Yaks Bend

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Now if you could only find a way to weed out the AFK population and the achievement population from the “active” WvW population.

Yaks Bend

in WvW

Posted by: Robin Da Poor.6281

Robin Da Poor.6281

So, I assume this means because YB has a few dedicated players that spend hours upon hours in WvW which would also assume they would also gain ranks at the same time doing things in WvW makes them highly populated but on the same level having servers that have lots of players spending a hour or so on WvW but not gaining ranks because they are not in long enough makes them underpopulated in your algorithm.
Does this not penalise a server for having dedicated WvW players?

So in theory, similar to a post above a server could have 500 players dedicated spending hours in WvW and then a server have 1000 players doing just a couple of hours, but the server with 500 gets locked.

SharpShooters [SS] | Yak’s Bend |
Pro With A Bow| Mes Mer Relda |

Yaks Bend

in WvW

Posted by: Nikola.3841

Nikola.3841

So, I assume this means because YB has a few dedicated players that spend hours upon hours in WvW which would also assume they would also gain ranks at the same time doing things in WvW makes them highly populated but on the same level having servers that have lots of players spending a hour or so on WvW but not gaining ranks because they are not in long enough makes them underpopulated in your algorithm.
Does this not penalise a server for having dedicated WvW players?

So in theory, similar to a post above a server could have 500 players dedicated spending hours in WvW and then a server have 1000 players doing just a couple of hours, but the server with 500 gets locked.

Unlike DAOC for example, WVW is not the end game, so of course its possible that there are big differences in percentage of players dedicated to WVW among servers

Yaks Bend

in WvW

Posted by: Immortal.1543

Immortal.1543

Yaks Bend building siege isn’t why they’re locked of course. If that were the case they’d be locking a lot more server other than them.

you guys are funny but im being serious

When MaL is asking why YB is getting the shaft then there is something seriously wrong that should be looked at.

As someone who has never left YB, I’m pretty happy with our current lack of a need for siege and having a solid 5-7 fights guilds running strong. But stereotypes never die haha

But to clarify on McKenna’s post, does this using player hours as the metric mean that guilds running 4-6hr raids could actually hurt their server if they make up a disproportionate amount of time zone coverage? For example, say we run a 4 hr raid at 30 people in our guild group, but for the two time zones after (OCX/EU) we have 30 total people on the entire server running that time, is there any way to account for this? I ask because we are heavily outnumbered (2:1 and even 3:1 at times) during 2/3 of active time zones, but we can’t get our population spread out if we’re always full.

Immortal – YB4LIFE – [AI] Chronicus of Nornia/Lady Chronica

Yaks Bend

in WvW

Posted by: Liston.9708

Liston.9708

moral of the story…. no more scouting…. those people can rack up some serious time….

while I may not like the explanation, I appreciate it. So next question – where is the link? We do okay by ourselves in T3/T4 matches, but have no chance for t2 as is. (okay if we never go t1 ever again)

YB→YB→YB→YB→YB→YB→YB→most likely YB

Yaks Bend

in WvW

Posted by: Choppy.4183

Choppy.4183

I see a lot of jumping to conclusions here.

For example, this new algorithm shouldn’t lead you to conclude “no more scouting” unless you have loads of people doing nothing but reporting enemy activities. If that’s all your scouts do, and you have so many that it can tip the balance and get your server locked, then you need scout training more than transfers because you’re sitting on a pile of under utilized resources.

Same goes with the notion that turtling up in structures is the way to go (to drive down ranks gained). The boredome factor alone is bound to drive down participation rates, and you’ll need the transfers in to compensate for the players you’ll be hemorrhaging.

Finally, expecting Anet to balance for things like skill seems a bridge too far too far to me. That’s “everyone should get a participation ribbon” thinking, as far as in concerned.

If you’ve got enough players to get locked and they’re under performing, work on commander and player skill development.

I’m Biff Rangoon, and I approved this message.
Ehmry Bay | Omg Brb Icecream Truck (ICEE)

Yaks Bend

in WvW

Posted by: Busty Bombsalot.4570

Busty Bombsalot.4570

So from what I am reading here, Anet is penalizing Yaks for not only play long hours to get to t-1 but then also penalizing us for playing good as well.
Let’s add another 4 to 6 servers for us to fight against instead of the 4 now. I’m sure I can kill at least one 50 man blob all by myself so why not send more so other people have something to kill.
Gee I wonder how we dropped to t-4 from t-1 after only 4 or 5 months of no match ups .

