Guild Wars 2

DanyK.3842:

There is no leaderboard or rankings anymore.

I see, you are really informed, can you tell me what’s this: https://leaderboards.guildwars2.com/ ?

DanyK.3842:

Any matchup can happen

Again you show superior knowledge of the game. Everything can happen for Gunnars Hold (not for Elona) but some are rather unlikely: https://forum-en.gw2archive.eu/forum/wuv/matchups/Potential-Matchups-6-21/2247983

How about: Vizunah Square [FR] – Baruch Bay [SP] – Gunnar’s Hold
No Germans involved, unfortunatly very unlikely only a 0.000042% chance to happen
Or do you prefer
Gunnar’s Hold – Fissure of Woe – Vabbi, much more likely with 0.363860%

azizul.8469:

perhaps the group selection should be divided further into 4 pools.
group 1 (rank 1-6)
group 2 (rank 7-12)
group 3 (rank 13-18)
group 4 (rank 19-24)

For anyone at the border of a group this is bad, e.g. rank 6, 12 and 18 will always get stronger opponents and rank 7, 13, 19 will always get weaker opponents.

As soon as the strength of some servers change the groups become inadequate

And last but not least: The ladder will disintegrate again, Glicko-2 ratings between groups will be incomparable, only showing the relative strength of servers within a group but not the relative strength of servers between groups.

Better would be: reroll until every match has a rank-difference of mostly 5 (or 4 or 6 or 7) 1-3-6 ok, 1-3-7 reroll all servers.

@DanyK: Why does a player on the rank 19 server cares about the opponents of the rank 2 server? The match was a nice intermezzo, but nothing to be repeated. We have to go back to defend rank 2 and play for rank 1.

Junkpile.7439:

I wanna face some real servers, not national servers.

You have to go down around 15 ranks to boost your chances to meat non-national servers a lot (top 12: 3 mixed vs 9 national; bottom 12: 9 mixed vs 3 national servers)

Someone else put them into:
https://forum-en.gw2archive.eu/forum/wuv/matchups/Potential-Matchups-6-21

Xerol.1578:

Sanctum of Rall (est. T1 after this week, ranking 2194.2, deviation 176.4) could not possibly roll lower than 2017.7, while Sea of Sorrows couldn’t roll higher than 1982.6 (est. T7, 1797.5, 185.1) so they could never actually match up.

That’s wrong and it’s in several of your possible matches as well, e.g.
SoR – BG – SoS 0.37%
SoR – TC – SoS 0.09%
SoR – DB – SoS 0.05%
SoR – JQ – SoS 0.04%

Here is a simple example showing how it can happen:
https://forum-en.gw2archive.eu/forum/wuv/wuv/Server-Match-up-is-TERRIBLE/2245792

The basic idea is: they do not have to overlap, it just must be possible that mostly one is left in between (and everyone else is able to “roll-out”). And of course match-alignment must fit.

Dragonbrand 1 958.001 rolls -200

Fort Aspenwood 1 874.367 till Maguuma 1 769.416
roll any positive number
Kaineng 1 687.742 till Ehmry Bay 1 655.349
roll at least +100

Borlis Pass 1 397.385 till Sorrow’s Furnace 1 113.774
roll -200 till not positve

Henge of Denravi rolls +200
and you will have Dragonbrand and Henge of Denravi in a match (over 800 points away)

Of course the more servers are in between the less likely it is.

Jong.5937:

Not quite sure what you mean here. Not quite sure I understand Anet’s leaderboard! But, to explain what I mean, I think the difference in real glicko of 2 servers matched up should not be more than 250 (ish!). So, if I understand their method correctly (which I may not) the adjustment for any one server should not be more than +/- 125 (or the method could be adjusted). If I am reading you correctly the current system could match servers up to around 500 apart.

Of course, this won’t help those at the very bottom of the table (and possibly the top) but there is nothing short of server merger or some very drastic changes to point scoring that is going to help them!

The real problem is that (not only due to the trippling in matches) in unlikely circumstances even a distance of 800 rating points can happen.

Assume everyone has a deviation of 160 (most have more), then add ANet 40 constant.
Every one can ROLL from +200 and till -200
Which means that 2 server 400 away can be rolled to the same number.

