase in population and some will drop.

 

I hope I was able to demonstrate though that using these particular tools to chart actual population numbers would be incorrect and can in fact lead to wrong information.

 

The only way they could be valid as a defenitive tool for population growth/decline is if we knew for a fact that the very light/light/medium/heavy/very heavy numbers in fact never changed regardless of population cap changes. We also don't know if Bioware can/does individually change those numbers for their on internal purposes. What if they lowered the cap on other servers for instance, then the graph would go up but population might be dropping? We just don't know based on the information these charts give us.

[zrinking]

Great thread

[Varnoc]

Your use of logic and fact is a breath of fresh air.

 

Sadly it won't stop people from misrepresenting those sites to reinforce their faulty logic.

 

I say let those that want to hate on the forums do so, since their $60 purchase and $15 a month go towards helping to expand this game.

Edited February 23, 2012 by Varnoc

[Alavastre]

Indeed, maybe the sky is not falling?

[Alavastre]

Wow, I'm surprised a thread like this would be buried, it's a very good read.

[Orduno]

No one really cares about anything that can't be fit into an angry tweet about this game.

 

Although I'm surprised we haven't seen "obvious BW fanboy lol" commentary yet.

[Ebbikenezer]

Yeah, good read, pretty insightful to say the least.

 

I'm not one to trust very many independent websites for statistical info anyways, good or bad.

[TheOnlyKyrenS]

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

[Goretzu]

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.

Edited February 23, 2012 by Goretzu

[muzzle_flash]

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 is 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 (I'm sure game companies would want to do both, of course).

 

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

 

+1 , agreed

[uidLucid]

As a developer for 2 mmos, one that was a success for a few years, one that a company tanked ( Thanks for that Rapid Reality... never let a gaming company be ran by devout Christians that want to make your female characters resemble 14 year old boys to avoid them being "Too Sexy". ), I can attest to one simple fact.

 

Unless the developers release to public knowledge the exact rounded ( Up or down, you decide. ) concurrent UNIQUE connections per day, every, single, solitary, statistic site is unequivocally.... you guessed it; wrong.

[Alavastre]

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.

 

Not really, you can only tell that if you know the population caps are steady and they don't change. Otherwise, you have no clue, population could be going up or down and you wouldn't be able to tell either way.

[Goretzu]

Not really, you can only tell that if you know the population caps are steady and they don't change. Otherwise, you have no clue, population could be going up or down and you wouldn't be able to tell either way.

 

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.

Edited February 23, 2012 by Goretzu

[Drewser]

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.

 

Maybe long term trends but a short trend coming out of holiday makes the SWTOR graphs useless at this point in terms of estimating subscriber base from them.

 

Then factor in server load changes and it makes them even more useless.

[Urael]

Great Post OP!

 

People that have an ax to grind will always use "statistics" as they see fit to justify the grinding of the ax. Logic does not work with people of this ilk. They have an adgenda to support soley for their own ends.

 

All claims of one extreem or another on the interwebs should always be taken with causion.

 

Again, OP thank you for the time you put into your post. I hope it does not get burried. And stays at the top of the boards to refute all the extreem posts forcasting gloom and doom.

[Emeda]

Well written but still have to ask.

 

you quote "We’re monitoring all of our servers around the clock, and we’re raising server population caps where required. However, to ensure our long-term goals of server stability and healthy populations, we do not want to raise population caps too rapidly. We want incoming players to populate lower population servers. At the same time, higher population servers will not be ‘locked’ because we want to allow people to join a server to play with their friends if they wish to.” but not where they say some servers are almost at their absloute maximum and when that is reached it will not be increased.

 

All those full servers that they listed did not have their server caps raised by much. They added new servers instead. So why are those servers that they raised the cap to absolute maximum that they said they would not increase over that now no longer full? The only reason is because people are not playing on them anymore.

 

Did Warhammer have a site like this? I would love to see their graphs and see how they relate to the SWTOR ones. I would believe they would be very similiar.

