could it be a mistake to seek balance in relation to playing time?

[Mabi black.1824]

2 hours ago, Dawdler.8521 said:

No it doesnt because you dont know how Anet weighs any of this in the algorithm.

This is the same as saying 1 power adds 1 damage and calling it mathematically correct.

probably you are right, in fact I asked for help to the administrator to try to understand how the algorithm reasons.

it is true that we write a lot of things in the forum, but sometimes a little help from someone who works in the anet team would really be self-driving and very welcome.😅

[Chaba.5410]

A lot of confusion in this thread.  Lemme try to provide some clarification.

 

1 - In 2017, the size of a world was changed from raw number of accounts to "WvW activity".
https://forum-en.gw2archive.eu/forum/game/wuv/World-Population-Changes-Are-Coming

2 - We have an idea of what "WvW activity" means.

"We have simulated other algorithms to measure world size and ultimately found that player hours gave us more accurate results because we are mostly comparing active WvW play."

3 - And we have an illustration of how measuring the size of a world through "WvW activity" determines server linking.  Notice that the sizes of hypothetical worlds in this dev post are listed as a percent.  That's the percent of "Full", which is a threshold that Anet can arbitrarily adjust and has in the past.

https://forum-en.gw2archive.eu/forum/game/wuv/New-Worlds

4 - What does this mean for World Restructuring?

"We already use play hours to determine population status of a world and when we refer to player value or evaluation or score we are talking about play hours with some kind of scalar adjustment. With this system, moving people around every two months we would like to track a few more things that can help us distribute players more effectively."

"Doing this can help us get more even matches. WvW is not completely a number of bodies game. A hundred veteran players will always beat a hundred casual players."

 

So no, the hypothetical 1000 high activity players vs. 2000 low activity players will not happen.  And it doesn't happen now with servers despite population being based on play hours because the system doesn't place players on servers so we're already mixed.

Edited February 3, 2022 by Chaba.5410

[Chaba.5410]

7 hours ago, Mabi black.1824 said:

hi chaba,

try to look at what I wrote with an open mind. do not take into account the special exceptions that may be present on all sides of the game.

do you think my reasoning could be correct? after all, there must be a reason and therefore an error in the algorithm that builds the teams.

You could've just gone back and read the World Restructuring FAQ since it addresses the main assumption in your original post.  /shrug

[Mabi black.1824]

1 hour ago, Chaba.5410 said:

So no, the hypothetical 1000 high activity players vs. 2000 low activity players will not happen.  And it doesn't happen now with servers despite population being based on play hours because the system doesn't place players on servers so we're already mixed.

you were really kind to reconstruct the speech that I had completely missed since not in the forum for a short time.

I also completely agree with you that the system today has no way to move players and therefore discriminate against those who play more on one side and those who play less on the other.

at this point with certainty we can say that it is impossible to build 1000 more active players vs 1000 less active players (because I count uses the time parameter and the world has in cap)

in the same way we can say (although with little probability) that we could have a team of 2000 players against a team of 1000 players and for the current algorithm they would be two perfectly balanced teams.

 

Of course, I am pushing him to the extreme in order to make the idea better. 

but how is it possible that all the players who spend a little more time find themselves in the same server? or in some few servers?

I do not know, but perhaps with the passing of months and years players conform to a certain style ; or simply in some servers there are a handful of more commanders who cover more hours than another or maybe on that server there is a group of pensioners unleashed etc etc.

this does not detract from the fact that the limit construction that I have indicated can happen . the algorithm rightly does its job but the result it gets could go in the opposite direction.

after all, there must be an error somewhere in the algorithm.

 

[Mabi black.1824]

1 hour ago, Chaba.5410 said:

You could've just gone back and read the World Restructuring FAQ since it addresses the main assumption in your original post.  /shrug

I have absolutely not thought about world restructuring because it has nothing to do with my reasoning, or attempt at reasoning.

the restructuring of the world adds the concept of alliances and their distribution by increasing granularity and that's fine. my thought is on how spiders the algorithm number of players for the time. if there resides an unwanted error we also find it in the new system.

I also add that someone here has rightly stated that it would not be correct to discriminate against a player because he devotes a little more time / passion to the mode. but to all intents and purposes that is exactly what the algorithm does at the moment.

[Chaba.5410]

6 minutes ago, Mabi black.1824 said:

I do not know, but perhaps with the passing of months and years players conform to a certain style

You're on the right track.  Every server at this point has an ossified reputation and players who identify with that reputation will seek that server out.  I've always maintained that WvW is at it's healthiest when there is a diversity of playstyles so these ossified servers are really bad for the game mode IMHO.