But it is even worser:
Assume server with the following ranks and with the following rolls
A 2000 +0
B 2000 +0
C 1800 -200
D 1401 +200
E 1200 0
F 999 – 200
G 600 +200
H 400 0
I 400 0

Would give:
A – B – D 599 difference
C – E – G 1200 difference
F – H – I 799 difference

So the worst case pairing in the middle is not only 2*(D+40).
Because servers in between (D and F) can “roll out” of the match.
It would work with any amount of servers un between B and D and between F and H.
Only if there are more servers between D and F the “worst case” goes down to a match with 800 difference.

So I think a (uniform) roll of +/-200 is far to much. But I like the general idea, possible fixes:
- Reroll all matches if a max rank distance (e.g. 4, 5 or 6) is violated in any match
- set D to -40 instead of +40
- Change rolls to gauss distribution (roll 2 dices in the range +/-0.5 and add them up)

jiddo.8041:

Meanwhile in eu 2/9 matchups are balanced. Last week 1/9 matchups was balanced, so yes great system anet…..

3-5-6 is not balanced because 6 clearly wins?

20-21-23 sounds very balanced as well, even if 20 dominates, but the surprise is 23 seem to beat 21.

24-26-27 is as balanced as it can be, would not be more balanced, if it would be 25-26-27

2-10-11 is not balanced, but all 3 sides like it. (as long as it is not repeated)

1-4-7 is not the dream match for 7, but they do well compared to 4.

Really bad are only 2:
8-15-16
and
9-17-18

For EU I would like to see the following Matches next week:
Vizu-Baruch-Piken
Elona-Jade-SFR
Kodash-Deso-Augury
Riverside-Abbadon-Millers

As long as the Zerg is the most effective way to let untrained people be very successful there will be Zergs.

Beat the Zerg and it will disappear

Complaining that it beats you makes it stronger

An All vs All results-matrix would make sense, e.g. like Elona has one just for Elona’s results on it’s homepage: http://myelona.de/ (titled “Season One”)

EU Start Rating are still only partially repaired on millenium.org.

Here the correct new Rank (1st Column), new Ratings (2nd column), Evolution (5th column) and if it happens Rank Change (6th column) from Thu 9:30 CEST:
1 Vizunah Square [FR] 2058.5159 189.1297 0.7602 -34.9699
2 Elona Reach [DE] 1946.2774 193.6014 0.7610 -40.3304
3 Jade Sea [FR] 1899.2935 183.4728 0.7494 +9.7789 +1
4 Seafarer’s Rest 1891.7438 172.6721 0.7420 -8.2514 -1
5 Baruch Bay [SP] 1879.6849 173.8715 0.7413 +38.2269 +1
6 Kodash [DE] 1842.2025 172.7100 0.7355 -29.2815 -1
7 Desolation 1807.4619 182.2936 0.7439 +22.2943
8 Piken Square 1754.0991 185.8030 0.7403 +9.8885
9 Augury Rock [FR] 1745.4565 187.7878 0.7351 +11.2028
10 Riverside [DE] 1733.2856 180.3342 0.7367 +12.4322
11 Abaddon’s Mouth [DE] 1689.6582 180.7611 0.7390 +21.2682
12 Miller’s Sound [DE] 1567.5631 172.5954 0.7309 +21.0421 +2
13 Gandara 1541.8530 172.8383 0.7342 -8.1870 -1
14 Aurora Glade 1537.4793 170.5015 0.7292 -12.4717 -1
15 Far Shiverpeaks 1502.5296 179.3928 0.7366 +2.6231 +1
16 Drakkar Lake [DE] 1493.2669 175.6798 0.7306 -11.6889 -1
17 Fort Ranik [FR] 1436.5372 180.3759 0.7337 -6.8133
18 Underworld 1417.1894 182.6890 0.7360 -3.2685
19 Gunnar’s Hold 1378.6821 185.4020 0.7574 -3.2744 +1
20 Ruins of Surmia 1333.7166 184.0650 0.7615 -70.0131 -1
21 Ring of Fire 1254.0551 185.0453 0.7542 +24.4671 +1
22 Dzagonur [DE] 1239.8275 178.3439 0.7378 -13.0427 -1
23 Arborstone [FR] 1213.1478 203.2605 0.8127 +26.3482
24 Whiteside Ridge 1183.9934 184.8667 0.7579 +48.1006 +1
25 Blacktide 1110.4259 231.9722 0.8227 -43.8148 -1
26 Fissure of Woe 680.5796 219.9253 0.8026 +47.5803
27 Vabbi 421.4563 231.2630 0.8198 -4.9862

Thanks for sharing your C# program, ken.

