Guild Wars 2

solution proposals

prerequisits None of the following proposals violates the two essentials of WvW as set by ANet:
- 24/7 availability of WvW
- no discimination of player from any time-zone of the world

1) Weighting of the score by number of participants

Compute the mean total (sum all sides) number of players inside WvW within the ticker-time-frame. Multiply the ticker-score by this number (divided by a fix constant ~400 to scale to the current score-range) i.e. the more people fight for the points the more worthy is each point, but every player is equally important every prime-time player has the same weight as every off-time player.

2) Limit the objective/score by number of participants

a) Overal limit
Compute the mean total (sum all sides) number of players inside WvW within the ticker-time-frame. No side can get a score higher than this number.

b) Limit by map
Compute the mean total (sum all sides) number of players inside WvW on this map within the ticker-time-frame. No side can get a score higher from this map than this number.

This adds some strategic aspects about distribution of players over maps. A large blob jumping to a map with only a few controlled objectives earns less points than a force distributed according to already owned objectives.

c) Limit by objective

Each objective needs a minimal human player garrision (mean over time) to count for the score. The required garrision depends on the value of the objective. E.g. SM needs 35/5 = 7 people to count, a supply camp (inside the area of the circle) needs 5/5 = 1.

This adds some strategic decisions. Do you want to score and defend or do you wana do an offensive at cost of score.

To find motivated garrisions: Succesful scoring of an objective is an event with a reward comparable to defense of a tower, it is granted to all that where at least 2/3 of the time (10 of the 15 min) inside the objective.

d) Revert un-occupied objectives to neutrality

It’s unreasonable that a handful of people can control a large area. And if an area is not controled it rewerts to neutrality (like SM at match-start). So each objective needs a minimal human player garrision (mean over time) to stay under this sides control. E.g. SM needs 35/5 = 7 people to keep control, a supply camp (inside the area of the circle) needs 5/5 = 1.

To find motivated garrisions: Succesful scoring of an objective is an event with a reward comparable to defense of a tower, it is granted to all that where at least 2/3 of the time (10 of the 15 min) inside the objective.

3) Better and More NPC activity

Each siege weapon placed (or build by upgrades like cannon and mortar) is automatically manned by a NPC as long as no human wants to man it.
All these weapons do auto-fire to any target in range of the weapon, with some intelligent priorisation (e.g. treb prefers to fire on trebs in range).

To make this work the siege limits need to be separated by objective (3 per area, with the following total: SM 100, keeps 75, towers 50, camps 25), they must not despawn, and sieges in the objective must be under the control of commited people. Asimple ay to ensure this is: Only people from the guild of the claimer can place and remove siege weapons inside this objective. A guild claim expires if no buf is provided and not at least one player of that guild is inside the objective.

4) partition the 24h into time-slices

Just have different leagues (or ranking) with different matches for the 3 mayor timezones (running in parallel, such that there is 24/7 a match you can join)!
Every server plays in three matches in three leagues during every week:
the league of 0-8 UTC – called NA-prime,
the league of 8-16 UTC – called Pacific-Prime,
the league of 16-24 – called EU-Prime.
At 8, 16, 24 UTC the matches are switched (new match maps restore from last save in the background, player on map and in queue moved to new match, and then maps of the suspended match saved).
There will be 24/7 WvW, and and in each time-slice each server will be matched against equally strong servers in that time-slice.

Alternatively instead of 3 slices a 8h, 4 slices a 6 hours or even 6 slices of 4 hours can be choosen.

For variance with a special league a weekend-slice-league and a weekday-slice-league can be made.

The main goal is reduce match time to managable amount. If I build up and garrision my tower till match switch, I am sure that I will find it in that state when my slice-match will resume. This should improve commitment of players/guild to the match-slice they choose for the week.

No one cares about australians playing against australians or americans playing against americans or ….

The only thing that players hate is that the amount of australians on your server is decisive for your (non-australian) players to win, whereas your 10 times as large amount of NA or EU players does not matter et al.

Just separate the matches:
8 hour for pacific prime, 8 hours for NA prime, 8 hours for EU prime, switch match every 8 hours, still each match runs 7 times 8 hours.

Xenesis.6389:

You cannot penalize points or play times for players, it’s not their fault the other servers they face don’t have the coverage.

Where is the penalty for them in:

Another idea:

ANY possession without a minimal amount of players (e.g. points/level, e.g. SM 35/T1 = 35 people SM/T3 = 12 people) inside turn neutral instead of producing a score at each ticker.

You can roam the map at night, but if you do not place garrisons into your possessions they turn neutral and do not produce a score.