(edited by Busty Bombsalot.4570)

Yaks Bend

in WvW

Posted by: Caliburn.1845

Caliburn.1845

Yes thank you McKenna.

The playerbase is going to be split on whether or not these are the proper metrics to use. But I think this is the sort of communication Anet should have with the WvW community. And I think the entire community would like to see more of it.

So a big thank you.

To the WvW community, I think there are a number of weaknesses to using the metrics mentioned as many of you have already noted. But what are better metrics to use?

Caliburn.1845, Monsters Inc.
Darkhaven>Dragonbrand>Blackgate>Maguuma>Yaks Bend>Stormbluff Isle>Yaks Bend

Yaks Bend

in WvW

Posted by: Choppy.4183

Choppy.4183

So from what I am reading here, Anet is penalizing Yaks for not only play long hours to get to t-1 but then also penalizing us for playing good as well.
Let’s add another 4 to 6 servers for us to fight against instead of the 4 now. I’m sure I can kill at least one 50 man blob all by myself so why not send more so other people have something to kill.
Gee I wonder how we dropped to t-4 from t-1 after only 4 or 5 months of no match ups .

How in the hell have you been penalized?

I’m Biff Rangoon, and I approved this message.
Ehmry Bay | Omg Brb Icecream Truck (ICEE)

Yaks Bend

in WvW

Posted by: Caliburn.1845

Caliburn.1845

It also occurs to me that we have seen a downturn in activity in NA T1 in recent months, T1 is often the least active tier in NA by kills/deaths(presumably also ranks). We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

Caliburn.1845, Monsters Inc.
Darkhaven>Dragonbrand>Blackgate>Maguuma>Yaks Bend>Stormbluff Isle>Yaks Bend

Yaks Bend

in WvW

Posted by: morrolan.9608

morrolan.9608

To clarify further this is the first week we are using this new algorithm. So some of the complaints that are being brought up were problems with the old algorithm.

We use play hours to determine the size. Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.

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. The past algorithms weighed more heavily on individual players, so we ended up with situations where JQ was ‘Full’ because they had a lot of players, just not necessarily ones that played as much as Blackgate.

TY for the change, albeit its come far too late, and this confirms exactly what I’ve been saying for months.

Jade Quarry [SoX]
Miranda Zero – Ele / Twitch Zero – Mes / Chargrin Soulboom – Engi
Aliera Zero – Guardian / Reaver Zero – Necro

Yaks Bend

in WvW

Posted by: ich.7086

ich.7086

It also occurs to me that we have seen a downturn in activity in NA T1 in recent months, T1 is often the least active tier in NA by kills/deaths(presumably also ranks). We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

That makes no sense at all, it’s the server population algorithm which has been updated not the matchmaking one …

I Am An Intruder – War 80
Ich Bin Marc – Thief 80
All Your Dolyaks Are Belong To [Us] (Fort Ranik)

Yaks Bend

in WvW

Posted by: Caliburn.1845

Caliburn.1845

Traditionally the full servers in NA have always been the top servers on NA, with only a rare passing exception here and there. Since their population status is linked to hours played in wvw, as guilds flee the dead staleness of T1, those servers might open up, as the bandwagon servers in the lower tiers close.

But as witnessed by TC currently, it is supremely hard to go down tiers sometimes now.

Caliburn.1845, Monsters Inc.
Darkhaven>Dragonbrand>Blackgate>Maguuma>Yaks Bend>Stormbluff Isle>Yaks Bend

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Also, since you’re re-evaluating your activity level algorithm, does this mean that you are going to review when a server is considered “full”?

Yaks Bend

in WvW

Posted by: ArchonWing.9480

ArchonWing.9480

So optimal play is to just turtle up on siege and defend only. This will minimize ranks gained (k-training being most ranks gained) and increase the likihood of gaining a linking.

YB siege-up and turtle meta revival incoming then?

But YB is full. So maybe YB doesn’t actually PPT.

AFK uninstallng. My perception of reality has been shattered.

For once you have tasted flight you will walk the earth with your eyes turned skywards,
for there you have been and there you will long to return.