Edit: Now (12:00) millenium seems to be corrected. Thanks.

Historical rewrite? Which side does it?
Vizu’s and Jade’s, hm, “reduced aggression pact” is still a topic look into your match thread.
Even vs Deso where you hardly need it.

And there are a lot screen-shoots and even discussions on the french forum where https://forum-en.gw2archive.eu/forum/game/suggestions/WvW-Suggestion-Restrict-view-angle-in-WvW is shown.

Elona shows up. Riverside BL at 18:20 today: https://forum-de.gw2archive.eu/forum/wuv/matchups/Elona-vs-Flussufer-vs-Abaddon/280068
I guess we are the worst of the 4 top German servers in defense. But even we improved our defense in our 10 week match against Kodash.
So much that we were able to beat Vizu twice on our own, till they discovered the GW2-Window mode
https://forum-en.gw2archive.eu/forum/game/suggestions/WvW-Suggestion-Restrict-view-angle-in-WvW
and got Jade Sea as reinforcements.

Amelia Knox.9362:

Do you think that more numbers is a more skillfull tactic?

I think the more number vs skills argument is far over-stressed. And a to cheap excuse for a loss.

By the way “large zergs” aren’t that easy to handle.
- Keep them together as large fight-group and not as a large farm-queue isn’t without challenge.
- Be careful on every teleport the numbers halves (even with a large TS percentage)
- Having bad equiped low-levels inside, renimate the enemy faster than you can down them.
- …

Putting large numbers effectively into fights is very skillful. For the lead as well as for the players. Defending 4 maps with one “large zerg” is a very challenging thing.

For effective sieges if well defended position you need a 2:1 better 3:1 superiority and you must be fast, or the numbers turn against you very fast.

You may see only that whereever you try you get overrun by a large zerg, but if you see only the number, you overlook the challenge and effort it costs to be there in time.

Snowreap.5174:

Riverside and Abaddon’s Mouth will have their hands full this week playing Elona Reach (although if RS and AM are anything like Ehmry Bay they’ll be welcoming this opportunity rather than complaining about it).

We enjoy it as it’s a local derby, i.e. we can complain about each other in German.
In fact it’s more seen as German #1 vs German #3 vs German #4 than anything else.
But I am also sure, we don’t want to repeat it every week

Amelia Knox.9362:

When i lose a fair fight ….

You seem to be very young. Neither is Life fair nor games and especially Fights are as unfair as possible. Time to learn that.

And fights tend to be even more unfair, if you set own rules for it, that are neither the rules of the game nor the rules the enemy accepted.

We aren’t your enemy Currently none of us is. If you wana know how your enemy think you have to post your question into https://forum-en.gw2archive.eu/forum/wuv/matchups/Gandalf-Vs-Aurora-Vs-Millers

PS: And hopefully we never meet in a match in the near future it will be a T9 style match. We have nearly 500 ranking points more than you, a similar relation like Blacktide and FoW.

@Amelia Knox: Why should I leave our match thread if someone from outside the match asks match irrelevant questions?

And yes, ACs – as they are now – are needed. They give well prepared defenders a chance to defend the walls and doors against stupid brute-force attacks, i.e. against attackers that only use their mass, but not their brain.

Amelia Knox.9362:

Starting a new conversation, you think that siege adds something valuable to the combat experience? I don’t think so, i hate the Ac buff and Tup said many times this game will 100% better without door and no siege.

