How does ANET Measure Population in relations to WvW?

[Morden Kain.3489]

Well, any way we slice things, we do not know the actual algo that ANet uses any more.  With that, I am going to have some fun in WvW.  GL to everyone tonight... If I kill ya, nothing personal 😁.  If you kill me, oh well 😏.

[Chaba.5410]

5 hours ago, Karagee.6830 said:

perhaps only specific time-slots or each time-slot is weighted (so prime time will have a higher weight)

It was suggested by an Anet dev in the past, and can be dug up on the old forum archive, to weight skirmish scores by activity levels.  NA and EU prime times would always be weighted at the max level so as to increase the prime-time player's value towards the score and bring it inline with off-hours players.

Edited November 19, 2022 by Chaba.5410

[Chaba.5410]

4 minutes ago, Morden Kain.3489 said:

Well, any way we slice things, we do not know the actual algo that ANet uses any more.  With that, I am going to have some fun in WvW.  GL to everyone tonight... If I kill ya, nothing personal 😁.  If you kill me, oh well 😏.

What would be the reason to suggest they don't use it anymore?  Because someone edited the wiki page?  That doesn't make much sense.

[Karagee.6830]

11 minutes ago, Chaba.5410 said:

It was suggested by an Anet dev in the past, and can be dug up on the old forum archive, to weight skirmishes by activity levels.  NA and EU prime times would always be weighted at the max level so as to increase the prime-time player's value towards the score and bring it inline with off-hours players.

I misread. The main problem I see is Anet not counting total play hours, but some weighted metric, when the score assumes a balance based on total played hours. However this would explain why for servers like Gandara there is such a disconnection between server status and actual participation...if your participation is concentrated at prime time and weekend and that is weighted more, then the metric is going to wrong by some margin.

Edited November 19, 2022 by Karagee.6830

[Xoranium.6481]

IMHO it's like this.. 

If you step foot in wvw for even 1 second during the week it counts towards population...

No other proposed calculation makes any sense

[Aeolus.3615]

On 11/18/2022 at 7:25 PM, Xenesis.6389 said:

Anet probably wouldn't even need to check on timezones either if players manage to organize themselves into alliances with round the clock coverage.

 

It's funny how the very first alliance in the game all the way back in 2012 showed the flaw of the system, but also showed how to win the game, yet barely anyone has managed to replicate it in 10 years, it might come back after WR if the stakes are high enough for players to bother. But most players don't care for anything else other than their own time of play is stacked.

 

Yeah but Gw2 players dont like to fight.. they want just to ktrain or blob up with mass minstrells perma all boons smaller groups and call themserves a wvw guild...

If theres fights in wvw the game is unplayable due skill lag and masive fps drop anyway... Maybe making fights be more PVD vs cap back,  wins who ktrain faster and harder  maybe thats the combat Anet wish towards wvw, (well it is how combat has been in this gamemode anyway) which is pretty bad if so... players will still have to organize to be on maintenance when alliance has its weak timezone, VS alliances  that will try to hoard all timezones possible.

Wont be much diference but will solve some cricual  issues towards WvW.

 

EDIT: players will always try to manipulate to fight empty servers, and Anet will always pair very heavy populated alliances vs aliances mostly with RNG players and huge timezones gap.

Edited November 24, 2022 by Aeolus.3615

[Morden Kain.3489]

On 11/18/2022 at 5:28 PM, Chaba.5410 said:

What would be the reason to suggest they don't use it anymore?  Because someone edited the wiki page?  That doesn't make much sense.

Well, unless you are an ANet employee with access to the source code, we are purely speculating as to the algorithm being used in the first place.  It does not matter if it was posted to the wiki, then subsequently removed.  Unless an ANet developer comes along and says "we do it this way" and fills in the blanks on the algorithm (in English, not code), we will always be hypothesizing, postulating, presuming, and speculating. 

All we know is that ANet uses WvW participation as the basis for the world population.  Until such time as this changes, feel free to continue hypothesizing on the algorithm.

[Chaba.5410]

21 minutes ago, Morden Kain.3489 said:

Unless an ANet developer comes along and says "we do it this way" and fills in the blanks on the algorithm (in English, not code), we will always be hypothesizing, postulating, presuming, and speculating.

They have. It's not enough detail to game the system, but it's been far more than "participation". And there's been no similar discussion by them to suggest they changed the system from what they first described. Like, no notification at all that there was any change.

Edited November 23, 2022 by Chaba.5410

[Strider Pj.2193]

On 11/23/2022 at 1:48 PM, Morden Kain.3489 said:

Well, unless you are an ANet employee with access to the source code, we are purely speculating as to the algorithm being used in the first place.  It does not matter if it was posted to the wiki, then subsequently removed.  Unless an ANet developer comes along and says "we do it this way" and fills in the blanks on the algorithm (in English, not code), we will always be hypothesizing, postulating, presuming, and speculating. 

All we know is that ANet uses WvW participation as the basis for the world population.  Until such time as this changes, feel free to continue hypothesizing on the algorithm.

What we ‘hypothesized’ about the ‘full’ status is that for a while they used one server from NA (and likely one on EU) that was their max server.  For a long time (and maybe even now) BG was that server.  No matter how many players left, they stayed full.  The hypothesis was as BG dropped? Other servers became full based on that benchmark.  It drove players to less populated servers.  Over time (and we are talking years) BG started to open again, but they only kept them open for hours at a time then likely manually closed them again.  Gradually, that ‘full threshold’ dropped.  So….  Now we have a much lower ‘full threshold’.  Which also means, some servers that were ‘open’ and a certain population/play hours, would now be full.  
 

But play hours has always been their stated process.