Yaks Bend

in WvW

Posted by: Frost.5017

Frost.5017

I see a lot of jumping to conclusions here.

My response, and I’m quite sure Liston’s as well, was tongue in cheek. I was making a joke based on a very common dev/player relationship in that it doesn’t really matter how unfun a game system or mechanic is, players will use it if it helps them win, gain an advantage, or is deemed optimal. Those same players will use that system over and over and be mad at the devs that it’s not fun. It’s just how it goes sometimes: a player wants to ‘win’ and sometimes dev’s put the ‘fun’ and optimal (i.e. win) in very different spots.

McKenna’s post suggests (at least without knowing absolutely anything about the weighting, tuning parameters and/or other considerations in the population trending) that the ‘best way to play’ is play an incredibly boring turtle up and defend type play style. Not something we’re likely going to see happen in play here, but an example of the ‘fun’ and ‘optimal play’ being miles apart.

TLDR: Jokes are funnier if you include a couple of paragraphs of explanation.

Yaks Bend

in WvW

Posted by: Frost.5017

Frost.5017

But YB is full. So maybe YB doesn’t actually PPT.

Once Caliburn dials it back on dat map queue k-train he runs these days then YB will eligible for a link again.

Boom! Got ’em.

Yaks Bend

in WvW

Posted by: Choppy.4183

Choppy.4183

TLDR: Jokes are funnier if you include a couple of paragraphs of explanation.

Yes, that makes sense. I read something about that during my research into what humans call “humor”. (Thanks for the clarification) :P

I’m Biff Rangoon, and I approved this message.
Ehmry Bay | Omg Brb Icecream Truck (ICEE)

Yaks Bend

in WvW

Posted by: Liston.9708

Liston.9708

Indeed I was half joking. we will see next Monday when afkers and people who don’t really do anything don’t count against you.

Either way, I bet our 1 server count of hours/ranks is less than the sum of the linked server groupings around us. If this is true, the linking is just as flawed……

YB→YB→YB→YB→YB→YB→YB→most likely YB

Yaks Bend

in WvW

Posted by: TorquedSoul.8097

TorquedSoul.8097

It also occurs to me that we have seen a downturn in activity in NA T1 in recent months, T1 is often the least active tier in NA by kills/deaths(presumably also ranks). We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

That makes no sense at all, it’s the server population algorithm which has been updated not the matchmaking one …

It makes sense if considering actual WvW related activity like caps and kills … BG, Mag, and TC are all on the bottom half of the activity level chart.

Attachments:

Yaks Bend

in WvW

Posted by: Tiawal.2351

Tiawal.2351

To clarify further this is the first week we are using this new algorithm.

We use play hours to determine the size. Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.

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. The past algorithms weighed more heavily on individual players, so we ended up with situations where JQ was ‘Full’ because they had a lot of players, just not necessarily ones that played as much as Blackgate.

Thank You for this change! This likely can be further improved & tweaked, based on results, but still a good step, and it’s exactly what WvW needed the most.

This won’t be enough to counter the massive player transferring movements into linked servers, which constantly imbalances the system, because some or most, prefer to dominate through sheer numbers, instead of having a fun and competitive environment.
You have to take control over who transfers where, and for that, new metrics are needed, a new system, based on a (new?) Contribution Rating: while everyone is welcome, fresh blood is needed, but not everyone has the ability to contribute equally. Being able to and doing it actively, or having less contributing periods should be monitored, just like your PvP ranking system does.

“If You Can’t Measure It, You Can’t Improve It.” You know that.

Thank You for posting these, much appreciated!

A wandering ronin, employed by [ENMA]

Yaks Bend

in WvW

Posted by: ich.7086

ich.7086

It also occurs to me that we have seen a downturn in activity in NA T1 in recent months, T1 is often the least active tier in NA by kills/deaths(presumably also ranks). We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

That makes no sense at all, it’s the server population algorithm which has been updated not the matchmaking one …

It makes sense if considering actual WvW related activity like caps and kills … BG, Mag, and TC are all on the bottom half of the activity level chart.

Read with me :

We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

What puts server into Tx ? Glicko and matchmaking algorithms. Not the new algorithm which computes a server activity. This is what i was pointing out.