In fact it is an advantage for them, they still have something to capture, even if they would have reached 695:0:0 in the current system.

It just results in less influence of PvD-Karma-Trains on match outcome. No penalty et al for their game fun.

Chris.3290:

This is an international game and it sucks when someone beats you through PvD, but as fortunes change and populations move around; most of the non-NA teams are migrating to the top tiers.

And as anyone want to be on the winner side we end with 1 server having all WvW-people. Great matches in future 695:0:0 from start to end, a win-win for all, and no more losers.

And it even works as everyone will go to EotM then, where you can play balanced matches all around the clock.

Another idea:

ANY possession without a minimal amount of players (e.g. points/level, e.g. SM 35/T1 = 35 people SM/T3 = 12 people) inside turn neutral instead of producing a score at each ticker.

You can roam the map at night, but if you do not place garrisons into your possessions they turn neutral and do not produce a score.

Apfelcreme.5048:

kodash zone blobs at 4am? are you serious? We were like 10 people on EB until 12AM.

If these 10 are together it’s not only a zone blob, but even a match blob

Fingolfin.7451:

I agree with the first post 100%..Especially the 1st solution is what i was thinking a long time ago.It’s the best solution to the problem and it makes perfect sense!

I do not:

a) why these times? Main play times are different from match to match.
=> dynamic scoring based on no of players is much better.

b) why should 15 people still score 50% of 1200 people? they are 1.25% of people, 1.25% of score is much more adequate.
=> dynamic scoring based on no of players is much better.

Bubi.7942:

You can get 10 000 – 30 000 points in just one night. There are (estimate – my server) 150-500 different people playing in daytime. There are (estimate – my server) 15-25 people playing at night.

That’s exactly why I proposed to limit the score by the number of people online.
If there are 1200 people online (3 sides a 4 maps a 100 ppl) you can get the full score of 695:0:0.
But if there are only 15 people (total) online the max score any side can get per tick is 15.
If all three sides are sleeping (0 people are online) no one can get any point.

I do not see any discrimination of off-time players in that. In fact they are still privileged, they count for a full point whereas the primetime player counts only for 695/1200 point.

Currently we have a discrimination of prime-time players they count only for a 695/1200 point whereas an off-time players counts 10 to 20 times as much.

Evening performance is already irrelevant, what do you want to nerf on irrelevance?

Hule.8794:

Dayra.7405:

There are already more advanced methods out in the forum.

Namely to weight (multiply score by |players| / |capacity| ) or limit (required garrison, i.e. max score = |players| ) the score by the number of people in WvW. (|player| is the total sum of all players from all three sides)

Really? Anet dont adjust population counter and gem transfer cost on lower tier servers and you think they will implement this or any similiar solution?

Yes, they are in the forum.
Yes, ANet is ignorant with respect to them.

And yes, I keep my signature till this changes

Offair.2563:

i “work” all night with my teammates to get stuff to tier3 just to lose it in the morning to 20 ppl with no effort, so lets do something about that for people who play at night due for not having a 9 to 5 job

The only solution to this are the time-sliced (3 slices a 8h or 4 slices a 6h) matches.
https://forum-en.gw2archive.eu/forum/wuv/wuv/Night-protection-1/3281118
So every server play 3 or 4 matches in parallel. And after the slice of 6 (or hours the map is saved, and the other match resumes. So if you leave your tower at the end of a slice you are guaranteed to find it in the same state 16 (or 18) hours later when this slices is restored.

Offair.2563:

So you want to penalize people playing the same game as you but not on prime time, then i want those nightwatch and nightshift being used 24-7 then

God all these people and solutions, am glad they dont work for Anet phew

There are already more advanced methods out in the forum.

Namely to weight (multiply score by |players| / |capacity| ) or limit (required garrison, i.e. max score = |players| ) the score by the number of people in WvW. (|player| is the total sum of all players from all three sides)

This is fact does not penalize people playing in non-prime time it removes the current penalty from people playing in prime-time and makes off-time and prime-time players equally important for match win. Currently only the off-time players decide a match, prime-time performance is mostly irrelevant for match outcome.

A third way is to just separate the different time-zones in different matches, that are switched around https://forum-en.gw2archive.eu/forum/wuv/wuv/Night-protection-1/3281118

GD should be running alone on a borderland during prime-time under the week.
Some guilds left, some broke, some are in the process of reforming, some just have not the size/power to face the RS and/or Vizu zergs so they do public raids. And of course you should know (or at least remember) the problem on Vizu: On weekend and all time on EB, you are not able to form a guild-zerg if at least half of your guild hangs in the waiting-queues.