[Mabi black.1824]

I personally believe that the new alliance system will make me lose a long-standing community with which I have shared a lot. but I also think it will bring a lot of novelty or freshness as you call it.

but that's another matter, here I'm thinking about balancing mechanics. number of players and playing time and a possible unexpected outcome.

[Chaba.5410]

32 minutes ago, Mabi black.1824 said:

but that's another matter, here I'm thinking about balancing mechanics. number of players and playing time and a possible unexpected outcome.

I think Anet is already looking at how to avoid the unexpected outcome you envision though.  At least they already know about it.

"when we refer to player value or evaluation or score we are talking about play hours with some kind of scalar adjustment. With this system, moving people around every two months we would like to track a few more things that can help us distribute players more effectively"

Edited February 3, 2022 by Chaba.5410

[Xenesis.6389]

Most likely Anet will not share the ins and outs of their sorting system, as that would lead to players trying to game the system, this is why they haven't released much information on population activity since they started using that to measure populations back in 2015. Only that they look at previous weeks of activity data to see if there's any abnormal readings like mass transfers.

They have mentioned there could/will be different values to players based on certain conditions, one of the main ones being commanding, they have even considered things like wxp gain and time zones, but they are not used currently as the data has major problems with it.

I would suggest for those who are interested in discussing world restructuring, to maybe go read up all the information released on the topic over the years. Some of it could be outdated, but at least you can get an idea for the foundations of it.

https://en-forum.guildwars2.com/discussion/26547/world-restructuring

https://en-forum.guildwars2.com/discussion/45856/world-restructuring-update-1

https://en-forum.guildwars2.com/discussion/61986/world-restructuring-update-2

https://en-forum.guildwars2.com/discussion/26877/world-restructuring-faq

The were two previous times for proposals to add more new smaller worlds, more pieces, for better links, before alliances came to be. Essentially alliances could have had their own worlds linked to the main ones, but it would have mostly been player controlled still instead of anet forcing everyone into the washing machine together with alliances. It would have been the best of both worlds, those that wanted their static server community, and those that wanted to play with their guilds and friends. This was the chance to avoid the alliance system, but a lot of people didn't like it, maybe because they didn't realize it's potential...

May 2016 https://forum-en.gw2archive.eu/forum/game/wuv/Hypothetically-Speaking-New-Worlds

Nov 2016 https://forum-en.gw2archive.eu/forum/game/wuv/New-Worlds

The population switched over from counting bodies to measuring activity in July 2015, some answers on how they measure are in the threads.

https://forum-en.gw2archive.eu/forum/game/wuv/World-Population-Changes-Are-Coming

https://forum-en.gw2archive.eu/forum/game/wuv/Transfers-Links-and-Population-Data/6664495

https://forum-en.gw2archive.eu/forum/game/wuv/Yaks-Bend/6513382

For a deeper dive you can check all the post of the developers who were involved with those discussions, and the other topics around wvw at the time such as mix borderlands, relinks, skirmish mode, etc. Note: these devs are no longer with anet, so again the information may be outdated on some things.

https://en-forum.guildwars2.com/profile/59-mckenna-berdrow2759/

https://forum-en.gw2archive.eu/members/McKenna-Berdrow-2759/showposts

https://forum-en.gw2archive.eu/members/Tyler-Bearce-3427/showposts

Edited February 3, 2022 by Xenesis.6389

[gloflop.3510]

I think that we are all more and more going into a theoretical discussion.

Is it theoretically possible that one world consists of 2000 accounts with a low activity and a second world consists of 1000 accounts with a higher activity? => Yes, it is.

Does it happen that one world has 2000 accounts while another has 1000? => Only anet has the data and can say something for sure but I doubt it due to two reasons: 1) It never felt to me like I had a matchup in which such an imbalance occurred. 2) Anet would have communicated differently. Of course they would not say the numbers straight but between the lines would have admitted that they see a problem like this. @Xenesis.6389 collected relevant posts by anet. Feel free to read yourself.

Will world restructuring make such an imbalance more likely? => It depends on how anet is programming their algorithm. However, from what I could read, they are aware of the problems surrounding population balance. It will just require some time for them to program.

[displayname.8315]

What do we think explains the no-activity shards during the beta? 

 

Were they the low playtime casuals that log in every now and then to leach off WvW zergs?