On the whole topic now :

Your data is just an another way to measure activity which has flaws and strengths as the way Anet measures it.

You are “locked” into T1 when your glicko rating is high enough and this doesn’t consider any form of server activity.

If you got a full map ktrain that just farm camp with full boosters againts a full t3 map that keeps ticking with like half of the defensors sitting on AC sup guess what : the ktrain server activity is higher (with both your and Anet’s metrics) but its PPT gain (thus glicko rating) is lower than the turtling server. The ktrain train server will get locked before the turling one (if that difference in behaviour is represented in average playtime). And guess what ? It’s perfectly fine.

Let me explain. The goal of wvw is to gain warscore. Servers are a black box that convert “player activity” in “warscore” (where warscore increase with player activity in ways that depend of server/coverage/…). The locking algorithm is just a cap in the input in that black box to prevent the “stack MOAR” mentality and tell players, ok now you got enough players to compete so stop complaining and git gud.

So let’s go back to the low activity in T1. Last time i checked, when a T2 server (with higher activity) goes into T1 it goes destroyed. That means that the T1 server output more warscore than the T2 (so they deserved to stay in T1) and at a lower activity (according to your charts we don’t know Anet’s numbers). They are more efficient than the T2 server. For the T2 to be locked into T1 has to output the same warscore than T1 server and at a activity lower than the lock thereshold (you can gain people/activity but not too much) : aka they need to play better (in terms of warscore).

So what’s wrong ?

I Am An Intruder – War 80
Ich Bin Marc – Thief 80
All Your Dolyaks Are Belong To [Us] (Fort Ranik)

Yaks Bend

in WvW

Posted by: Thea Cherry.6327

Thea Cherry.6327

We are playing against Drakkar this week, we farm them, get exp, rank up… Oh my god we are doomed. o.o

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Just had another thought.

Have you considered instead of using just player-hours to determine a server’s activity level, that you modify it based upon a servers kdr numbers?

Yaks Bend

in WvW

Posted by: Jayne.9251

Jayne.9251

Just had another thought.

Have you considered instead of using just player-hours to determine a server’s activity level, that you modify it based upon a servers kdr numbers?

You just want more loot bags.

Bad WvWer. Bad. Bad.

Thwump.

L’enfer, c’est les autres

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Just had another thought.

Have you considered instead of using just player-hours to determine a server’s activity level, that you modify it based upon a servers kdr numbers?

You just want more loot bags.

Bad WvWer. Bad. Bad.

Thwump.

I just thought that this adjustment might discourage four types of playstyles that are boring to play against.

1. The siege humping PPT style
2. The bait the enemy into siege range style
3. The attack only when you outnumber style
4. The map queue blob style

After I thought about this for a while, since BG is known for gaming the WvW system, I’d think they’d try to get their server open by having “wear no armor” raid weeks.

(edited by Swamurabi.7890)

Yaks Bend

in WvW

Posted by: Deemo.6094

Deemo.6094

Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.

Does this mean that a player who plays 8hours with exp boosters can count for a lot of spaces within the server? My question being, do exp boster effect server population?……………

Persistence overcomes resistance. In the end, all things you work for can be yours, except in gw2.

Yaks Bend

in WvW

Posted by: TorquedSoul.8097

TorquedSoul.8097

It also occurs to me that we have seen a downturn in activity in NA T1 in recent months, T1 is often the least active tier in NA by kills/deaths(presumably also ranks). We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

That makes no sense at all, it’s the server population algorithm which has been updated not the matchmaking one …

It makes sense if considering actual WvW related activity like caps and kills … BG, Mag, and TC are all on the bottom half of the activity level chart.

Read with me :

We could reach a point under the new algorithm where servers are far more likely to be locked below T1 then within T1 itself.

What puts server into Tx ? Glicko and matchmaking algorithms. Not the new algorithm which computes a server activity. This is what i was pointing out.

On the whole topic now :

Your data is just an another way to measure activity which has flaws and strengths as the way Anet measures it.

I think what Caliburn was suggesting is that if you measure population by some form of activity and have an arbitrary “population”/activity cap then if the lower tier servers are more “active” than tier 1, Anet might see this as excess population and lock those servers. And they may see the bored and stale T1 low activity as reduced population and open those servers up.