This IS the game, if you don’t like it you are in the wrong game

Raniri.6178:

And the winner are: All Servers who’re not playing against you

Yeah, you successfully hide behind the random-roll again, a pitty.
Still no opportunity to show SFR that they are only a little bit remaining “snow of yesterday”.

It’s not the match-thread , it’s only the international press-lounge for the match.
The thread is of course:
https://forum-de.gw2archive.eu/forum/wuv/matchups/Elona-vs-Flussufer-vs-Abaddon

Popcorn only for german-speaking

Whow a lot of international Attention for out little local-derby

Looks like millenium “believed” an API response from the reset-in-progress and by that trashed last weeks matches: http://mos.millenium.org/eu/matchups/history/76

The current start ratings on millenium have nothing to do with the real ratings at https://leaderboards.guildwars2.com/de/eu/wvw

Junkpile.7439:

Well you get more bags than ever. EB is full of viz players who die like ants.

And if you run out of Viz, you can kill the Jade reinforcements
I feel sorry for Deso, too

@kursk and mini:
Do you see a preference for night, morning, afternoon or evening PVD in this statement?

Dayra.7405:

But I do NOT like (hate?) that WvW is a game that by design favors PvD over PvP.

I guess you found the solution for a better (less PvD) match:

I believe morning capping is by far more profitable since it also provides with full upgraded towers and keeps that allows you to tank even more points all the way until you go to sleep.I just can’t believe how you chose or pretend to ignore such obvious facts.

Should morning capping be “by far more profitable” than night-capping you should switch your strategy from night to morning. By that you would turn the current: 2/3 PvD and 1/3 PvP match into a 1/3 all sleeping, 2/3 PvP match.

I would like it.

Tobias Steele.2071:

Actually in the current tier 2 match up sor players got bored and transferred back into t1 instead of doing pve and that was against the 4th and 7th ranked servers they would probably leave and never come back against t8 lol. No such thing as loyalty in tier 1.

Would be a success of the new system: spread out of the T1-population to create more more balanced servers.
But they better spread out to all servers (not no other overpopulated serves) to avoid creating the next unbalanced match that bore them.

MrsDelicious.4120:

So much hate from elona, and so much flame to the nightcapping when u guys do the same at morning. I will take the words of th few guys from elona that seemed correct and ignore all the haters that from some reason cant stand the fact that they are not 100k points over us.

If you interpreted hate against BB into my posts you are wrong.

I do not see it as BB faults that night capping is a very successful strategy, but as ANets fault. I only stated the fact that you use it quite extreme, and that it’s successful use is the main reason why you came that close in score to Elona as you did. But I did not judge it, and especially I did not judge BB for it.

But I do NOT like (hate?) that WvW is a game that by design favors PvD over PvP.
But as said: This is a “not like” against ANets game-design, and nothing against BB.

And i find the latin-americans the more charming explanation, than any other. And again it would be not BB or it’s players fault, they would be there because the like to speak spanish, a perfect reason, that does not need any justification. If it would be true, it would only open a question to ANet: Why isn’t there a spanish speaking server in NA, if 80% of spanish speaking live in NA-timezone.

SoR looses rating now and all former T1 server will do sooner or later, as the old system was stuck in it’s design error for much to long.
The last NA T1-T2 exchange of the old system happened in week 08.
It needed already both
A T1 domination of SoS by 325 705: 168 393 : 117 061
AND
A T2 domination by Blackgate: 302 661 : 152 840 :149 939
for a match change to happen.

The distance between T1 and T2 ratings accumulated to very extreme result already.
These extremely boosted ranks are just the artificial result of 9 month buggy old system, and cannot be defended in normal matches as the current SoR match show.

They will adapt now in the new system, and the order in which the servers are hit by it is random, but it will nevertheless hit all, some sooner, some later.

The worst on the “half second or so of invulnerability”, is that these invulnerable are still selected as 1 of the 5 AoE targets, even if there are countless vulnerable targets available as well.

Yes, if these are the only servers E & F surely play again any of A, B, C or D
Which depends on the roll of A, B, C and D.
A, E, F can happen (e.g. A roll -1 everyone else rolls 1)

They do not have to overlap! It must only be possible that everyone in between is able to roll out. And this can definitely happen if the servers are only 4 * (“min server deviation” + 40) apart.