 

I just know my main account with diamond raider rank was put in a hot match against the most active(mag).  The team color I was on seemed very active pretty much any time I checked.

[RisingDaewn.8703]

In Deso's case, a still pretty active pug and active at (EU or NA) night server like Deso; who usually ends up in T1 with a link; even with links like Kodash and Fort Ranik, it was kind of expected that no link would happen, compared to 2 years ago when over 80+ players left, including most of not all of the GvG and Open field WvW Guilds.

Even then it took two server relinks for the the Linking system to adjust back then, until Deso got no links for a majority of 2020 and stayed full for a long while, after two stacked Links in Blacktide and Underworld in early 2021. 

So in the case of trying to fix the current Linking system,, balancing populations between the links and hosts (which is not that simple as just changing gem costs) and preventing transfers from imbalancing things, it is most likely too much work to attempt to fix everything from the old coding and such, without a proper rework like World restructuring.

All the more reason, as  I don't think they've disclosed how they decide the Server Links, which has created more issues than bringing balance to WvW.

 Frequently Asked Questions - World vs. World - Guild Wars 2 Forums

Q: When is a world considered "Full"?

 

A: Worlds are considered full when the average weekly playtime in WvW exceeds a set threshold. The status of worlds is periodically checked and then the “Full” status of worlds is adjusted.

 

Q: But my world never has a queue but we're still full?

 

A: We adjust based on average weekly playtime. This does not mean that your world will always have queues if it is marked as full. A world with high peak time population and a world with more spread out activity could both exceed the playtime threshold.

 

Q: But I want to play with my (guild/friends) and they are on a "full" server.

 

A: This is an unfortunate side effect of locking worlds and we agree that it can feel bad. However, if we do not lock worlds, worlds become overpopulated and it becomes impossible to have balanced match-ups.During any given period, you will find that some host worlds and links are open. While this may require some members of a group to make changes, there are options for everyone.

 

Q: When will my world be open?

 

A: This depends on the players on that world. If players transfer to a different world or stop playing for a period then they no longer will affect the thresholds and the world could open back up.[This is why it is against forum rules to ask us to open a world since world playtime determines this, not ArenaNet]

 

Q: Why are some worlds allowed to remain open over the threshold?

 

A: There are several reasons for this:

1. We adjust world population status periodically, but it does not update constantly. We are working on automating this process, but at this time, worlds occasionally can get more players on them because the threshold is not updated every time someone transfers.
2. Thresholds have been adjusted. As the populations have adjusted so have the thresholds. We switched from using WvW players to WvW play hours, and because of this the thresholds needed to change. Adjusting thresholds has caused some worlds to have populations temporarily over the threshold.
3. Players coming back to the game can put a world over the threshold. When a player has stopped playing for an amount of time they are no longer considered in the world thresholds. When they return to the game this can cause the world to go over the threshold since we do not remove people from worlds. We also have no plans to remove people from the worlds they play on because this would be a bad experience for those players.

Q: Are thresholds based on the top populated world?

 

A: Yes and no. When we adjust thresholds, we always take into consideration the worlds that have the highest population and play hours and make sure they are “Full”. However, the thresholds do not change based on the biggest worlds because this could be more easily manipulated by players, where they purposely do not play for several weeks to lower their population and open the world back up.

 

 

Edited February 5, 2022 by CrimsonNeonite.1048

[Mabi black.1824]

a lot of useful information, thank you.

at this point I wonder and I ask you, we could build the teams are in reference to the number of players. give as a goal to the algorithm to build teams that have the same number or something that comes as close as possible to the same number.

let's give ourselves the certainty of finding teams of 1000 vs1000  clearing the field of any doubt.

the time factor that in fact can have its relevance we make it work only on the points of war.

we say that the system checks online players every 60 minutes and consequently gives a value to their war points. if all 1000 players are at work their points have value 1, if only 500 players are at work their points take on a value 2.

the time factor is too volatile to build an 8-week match, with the extreme risk of being in 2000 vs 1000.

I mean this week I can be covid positive and I will have a game time next week I'm at work and I'll have another one the week after in the company they ask me for overtime and I'll have a different game time etc etc.

 

also in the same way we could reason about the concept of coverage, even the coverage will have an influence on the value of the war points, not on the number of players of a team.

 

in short, give priority to the algorithm to build teams with certainty and obtain in any case 1000 vs 1000.

 

everything else you want to consider (time, coverage, numerical inferiority on the map, and everything you want to take into account) are just variables that will affect the value of the war points.

we also consider that fixing how the algorithm works is necessary today as tomorrow with alliances.