They spend to much time in the forum with SFR and Piken and captured their text block library

So new year starts like old year ended. Let’s have fun again.

Most interesting question of the week: Will Vizu be without Medal for the first time in history this week, or are the Elonian/Riverside forces to weak (and/or to busy fighting each other again) outside holiday to achieve this?

Zikory.6871:

Avatar of Belle.9623:

You know there is zero chance of TC fighting any T1 server, right?

That’s not true. While the chance is very low, it’s certainly not “zero” and out side of the current T1 servers has the highest chance to be subbed in.

Deviation (https://leaderboards.guildwars2.com/de/na/wvw) for both is below 160 so max roll is smaller than (0.45 * 160)+10 = +/- 82

Predicted new Ratings (http://mos.millenium.org/na/matchups) are:
Sanctum of Rall 2 088.944 -82 = 2006
Tarnished Coast 1 915.588 +82 =1997

So it is NOT possible that SoR rolls below TC. so TC has no change to roll into T1 and SoR has no change to roll out T1.

Hi Ken,

It’s Thursday !!

I want Elona-Vizu-Piken for next week

But it still looks like SFR is not winning, so they are desperately looking for reasons for that …
But when I read Ylias post again, my imagination suggest me a totally different story:

A drunk player mixing up former (SFR) and current (Piken) TS. Lets assume he sits in wildcreek as scout:
Scout: I need help 20 Kodash incomming in wildcreek!
SFR com: Aeh, wildcreek does not belong to us.
Scout: Shure it does I am in it! Please help!
SFR com: Ok we come!
Scout: Ahh help help, now the 20 Kodash got crushed by 40 SFR, come fast!
SFR com: hehe you called us, now we are here

What’s going on on EU servers?

Baruch Bay -9 ranks, -210 rating
Jade Sea -3 ranks, -120 rating

T1 was a protected area during the last weeks, but this week:

Piken -3 ranks, -40 rating
Vizu -2 ranks, -80 rating

SFR constant but mostly only due to the free pts Piken give away.

Many WvW-player lost motivation after league?

Last year it was SFR with the sylvester-night-zerk, neither Elona nor Vizu

Doesn’t look like at the reset anymore, but still it looks like Vizu is really lucky that Elona and Riverside prefer to fight for rank 4-5 instead for rank 1-3

NA has to do it, we already have our new matches in EU

Let’s share a thread for the last 4 days in the forum.

HF & GL to all 6

Btw. nice Prediction by MOS

Edit: As the others prefer their own thread I changed the title.

Junkpile.7439:

Some non DE or FR servers.

If you transfer to Vizu, you would get SFR and Piken, voila
If you transfer to Vabbi you likely get only EU-opponents as well.

If you stay on SFr you have to drop 100 – 250 ratings points AND roll a -1 – 0 to fight BB, Aurora, Deso, Gandara.

Just increasing randomness without making WvW more competitive with respect to something else than pure numbers of players will only lead to
easy wins for the WTJ and blow-outs (or boredom) for all others
And the size of this thread is still much smaller than the size of
https://forum-en.gw2archive.eu/forum/wuv/wuv/Server-Match-up-is-TERRIBLE
this clearly shows balance is more important than variance

Straegen.2938:

At the very least grade on a curve so there is always a Low/Medium/High/Very High. Why have several ranks if every server is High/Very High.

I never saw a low-pop server. I am in doubt that low-pop is a possible value. It sounds to negative.
Only medium – high – very high and full exist.

lunoh.1963:

Back in spring Anet introduced a very randomized matchmaking system, the number of different opponentswas very high. This system was introduced for a very good reason. The problem now is that the t1 servers are locked exactly like they were pre random matchmaking. So we are back to square one here in t1.

No we are not back to the original system. The original system locked the tiers, even for imbalanced matches due to a mathematical flaw in the design. This flaw is resolved.

But we are back to ANet original intention with match-making:
First priority is to generate balanced matches.

And I think it does so very well at moment. And EU-T1 is not locked, the random system can always roll the servers out. And if a T1 server looses power or a T2 server gains power match-making will react fast to it. The spring situation where Deso and SFr lost their power and were still locked in T1 will not happen again.

The imbalances in NA are even stronger (more concentration on fewer servers, as there are no national-servers that counter the WTJ-movement.). Therefore NA-T1 is temporarily locked. And league matches proved that this is in fact adequate.