 

All this thread is is a well written way of saying dont listen to those sites because they dont show what I want them too.

 

Why would you work on one of those sites and then still deal with them if they are misleading?

[Palenen]

Imperial: I had the idea of flying to each planet and getting those pop levels, ship, then to fleet. See what ya have. Then by a gross figure, perhaps double that to include Republic or rather if you have a Republic character, doing the same. Of course it wont be correct but you can see it visually what the planets are at. Honestly, I have not done this. It will grossly only show current connections but the planets I am seeing, they rarely have 50 or more people on them.

[Goretzu]

Maybe long term trends but a short trend coming out of holiday makes the SWTOR graphs useless at this point in terms of estimating subscriber base from them.

 

Then factor in server load changes and it makes them even more useless.

 

 

Yup you can't see much outside of months (or several weeks at best), but there is months of data now.

 

Variability of server load change is fairly irrelevant in this context, if it wasn't much of the data the world relies on to function simply couldn't be used. We've been dealing perfectly well with cyclical data for centuries now.

[Alavastre]

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.

 

You're correct in that, but since I'm not working for BW, nor do I know how easy or how constantly something like that would change, I cannot make that assumption. Especially two months out. Now a year from now, I would say you're right. It's possible they will tweak the numbers constantly, or maybe not. Again, it's hard to tell.

 

But yeah, hard trends you should be able to tell. But from what I see, there is nothing significant with the trends atm. And I personally don't see a difference, if anything there seems to be more people on my server. Just my experience though.

[Lethality]

The only way they could be valid as a defenitive tool for population growth/decline is if we knew for a fact that the very light/light/medium/heavy/very heavy numbers in fact never changed regardless of population cap changes. We also don't know if Bioware can/does individually change those numbers for their on internal purposes. What if they lowered the cap on other servers for instance, then the graph would go up but population might be dropping? We just don't know based on the information these charts give us.

 

You know of Occam's Razor, I presume? The idea that the simplest explanation is probably the correct one.

 

We could explain a million reasons around why those charts are wrong... but that, too is speculation and unlikely speculation.

 

I doubt BioWare every changed the threshold for what "heavy" or any other designation meant, even when adjusting the server caps. It would have been disingenuous and if they did it to hide population drops, well, that's another story all together. Safe to say that didn't happen.

 

The simplest explanation here is that there are less people playing less often than before.

[Drewser]

Yup you can't see much outside of months (or several weeks at best), but there is months of data now.

 

Variability of server load change is fairly irrelevant in this context, if it wasn't much of the data the world relies on to function simply couldn't be used. We've been dealing perfectly well with cyclical data for centuries now.

 

What? There is not "months of data". We had a mid December launch. That rolled right into the holidays which will obviously create outlier points due to people's work/school schedules then we had the announcement that they have been playing with server cap limits in early Feb.

 

At best, we have about one month of data to go on so far, which shows a very slight decline from Jan 24 to today.

[Darka]

Why do I get the feeling that the next time I check this thread that its going to be a 30+ page flame fest, I here the Drones and Fanbeez swarming their way here

[Apocalypsezero]

You know of Occam's Razor, I presume? The idea that the simplest explanation is probably the correct one.

 

We could explain a million reasons around why those charts are wrong... but that, too is speculation and unlikely speculation.

 

I doubt BioWare every changed the threshold for what "heavy" or any other designation meant, even when adjusting the server caps. It would have been disingenuous and if they did it to hide population drops, well, that's another story all together. Safe to say that didn't happen.

 

The simplest explanation here is that there are less people playing less often than before.

 

On a static situation yes, but server caps are not static.

[Dezzi]

EDIT - Ignore Edited February 23, 2012 by Dezzi

[iceperson]

nice post.

 

i remember when Rift lowered the label thresholds for all their servers not long after release instantly making all those charts go nuts looking like player pop jumped when it was obvious only the full servers had healthy pops. I think they've culled 60+ servers since that time.