 

As to all of the specifics, they won’t release it.  Period.  So, to some extent, we are speculating.  But it does make me laugh when people keep asking Anet to tell us…. They have NOTHING to gain by posting the actual activity numbers.  

[Karagee.6830]

48 minutes ago, Strider Pj.2193 said:

What we ‘hypothesized’ about the ‘full’ status is that for a while they used one server from NA (and likely one on EU) that was their max server.  For a long time (and maybe even now) BG was that server.  No matter how many players left, they stayed full.  The hypothesis was as BG dropped? Other servers became full based on that benchmark.  It drove players to less populated servers.  Over time (and we are talking years) BG started to open again, but they only kept them open for hours at a time then likely manually closed them again.  Gradually, that ‘full threshold’ dropped.  So….  Now we have a much lower ‘full threshold’.  Which also means, some servers that were ‘open’ and a certain population/play hours, would now be full.  
 

But play hours has always been their stated process.

 

As to all of the specifics, they won’t release it.  Period.  So, to some extent, we are speculating.  But it does make me laugh when people keep asking Anet to tell us…. They have NOTHING to gain by posting the actual activity numbers.  

That is not correct. Posting some indication (not the actual numbers) would show they have a system in place that makes some sense. Even if they did tiers. In a sense they already do with the server status, but they don't want to say anything that involves numbers because what happens if the metrics show something very different than server status? People would be ready with pitchforks especially on EU where 3 server get shafted for 2 months every relink.

Or, you know, they could just openly say that server status DIRECTLY reflects their metrics over a certain number of weeks and we should believe it. There is no harm in doing this since nobody can check, especially if it was true. However, people are often reluctant to outright lie when they can simply avoid the topic altogether.

Edited November 25, 2022 by Karagee.6830

[Strider Pj.2193]

20 minutes ago, Karagee.6830 said:

That is not correct. Posting some indication (not the actual numbers) would show they have a system in place that makes some sense. Even if they did tiers. In a sense they already do with the server status, but they don't want to say anything that involves numbers because what happens if the metrics show something very different than server status? People would be ready with pitchforks especially on EU where 3 server get shafted for 2 months every relink.

Or, you know, they could just openly say that server status DIRECTLY reflects their metrics over a certain number of weeks and we should believe it. There is no harm in doing this since nobody can check, especially if it was true. However, people are often reluctant to outright lie when they can simply avoid the topic altogether.

🤷

 

Join a stream with them and ask.  Good luck.

 

Maybe when your server does another protest sit out you can take the time and join one.  
 

Yeah.  EU is abused.  😬

[Chaba.5410]

6 hours ago, Strider Pj.2193 said:

What we ‘hypothesized’ about the ‘full’ status is that for a while they used one server from NA (and likely one on EU) that was their max server.

I assumed @Morden Kain.3489 was talking about the algorithm that works off of playhours and not the Full/Very High status thresholds.  Maybe he was lumping them together though?  I consider them distinct.

I suspect thresholds have always been manually set, or at least there's very little automation evident in it.  If you recall, the host servers used to all be marked as Full even if they weren't when server links first rolled out.  The intent was to get people to move to the lower populated linked server.  So at least there's a method they have to override any automation of a server's status.

And as you noted, BG has opened in the past.

Edited November 25, 2022 by Chaba.5410

[Strider Pj.2193]

1 minute ago, Chaba.5410 said:

I assumed @Morden Kain.3489 was talking about the algorithm that works off of playhours and not the Full/Very High status thresholds.  Maybe he was lumping them together though?  I consider them distinct.

I suspect thresholds have always been manually set, or at least there's very little automation evident in it.  If you recall, the host servers used to all be marked as Full even if they weren't when server links first rolled out.  The intent was to get people to move to the lower populated linked server.  So at least there's a method they have to override any automation of a server's status.

And as you noted, BG has opened in the past.

You are likely correct on each count.  🙂 

 

 

[Chaba.5410]

3 minutes ago, Strider Pj.2193 said:

You are likely correct on each count.  🙂

I wanted to add that BG used to not get a link for a real long time.  That's how overstacked above the Full threshold they were.  They had excellent historic coverage because of that overstacking that kept them from dropping tiers.  EU servers don't really have much coverage compared to NA servers.

Now BG regularly gets a server link.  So the intent of getting people to move to other NA servers by keeping BG Full seems to have worked.

Edited November 25, 2022 by Chaba.5410

[Strider Pj.2193]

3 minutes ago, Chaba.5410 said:

I wanted to add that BG used to not get a link for a real long time.  That's how overstacked above the Full threshold they were.  They had excellent historic coverage because of that overstacking that kept them from dropping tiers.  EU servers don't really have much coverage compared to NA servers.

Now BG regularly gets a server link.  So the intent of getting people to move to other NA servers by keeping BG Full seems to have worked.

🙂

[Kal Durak.2830]

On 11/24/2022 at 2:48 AM, Morden Kain.3489 said:

Well, unless you are an ANet employee with access to the source code, we are purely speculating as to the algorithm being used in the first place.  It does not matter if it was posted to the wiki, then subsequently removed.  Unless an ANet developer comes along and says "we do it this way" and fills in the blanks on the algorithm (in English, not code), we will always be hypothesizing, postulating, presuming, and speculating. 

All we know is that ANet uses WvW participation as the basis for the world population.  Until such time as this changes, feel free to continue hypothesizing on the algorithm.

I don't need to know anything about algo, Anet employee, etc... I know this:

30 of us plus 10 Pugs are on a BL map, and we are facing 70 x 2 from two other servers.  = Not balanced.  Cannot play against giant blobs. 

I'm told they consider your rank in total play time as well.  That's why I have seen entire blobs of scouts. No diamonds etc... Who knows this last fact.  It does match up to our numbers versus theirs in WvW. I can see that easily without any further knowledge.