If you want more variance you need to fight for a match-score system that allows more servers to compete against each other, i.e. reduce the dominance of coverage-wars on match outcome. (cf. offtime-capping, e.g. https://forum-en.gw2archive.eu/forum/wuv/wuv/Night-protection-1/3281118 which has more variance built-in every server always play 3 different matches at any day)

Looked like FSP has a strong night-shift. Maybe early in the morning (before Work/school) is the best time.
But this week is XMas-match. Things may be very different than usual.

Yeah. You should not have rolled ahead of Riverside, we all wanted Elona-Kodash-RS!

The EB Keep of the winning side (i.e. the side with highest coverage) is even harder.

@splatticus: In case none told you so far: Every spawn has THREE exists, just take one of the two others if one is besieged.

He may also have hit something before, that was now finally killed by other people or he repaired something and the defense event finished just now or trebbed a wall and the conquest just happened or ….

There are enough possibilities to get exp quite a while after action.

Lord Kuru.3685:

Zikory.6871:

Though keep in mind you will want to discard any soul bound items before deleting the character as other characters can’t pull them out of the bank and you will not be able to delete them (the Items).

That’s hilarious!

No, it’s not true.
True is: you cannot move them out of bank, if they are soul bound to a different char.
False is: you cannot delete them.
Just, right-click the item while it is in the bank, and choose delete item.

Still you may want to take them out before deleting the owning char, i.e. to sell them to a NPC or to scrap them. Both things that are no longer possible, when the owning char is deleted.

@Luna: I generally agree with your post: Variance is better than repetition.

But I must also say: balanced fights are better than domination.

The difficult choice is that between Variance and Balance. A pity that we have to make such a choice. ANet could simply nerf the coverage war to something more meaningful, such that we can have both Variance and Balance!

Third call for: Kodash – Elona – Riverside

I hope it happens before MU threads are gone.

Snowreap.5174:

it looks like the rating gap between NA #3 and #4 is now large enough that NA tier 1 has no randomness at all — it’s guaranteed to be BG vs JQ vs SoR.
…
-ken

Yeah with TC not winning T2 (-32 rating) and SoR not loosing T1 (+30 rating) this is probably adequate (Gap increase by 62).

Tribio.8531:

Mapchat/Teamchat can get the same results, no? Just enter a BL and mention on map “Looking for people to go cap us some towers/camps”. Depending on your server rep you’ll get some people from that map to join up..

If you are alone on an empty borderland or outnumbered defending the WP-keep at EB or …, it would be nice, if the lfg-tool could be used to to grab people from PvE!

titanlectro.5029:

ptitminou.6489:

titanlectro.5029:

Oozo.7856:

Zander.4763:

Welcome back guys!

http://www.guildwars2guru.com/forum/43-stonemist-castle/

~GW2Guru

Far too much moderation on GW2Guru. It was even more carebearish than these forums when it came to “insulting” other players, guilds, and servers.

Sounds great.

Seriously, an unmoderated match-up thread will be pointless.

No, it will be great. Every adult can read over and ignore the bad comments.

Yes, every adult can. But does every adult want too?

Why are you looking into MU-threads if you do not want too?

Please consider this one, if you reorganize the Forum.

Current population measures total number of players that declared this world their home since beginning of GW2, if they play during the last month/year or not is not distinguished.

This of course is totally unrelated to how many people actually play and even less related to how many people actually play WvW.
On the other side world-transfer is nearly meaning less for PvE people: they can guest everywhere for free.

Therefore transfers are mainly WvW-caused and the costs should be computed based on active WvW-measures:
- total number of person-hours spend in the last match (1 person on any map for 1 hour = 1 person-hour)
- total number of person-hours spend in the queue
- glicko-score/rank
- total WEXP acquired by people on this server during the last match
- sum of WEXP-score (at transfer) of newcomers during the last week
- ….

Any of them would work better to compute transfer costs than the currently used "amount of aktive and inactive (rotting in the catacombs of your server ) PvE+PvP+WvW-players " and a combination of several of them would be even better.

1) It’s the 3rd week and not the 4-5.
2) the 1st of the 3 was the last league match not random
3) random means within the range of competition, after all the movements to SFR and Piken (e.g. half Deso) these is just no one else the top 3 can play against in a fair match.

4) The best way to get out of this match is to undo the stacking. Go back to your previous server and immediately you are in a different match.

5) If enough people unstack, the competition (to T2 servers) may be restored and the random-system will let Piken and/or one of the others out of T1.

6) This is great! No one of T2 want to match any of the T1 servers at the moment. It would be much to unbalanced.

Admittedly this week NA rolled quite tier-aligned, but this is always a possibility and does not mean that it will always do it.
Admittedly NA has many holes in ranking. But if I remember the million posts against the random system everyone argued that these differences in ranking are adequate and matches over these differences will only result in blow-outs.