Another thing that is important to note is that overall activity in a tier seems to calibrated by the lowest activity server in that tier. So if you have TC bombing … which they are in terms of activity, that will drag down the activity of both BG and MAG. This is simply because of the highly reciprocal nature of WvW. You can’t take an objective back if it was taken from you. You can’t kill a player that is not on the map. And so on.

I don’t know if Caliburn was mentioning “performance” or skill as those probably don’t show up in the measure. In the cases where an “unskilled” server ends up in tier 1 and gets crushed, their activity for that week will be low but on average their activity will be higher because they will spend most of their time in the lower tiers and not tier 1.

This problem is further aggravated because there is really only two genuinely tier 1 servers which means that any 3rd server added is likely to bring the activity level down especially if they get stuck there for a long period of time … like TC. And Ironically, TC is not being rewarded for its lack of interest by getting into a lower tier but by receiving links. I think they currently have two.

And yes, all these types of derived population measures are flawed which is why I support things like battlegroups and fixed REAL population caps of some kind rather than the derived and manually curated process.

As far as population vs glicko, many see glicko score as a function of population. It may not be as bad as it was before linking but it is still a factor and will always be a factor in the minds of the players until there is a mechanism for a hard population cap.

(edited by TorquedSoul.8097)

Yaks Bend

in WvW

Posted by: TorquedSoul.8097

TorquedSoul.8097

Just had another thought.

Have you considered instead of using just player-hours to determine a server’s activity level, that you modify it based upon a servers kdr numbers?

I like that idea … it might punish blob servers.

Yaks Bend

in WvW

Posted by: TorquedSoul.8097

TorquedSoul.8097

Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.

Does this mean that a player who plays 8hours with exp boosters can count for a lot of spaces within the server? My question being, do exp boster effect server population?……………

Good point. Given that they measure ranks gained, they most certainly would skew the population measure.

Everyone stop taking experience boosters if you want your servers opened.

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.

Does this mean that a player who plays 8hours with exp boosters can count for a lot of spaces within the server? My question being, do exp boster effect server population?……………

Good point. Given that they measure ranks gained, they most certainly would skew the population measure.

Everyone stop taking experience boosters if you want your servers opened.

They used to measure ranks gained, now it’s used for comparison between servers.

So in the past, if your server had a lot of ktraining blobs gaining ranks, your server would be full.

If your server was full of a mix of PvP players and PPT havoc/small squads, your server was likely to be open.

Yaks Bend

in WvW

Posted by: Deemo.6094

Deemo.6094

Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.

Does this mean that a player who plays 8hours with exp boosters can count for a lot of spaces within the server? My question being, do exp boster effect server population?……………

Good point. Given that they measure ranks gained, they most certainly would skew the population measure.

Everyone stop taking experience boosters if you want your servers opened.

They used to measure ranks gained, now it’s used for comparison between servers.

So in the past, if your server had a lot of ktraining blobs gaining ranks, your server would be full.

If your server was full of a mix of PvP players and PPT havoc/small squads, your server was likely to be open.

You get a lot of wxp from kills trust me. Constantly farming groups gives a lot more wxp than ktraining. Then you can add +100% exp from kills onto that. It adds up fast. What im saying is even guilds that dont PPT get a lot of levels.

Persistence overcomes resistance. In the end, all things you work for can be yours, except in gw2.

Yaks Bend

in WvW

Posted by: McKenna Berdrow

Previous

McKenna Berdrow

Game Designer

Using exp boosters does not affect play hours and thus does not affect world sizes.

Yaks Bend

in WvW

Posted by: Aurika.6751

Aurika.6751

Using exp boosters does not affect play hours and thus does not affect world sizes.

It speeds of WvW rank gains, which were another metric that you mentioned.

Sladi [TW]

Yaks Bend

in WvW

Posted by: Chaba.5410

Chaba.5410

Using exp boosters does not affect play hours and thus does not affect world sizes.

It speeds of WvW rank gains, which were another metric that you mentioned.

Just because something was mentioned doesn’t mean you understood it.

“Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.”

Chaba Tangnu
Founding member of [NERF] Fort Engineer and driver for [TLC] The Legion of Charrs
RIP [SIC] Strident Iconoclast

Yaks Bend

in WvW

Posted by: Aurika.6751

Aurika.6751

Using exp boosters does not affect play hours and thus does not affect world sizes.