And yes, if your constants (+40 instead of +200) are right, BG can currently neither match FC nor ET

That was the “patching the patch” downtime?

If yes, I can only repeat:
https://forum-en.gw2archive.eu/forum/wuv/wuv/Only-a-WvW-player/2119701

Snowreap.5174:

the math here is wrong.

I would say my math is correct, but somehow I got the constants wrong, so the worst case match is one with ( 4 times (EU deviation 170-231, NA deviation 171-193) + 40)): servers of 800-1000 points away.

Snowreap.5174:

BG cannot possibly end up with a match rating any lower than 1975 (2188 – 40 – 173).

Bg cannot be rolled less than 1975,

But it can match anyone up to 1375.
1375+ roll = 1575, everyone (all but one (i.e. blue) to be precise) in between can roll out (> 1775 rolls +1, < 1775 rolls -1)

Still a lot isn’t it?

MrsDelicious.4120:

You have no point saying that our nightcapping is because we have players from south america

Dayra.7405:

I only said that it is a likely explanation for your night dominance.

MrsDelicious.4120:

Also in the population numbers you write above, you are counting with Brazil which doesnt speak spanish.

In the list of spanish speaking I don’t have Brazil included, in the Internet connections I was to lazy to sort it out. So Brazil 88,494,756 internet users 58,565,700 FB account.
South-Amerika (minus Brazil, but still with Falkland Islands, (French & non-french) Guiana & Surinam) has around 100mio internet users with 70mio Facebook accounts

majos.8503:

Repuran.1859:

So the majority of BB’s points is from nightcapping non-europeans? Vizunah V2 then.

I think you didn’t read anything before you posted this…

He doesn’t have to read. A picture says more than thousands of words
And the new APi allows very nice pictures showing whats going on in matches:
(Source: http://gw2tracker.net/wvw/stats/live points per hour of the day, averaged over all days of the so far match)

@Snowstorm: Do you know the difference between a claim and a prove?

@Atoss: Maybe later when he gave a prove for his claim

@majos: I don’t want anything, you don’t like the rumor, I am happy with it

Snowstorm.3897:

And as usual, forum users only quote 10% of a post they’re quoting ._.

As usual I quote the part I want a say something about, and I leave the parts out for which I have no opinion. Quoting all – in a big blob- is quite useless, it leaves it totally open what you are referring to. And it was already written, why should it be repeated?

Snowstorm.3897:

…

Maybe, if you do not like this rumor, you can try to prove that it is wrong and instead of just saying that you don’t like it

I only said that it is a likely explanation for your night dominance.

Fact is you make the points only when we are sleeping and never when we are there.

Why don’t you make points on weekend?

Because you have to work on weekends or because we are there as well?

Yeah the probability is low, but that theoretical worst case example means that ANY pairing is possible with the current range and some (still bad ones) are less unlikely.

If Blackgate – FC – ET should be 1 in a billion
a T1 vs FC – ET is 3 in a billion
a T1 -any server – A T8 is more than 3×21×3 = 189 in a billion
…

If you sum up all the single probabilities of very bad matches you get the probability that each weak contains a very bad match and that isn’t that unlikely anymore.

Look at the average points per hour of the day at http://gw2tracker.net/wvw/stats/live
You only get more points 0-7am, from 7am-23pm Elona is ahead of you.

And this doesn’t need an excuse, it is just boring game design. But after month of denying it ANet at least recognized it for the first time in their latest note about WvW.

You get more points than elona from 0 and 7am and Elona get’s more points between 7am and 23pm that’s what the game statistics show. http://gw2tracker.net/wvw/stats/live

I would not call getting less points in EU prime-time when all servers have the same manpower on map more organized.

Snowstorm.3897:

Looking at the list of countries you posted there, do you really think they’ll be buying a good PC and GW2, when they have basic essentials like food/medicine to worry about?