[Dawdler.8521]

And in conclusion, absolutely nothing has been accomplished here and Anet will continue with whatever intended algorithm they have while not giving us any details, like normal.

[SweetPotato.7456]

1) I actually do not know why player's play time would be something to be considered in the matching (for population balance?), when we are already forced to join guild in order to play with our preferred team. 
2) Why would it be needed when unless they are giving players extra something for playing "overtime".

3) Why would it be needed if  playtime cannot solve the time zone population balance.
4) Why would it be needed if  player can choose to drop tier and not log / play less whenever they  plan to manipulate the tier. (one up one down scoring system)

[Chaba.5410]

1 hour ago, SweetPotato.7456 said:

1) I actually do not know why player's play time would be something to be considered in the matching (for population balance?), when we are already forced to join guild in order to play with our preferred team. 
2) Why would it be needed when unless they are giving players extra something for playing "overtime".

3) Why would it be needed if  playtime cannot solve the time zone population balance.
4) Why would it be needed if  player can choose to drop tier and not log / play less whenever they  plan to manipulate the tier. (one up one down scoring system)

Those sound like questions you can research and answer.  Plenty of links to historic forum posts have been provided.

[Chaba.5410]

5 hours ago, Mabi black.1824 said:

also in the same way we could reason about the concept of coverage, even the coverage will have an influence on the value of the war points, not on the number of players of a team.

 

I'm not sure what you mean by "coverage" here.  Why are you treating coverage as something separate from numbers?

You claim that number of players on a team does not have an influence on the score.  So let's make an example 500v500v500 player match.  Sounds balanced?  On one team, the 500 players all play 6 hours before logging off.  On the other two teams, their 500 players can only play 2 hours before logging off.  On the first team, 500 players created coverage by their longer play hours.  How is that balanced?  How does feeling forced to play longer hours to remain competitive incentivize wanting to play and making winning matter?

You do seem to understand this concept though that non-primetime players have a higher value to the scoring system than the "dime-a-dozen" primetime players when you wrote:

6 hours ago, Mabi black.1824 said:

we say that the system checks online players every 60 minutes and consequently gives a value to their war points. if all 1000 players are at work their points have value 1, if only 500 players are at work their points take on a value 2.

A never-implemented score multiplier proposal did exactly that.  It is considered controversial though because low action level players would feel like they are being measured differently (despite their overtuned value).  Isn't adjusting an individual players' value in points the kind of "if I choose to play 2 hours at gw2 I do not understand why I have to be counted differently from my brother who plays 1 hour at gw2"  issue you had a problem with over measuring by play hours though?  (See your comment on the World Linking thread.)

 

"

• While the above change takes steps to bring the value of off-hours coverage in-line, there’s a good chance it’ll still be overvalued. If that’s the case (and we’ll eventually poll on this), then we have plans for an additional system.
• This is the Action Level – Victory Point Multiplier system
  • This system would multiply the Victory Points awarded by Skirmishes based on map populations and time of day.
  • During prime time hours, the multiplier would always be at it’s maximum of 3.
  • During off hours, the multiplier might stay at 3 or drop to 2 or 1, depending on on activity level.
  • It’s important to include map populations as a factor, to make the system more fair for off hours players and its important to include time-of-day as a factor to prevent a winning team from trying to keep the score muliplier low by exiting WvW"

https://forum-en.gw2archive.eu/forum/game/wuv/Let-s-Talk-Scoring

[Mabi black.1824]

26 minutes ago, Chaba.5410 said:

I'm not sure what you mean by "coverage" here.  Why are you treating coverage as something separate from numbers?

You claim that number of players on a team does not have an influence on the score.  So let's make an example 500v500v500 player match.  Sounds balanced?  On one team, the 500 players all play 6 hours before logging off.  On the other two teams, their 500 players can only play 2 hours before logging off.  On the first team, 500 players created coverage by their longer play hours.  How is that balanced?  How does feeling forced to play longer hours to remain competitive incentivize wanting to play and making winning matter?

hi chaba,

you probably misunderstood me, or I explained myself badly.

in practice I build identical teams repeated to the number of players.

in reference to my first post for statisctica I predict that the majority of them will have a similar and balanced gaming experience (the concept that we play almost all in our free time)

they will all have the same playing time ----> no. how do I monitor it? -----> the algorithm checks every hour how many players are active compared to the limit imposed by the team. consequently it defines the value of the war point of one team and another.