It speeds of WvW rank gains, which were another metric that you mentioned.

Just because something was mentioned doesn’t mean you understood it.

“Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.”

Thank you Chaba, I missed that part. I was basing my post off of, “Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.”

However, it being in the algorithm at all should be concerning to people simply because Anet breaks things all the time (red rings for friendly AOEs as a most recent example) and I don’t trust them not to break this and use ranks gains as part of the population calculation.

Sladi [TW]

Yaks Bend

in WvW

Posted by: Fatherbliss.4701

Fatherbliss.4701

Right, looking at the Rank gains in order to identify or isolate anomalies in the data would be a reason to look at that as a measure. Makes sense.

I actually think this new calculation is good. There has to be some sort of metric beyond just a gut level instinct. Total hours played will take away the illusion that is gained by assuming that just because a server has a ton of people that automatically means they are playing non stop. That isn’t necessarily true.

And yes, I agree with the above posters that in theory the server with the highest rank gains is hitting doors. Again though, that is an assumption.

We have arguments over whether or not servers are full, when that may not in fact be the best way to look at this. We want competitive matchups correct? That doesn’t always equate to which server is full. (It certainly helps.)

As with any change that is implemented, I’d like to see people give it time and see what happens over time.

Leader of Goats of Thunder [GOAT]
Tarnished Coast: Bringing the Butter to you (no pants allowed)

Yaks Bend

in WvW

Posted by: TorquedSoul.8097

TorquedSoul.8097

Using exp boosters does not affect play hours and thus does not affect world sizes.

It speeds of WvW rank gains, which were another metric that you mentioned.

Just because something was mentioned doesn’t mean you understood it.

“Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.”

be nice … the first dev post said this …

“Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.”

It was “clarified” later in the third dev post.

Yaks Bend

in WvW

Posted by: Rambitshouse.8712

Rambitshouse.8712

Using exp boosters does not affect play hours and thus does not affect world sizes.

It speeds of WvW rank gains, which were another metric that you mentioned.

Just because something was mentioned doesn’t mean you understood it.

“Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.”

be nice … the first dev post said this …

“Yak’s Bend is among the highest worlds in terms of play hours and ranks gained, which is the primary metric we use to determine which worlds we lock each week.”

It was “clarified” later in the third dev post.

I thought the same as Aurika in first reading, guess I’m only human lol

Thanks for the replies McKenna.

Dtox

Yaks Bend

in WvW

Posted by: Liston.9708

Liston.9708

so afk and trolls (one person spams inventory all the time) DO count toward hours played? How do you measure “hours played” versus “hours contributed”?

YB→YB→YB→YB→YB→YB→YB→most likely YB

Yaks Bend

in WvW

Posted by: Spartacus.3192

Spartacus.3192

So, I assume this means because YB has a few dedicated players that spend hours upon hours in WvW which would also assume they would also gain ranks at the same time doing things in WvW makes them highly populated but on the same level having servers that have lots of players spending a hour or so on WvW but not gaining ranks because they are not in long enough makes them underpopulated in your algorithm.
Does this not penalise a server for having dedicated WvW players?

So in theory, similar to a post above a server could have 500 players dedicated spending hours in WvW and then a server have 1000 players doing just a couple of hours, but the server with 500 gets locked.

I don’t understand why you consider that to be “penalizing”

Server 1: 100 players playing on average 2 hours each = 100×2=200 player hours

Server 2: 75 players playing on average 6 hours each = 75×6=450 player hours.

It is obvious that server 2 is more active that server 1 and as such could theoretically be out numbering server 1 quite often during the day as the server 1 players logoff after only 2 hours playing.

In this situation its fair to allow Server 1 to not be FULL and for server 2 to be Full. As people transfer to server 1 the player hours should increase.

Your typical average gamer -
“Buff my main class, nerf everything else. "

Yaks Bend

in WvW

Posted by: Artaz.3819

Artaz.3819

McKenna, can you provide any correlation to WvW track gains (points unmodified by WXP boosters) with hours played?

The concern is hours played includes a ton of afk players for specific servers as there is a common tactic to use secondary accounts to just stand around in other WvW enemies areas for spying and/or troll other servers (especially tags).

In essence, you could be letting the spys and trolls severely influence your current algorithm.