I think you under estimate the situation in Latin-amerika. Most of them are second world not third. And even if they would be third world, they would have a lot of rich as well. (e.g. Carlos Slim, Mexico, is the richest person in the world according to Forbes)

Anyhow I looked it up: http://www.internetworldstats.com/stats.htm
Spain has 31,606,233 internet users with 17,590,500 Facebook accounts
South-Amerika has 189,982,457 internet users with 134,629,940 Facebook accounts
Central America has 51,452,595 internet users with 47,035,580 Facebook accounts
The Caribbean 13,480,693 internet users with 6,674,100 Facebook accounts

How many of them play GW2, only ANet may find out

Unfortunatly, theoretically this can happen (even if it is very unlikely)
Assume Blackgate rolls a -1 and FC and ET both roll a +1
This would give
Blackgate 2193 – 400 = 1793
Fergusson Crossing 919 + 400 = 1319
Eredon Terrace 879 + 400 = 1279

As 1793 is less than 800 away from 1279 everyone in between can “roll out”:
A server with 1400 points or more can roll +400 resulting to be out with 1800
A server with 1600 points or less can roll -400 resulting to be out with 1200

As said it is unlikely, but I think it should be impossible.
A possible roll of +/-400 is just to much! It means that theoretically servers with a rating distance of 1600 points can meet.

Already time to reduce the range I would say.

majos.8503:

Already clearly decided… Yes…

Yeah my fault, I forgot your night-crew (latin-americans?) only play under the week. But I see, it makes perfect sense, under the week the chance for opposition in the night is lower than at the weekend

PS: Only 9% of the Spanish speaking are living in Spain, whereas over 80% are living in America (north, middle, south, caribean) and BB is the only Spanish speaking server …

PPS: I always wondered why there is no Spanish speaking server in the NA, all in all there are much more spanish speaking (>400mio) living in America than there are English speaking (<300mio).

I did not wanted to predict a match outcome (even if the score expectation value does something like that) I wanted to express my concern that the rating will never stabilize to an adequate value that reflects the servers strength.

If such constellations occur the rating will not converge, but will keep a high volatility and deviation. The more servers are involved and in such a “circle of circumstances” the higher will be the volatility.

The actual rating of any of these server depend only on luck (who got a longer series of good matches in the recent past).

My last math-lecture is to long ago to decide if
- they converge to the same value with a very high deviation in the limit, or
- if they even swing around each other forever (build a Polyphase system) .

And the reason for that is: they do not have an absolute strength only a relative strength.

Snowreap.5174:

the algorithm is accurate by definition, because that’s the only purpose a “rating” has. it’s a predictor of probable score outcomes and if the actual scores don’t match the predicted scores, the ratings are adjusted appropriately.

I think this is a bit to optimistic. Let me make a simplified example
Server A play 0-8 (and only at that time)
Server B play 8-16 (and only at that time)
Server C play 16-24 (and only at that time)
lets assume dual matches for the moment.

If A plays against B, B should win 2:1 as it takes all from 8-16 and keeps it while the map is empty from 16-0.
So B should have a higher rating than A. If B plays against C, C wins 2:1. So C should have a higher rating than B. As rating scale is linear one would predict that C should be also much better than A. But of course when A play vs C A wins 2:1.

Problem in this case: better-than is NOT adequately projectable to an 1-dimensional ranking.

Another problem results from the 3-sided matches. The outcome of a match may be strongly influenced by the group dynamic in the match (2:1 or all vs all or 1 vs 1 with irrelevant 3rd). Any method that tries to reduce this group dynamic to server-rank necessarily fails, i.e. produces highly unstable results.

Hamham.7504:

Though I’m far too shy to even listen in. :<

You shouldn’t, it’s like listening a news/talkshow in radio

Zerg-radio for the win

Timpsu.9324:

We also have ridiculous queues on EB, I think we could actually fill a border sometimes with the people queueing to EB. It’s just that Desolation has a stronger history/presence in EB and people have the feeling that borders aren’t worth going into since there’s little to be conquered there.

EB is for randoms, BLs are for guilds (and randoms while they are in tkitten queue), isn’t that everywhere this way? At least on Elona it is this way as well. Your guilds need an recruitment officer on EB.

(lol, what makes the_ EB kittened )

/emote “text”
is visible to anyone (also opponents) in range,
but I’ve no idea if you can use it downed or dead.