If I look at Ken’s probability list for NA
https://dviw3bl0enbyw.cloudfront.net/uploads/forum_attachment/file/125866/na1-20131213.txt
then nearly every server outside T1 has 4-5 server with probability over 10% to meet.
3 out of 5 are 8 different matches possible per server.

I don’t think ANet did over-do in nerfing variation.

The player did over-do in concentration on the top-server.

There is just no one that want to play against top 3, be it NA or EU.

I would say: beside 1-3 all servers (especially 4-10) are very happy that 1-3 are placed together
So ANet manages very well to make 7 servers happy and only 3 unhappy.
Additionally: a lot of people asked ANet to enforce better spread out of people.
From that point of view it is also good: put 1-3 together till the people leave their server to leave their match.

Cottontail.8526:

I just want to know why SOR and TC aren’t matched against one another as the rank 3 and 4 servers.

Given the current ratings, these 2 are the only ones that can take the 3rd spot in T1. (SoS and FA cannot “rollup” enough to be ahead a “downrolling” SoR)
But there is only 1 spot free (Neither JQ nor BG are able to roll out of T1) so it’s either SoR or TC in T1 and the other in T2. Therefore they cannot match.

One has to be the punching ball of JQ and BG.

Miyu.8137:

Kinda tired of this match already. Tired of endless VS blobs 24/7. Tired of them camping towers and keeps and not going out for fight unless they outnumber you 2:1. …

Lol, tired of reading the endless repetition of this complain now since 16 month several times per week

PS: Yeah I know I could avoid it easily by not reading T1 thread

“Explore their system” probably fits better than “exploit their system”

Lord Kuru.3685:

Yes. Monte Carlo is legit. But using Monte Carlo here indicates to me that they don’t really understand the math, that they’re just randomly tweaking numbers for a few examples until the example cases “look ok” — and that’s generally a bad thing. (I don’t think the math is so bad that you require Monte Carlo in this case.)

If you want to do correct math, things become very complicated very fast.
Let’s start with what’s the probability of a specific roll? Mathematically it’s infinitely small but > 0, on computers you don’t have infinite, so it depends on the precision of your numbers (and the quality of the random-number generator).
But you can also turn it around. How much precision in the random number matters? I would say only 2-3 digits do, lets assume 2.

With 2 digits precision you have 201 different possibilities drawing a random number between +1 and -1. All of them should be equally likely so the probability of any number is 1/201. (1/2001 if you prefer 3 digits precision. Testing all 2001*24 or 2001*27 possibilities, would be an exact and sufficiently precise estimation btw.)

An quite easy computable probability at moment is T1.
Is it JQ, BG and TC or JQ, BG and SoR? Other possibilities do not exist.
This depends on 2 dice rolls, the one of TC and the one of SoR.
If TC rolls negative and/or SoR rolls positive, the match will be JQ, BG, SoR for sure.
That’s already 3/4, only 1/4 left where SoR rolls neg and TC rolls pos.

The rating difference between SoR and TC is 108. To get the probability you have to count how many rolls bridge that difference.

SoR deviation is 161, TC deviation is 163. Lets make it a bit easier both are 162.

Roll_SOR * (162*0.45 +10) + Roll_TC * (162*0.45 +10) > 108
<=>
Roll_SOR + Roll_TC > 1.3

35% of all rolls satisfy this, so the probability of JQ, BG, TC is (both have the right sign) 25% * (the sum is large enough) 35% = 8.7%

This was (rounded) exact math to compute the easiest of all matches. Do you really want do this for all of them, most of them depend on many more rolls, than just 2?

But to access the quality of the system, it is usually sufficient to access the mean (all roll 0) => you get the initial match-making system, and whats the worst case (all roll +1 or -1), i.e. how imbalanced can matches be at worst. And these worst-case imbalances happen: BB-GH-AR is a nearly worst-case match.

Valik Shin.9027:

I’m not taking about different times of the day I’m talking about different times of the week. Some servers are really strong during the weekend and not during the week some are the inverse and some are just consistent in strength all week

A weekend match in the weekend ranking/league followed by weekday match for the weekday ranking/league would be a nice thing as well.

Valik Shin.9027:

This wouldn’t make the ratings any more accurate for the simple fact that some servers are much stronger at certain times during the week.

And if you look at my proposal, I proposed to have different rankings (or leagues) and matches for the different time-slices as I expect that a servers strength in NA-prime is quite unrelated to it’s strength in EU-prime.