in this way I have associated the number of players and their variable coverage, I am not separating them.

it is not necessary to feel forced to play more time than your opponent, the algorithm takes care of it by working on the value of the war point. if you are 500 vs 1000 you can rest assured because your score is worth twice that of your opponent. capture or defend half of your opponent .... no problem the points of war will be properly proportionate.

with the logical styessa you can add as many variables as you want...... there are more commanders on the one hand, I modify the coefficient points of war, you want to balance the time zone, I divide the 24 hours into 4 time slots and modify the coefficient of the war points etc etc.

on the other hand you will have the certainty of having three numerically identical teams. the variables (precisely because they are variables) does not affect the number of players with whom I have built the teams.

[Chaba.5410]

15 minutes ago, Mabi black.1824 said:

it is not necessary to feel forced to play more time than your opponent, the algorithm takes care of it by working on the value of the war point. if you are 500 vs 1000 you can rest assured because your score is worth twice that of your opponent. capture or defend half of your opponent .... no problem the points of war will be properly proportionate.

I'll play devil's advocate here for you.

"Why are my actions being scored differently than another player?"

"Let me get this straight?  If I kill a player, I'm only getting half the amount of points as the guy on the other team for the same effort?  This game sucks, I quit!"

"GET OUT OF WVW!  WE'RE ALMOST AT 2x multiplier and just need a few more people to exit!"

Edited February 4, 2022 by Chaba.5410

[Mabi black.1824]

3 minutes ago, Chaba.5410 said:

I'll play devil's advocate here for you.

 

😊 I'm just pleased to compare myself with other players.

let's say that 500 vs 1000 in theory if you killed a player you will have committed a little more than your opponent I guess you will have found a complete group of 50 vs a lower group of 30 maybe.

and then when I talk about war point I mean the score of the team . the one that defines one on and one down.  no influence on the personal score of the individual player, the experience or gold you earned before remains identical.

[Chaba.5410]

22 minutes ago, Mabi black.1824 said:

😊 I'm just pleased to compare myself with other players.
 

This seems contradictory to what you wrote previously so you have me confused as to why you would rather have population counted by number of players than play hours.  The length of time a player plays is being compared to the length of time other players play.

Your proposition completely falls down if any one part of your supporting points becomes invalid.  Like removing a stone from an archway.  The entire thing falls down.

22 minutes ago, Mabi black.1824 said:

and then when I talk about war point I mean the score of the team . the one that defines one on and one down.

Yes.  Players would wonder why is the score being counted differently for them?
 

22 minutes ago, Mabi black.1824 said:

no influence on the personal score of the individual player, the experience or gold you earned before remains identical.

That's how it is now.  I touched on that here:

"You do get more results if you play longer.  You reach Diamond chest before someone else.  You progress through reward tracks faster.  You gain more WXP and loot because you kill more players and cap more structures than someone who doesn't play as long. "

Edit:  All that said, I personally think the action level idea with the victory point multiplier as detailed in the "Let's Talk Scoring" dev post should be tried.  It's a scoring change that operates at a more generalized skirmish level and doesn't involve the warscore.

Edited February 4, 2022 by Chaba.5410

[Mabi black.1824]

and then the efficient multiplier of war points as a player you do not know what value it has at that precise moment.

it is clear that I would not give you a way to abuse it. only the administrator has a way to count the active players every 60 minutes you will never know.😁

[Mabi black.1824]

well of course somehow you have to decide how to build balance.

I think I do not remove bricks but add them and make it more stable 😅

in fact if I go online and I find myself on a map outnumbered by 50% constantly throughout the week, what do you predict that I will have earned the same wxp points as the other player my opponent? I don't think so. and you can find it at any time with the k/d parameter.

if I look at how many players we killed and how many players died this week my team comes to half of those of my opponents.

wax perhaps a 50% lower activity in my parts? so we were maybe half the players compared to my opponents? is it right that I earn less wxp than my opponent because I am on the side of numerical inferiority?

no. that's why I use a coefficient that corrects the parameter of war points.

[Mabi black.1824]

Yes.  Players would wonder why is the score being counted differently for them?

 

because at that precise moment on the map you are twice as many as your opponents.

[Chaba.5410]

1 hour ago, Mabi black.1824 said:

Yes.  Players would wonder why is the score being counted differently for them?

 

because at that precise moment on the map you are twice as many as your opponents.

Which leads to...

"GET OUT OF WVW!  WE'RE ALMOST AT 2x multiplier and just need a few more people to exit!"