Yaks Bend

in WvW

Posted by: Kovu.7560

Kovu.7560

Yeah, Chaba, be nice. :<

~ Kovu

Charr Ranger, Necromancer, Thief
Fort Aspenwood. [CREW], [TLC], [ShW], [UNIV]

Yaks Bend

in WvW

Posted by: Taobella.6597

Taobella.6597

i can not speak for most sever i know from yak that people i started game with 1400 days ago still play today most of them avring around 7000 – 15000 hours LOL

Yaks Bend

in WvW

Posted by: SkyShroud.2865

SkyShroud.2865

To clarify further this is the first week we are using this new algorithm. So some of the complaints that are being brought up were problems with the old algorithm.

We use play hours to determine the size. Rank gains is tracked for comparison purposes since they usually follow a similar curve, but isn’t actually used to determine the world size.

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. The past algorithms weighed more heavily on individual players, so we ended up with situations where JQ was ‘Full’ because they had a lot of players, just not necessarily ones that played as much as Blackgate.

Is the pairing done base on the data from old algorithm or new algorithm as well?

Also, is this play hours weekly or bi-weekly or monthly? If it is weekly, it might not be able to account for surge of population regardless the causes of it.

Also, how is this new algorithm able to help with balancing given that even if players do play short hours, they could come online at the same time therefore giving birth to possible situation like bigger numbers over others in that time slice.

Edit: I am not talking about winning or losing in the scoreboard but actual gameplay. If there is a 30 ppl logging on for short hours at a time slice, they can easily roll over 15 people who log on for long hours. It might be fun for that 30 people to roll over the 15 but the 15 will be rolled over, emo over it then next hour, they ktrain and bored to death.

Founder & Leader of Equinox Solstice [TIME], a Singapore-Based International Guild
Henge of Denravi Server
www.gw2time.com

(edited by SkyShroud.2865)

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Is there a /age type command you can use in WvW to figure out your own player-hours for the match?

Yaks Bend

in WvW

Posted by: Chaba.5410

Chaba.5410

Yeah, Chaba, be nice. :<

~ Kovu

He’s a big boy.

Chaba Tangnu
Founding member of [NERF] Fort Engineer and driver for [TLC] The Legion of Charrs
RIP [SIC] Strident Iconoclast

Yaks Bend

in WvW

Posted by: Fridgemomo.3750

Fridgemomo.3750

lets solve this once and for all. First off all, take the relinking away just merge the bottom servers to the top ones already, waste of time relinking them every 2 months when people will complain about relinks anyways, and could spend that time on other things. Second of all if you ever want to make scores interesting even though there are population differences between servers is to bring in a handicap system. Based on the glicko differences in the servers in a matchup the lower server would get a certain amount of victory points per the difference in glicko between them and the top server in the matchup.

So say server A has 2000 glicko
server B has 1980 glicko
server C has 1950 glicko

I would say for ever 2 points difference in glicko the server would get 1 bonus victory point. So server B would get 10 extra victory points and server C would get 25. Now some times this wont work because large groups will transfer and it will take glicko some time to adjust, but if populations and the way servers play don’t change much this should bring closer scores.

Yaks Bend

in WvW

Posted by: Swamurabi.7890

Swamurabi.7890

Some more random thoughts:
1. Are you going to count player-hours when in spawn?
2. Way back before squads there were parties and I remember someone in my party logged out of GW2 from the WvW map and their blue dot stayed on the map for some time. Is there a difference in terms of player-hours between logging out of GW2 inside a WvW map and using a portal to go to Lion’s Arch?
3. Should you count player-hours when you are outmanned?

Yaks Bend

in WvW

Posted by: Justine.6351

Justine.6351

Probably because siege counts as population

this is 2017 not 2015…..

Yeah, but disloyal, bandwagoning “fight” guilds ignorantly perpetuating an old and now highly inaccurate stereotype, whatcha gonna do?

Lol are you trying to say YB doesn’t still hug siege? Because that’s not what I’ve been seeing this week.

Every server has those players that like siege. I see it in the same match up as you. I will say this about YB, on the first day that I came to YB, ( 2 Years ago ) up to now, YB has changed for most of us. We don’t use the amounts of siege that most of you think we do.

Ohhhkayyyy

https://youtu.be/jarKFnCQqhU