Guild Wars 2

90 for me, please

166 for me please (main character: Jolli)

102 for me, please

Mausler.3748:

[about WULD = winner up, loser down]

• The next match-up is “earned” rather than determined by RNG; in many cases this means a tense fight for the result until the very end of a match (this could also be achieved by publishing the RNG values for the next matchup early(!) during the preceding match-up)

My own tier, EU T5, is a perfect example of this right now: At the current scores (142K, 152K, 145K), the sever’s ratings will change between -14 and +12. With the RNG, that means very little for next week’s match-ups, so the battle is quite relaxed.

With WULD, there would likely be a tense battle about who ends up first/second/third in this tier.

3, which I understand to be “winner up, loser down” (WULD).

The basic problem with the purely Glicko-based matchups was that there could be ratings gaps between tiers that were never tested anymore: if no server ever crossed the gap, there was no way to ever find out whether the ratings gap was actally justified (one might also call this “incestuous match-ups”).

WULD makes it impossible for such an untested gap to exist, because by definition, each tier boundary is crossed every week, so that ratings from the different tiers are tested against each other.

More pragmatically, here are more reasons for WULD:

• More variety than the old system, but more balanced pairings than the current system (could also be achieved by reducing the random factor in the current system),

• The next match-up is “earned” rather than determined by RNG; in many cases this means a tense fight for the result until the very end of a match (this could also be achieved by publishing the RNG values for the next matchup early(!) during the preceding match-up),

• The fight against your current opponents becomes more meaningful, as you are racing against the servers you are currently fighting, rather than server(s) elsewhere whose expected rating gets close to yours,

• No complex mathematics involved for the next match-up (the Glicko rating can continue to be tracked for server ranking and prestige).

fivekiller.1432:

as pointed out in other threads, this is not true. The WvW system is not capable of balanced matches. Nearly every match is a blowout.

According to the current Europe WvW scores, every tier but T9 is well-balanced, many tiers even perfectly so. So your “other threads” appear to be talking nonsense, as do you.

Dranul.2094:

If pugs keep following you, ditch them its simple to do.

Pray tell, what is wrong in having your leet guild group augmented by pugs (say 15 + 15)? Sure, it makes winning against 30 opponents less impressive on video, but apart from that, it seems not a bad idea to win battles.

[I noticed an attempt to ditch pugs last night on AG BL, but failed to understand it, as the opponents ran groups of around 30.]

Nemesys.7251:

too bad sems FS going down,comeon FS we love your lootbag fight and remain here with us.

Don’t worry – at the moment it seems we have four perfectly-matched servers competing in/for T4, and the one going down can be expected to be back the next week. I just don’t want my server to be the one going down

Rookni.2469:

I really dont hope people will play the way they did tonight. The numbers of people running together was just insane! FSP borderlands

You now can see what numbers the opponents run around with, and you don’t feel safe until you have found a similar-sized group of your own

Terrahero.9358:

Observe.
http://www.youtube.com/watch?v=xTTzzg7UBqo

Result:

Unfortunately, this video is not available in Germany because it may contain music for which GEMA has not granted the respective music rights.

Folks, you are not doing your recordings any favour by adding proprietary music. I know I can use some proxy, but that usually involves severe quality degradation of what I can view (bad resolution , sub-par player software, bad performance; often all of that).

Isn’t that what the Borderlands are about? Each Borderland being unbalanced in favor of one server?

CorliCorso.6254:

So the next time a load of bunny guilds to join a low-pop T9 server, rather than facing decent competition within two weeks or so they will instead have five weeks of easy matches & all the farming they like. Yep, great idea.

The last weeks in Europe have shown that with ratings, it takes several weeks for a recently-fortified server to rise and for a recently-deserted server to drop, so there does not seem to be much difference either way. We have seen jumps over more than one tier as well as rising servers stalled in a tier, as it took time for them to overcome the ratings gap.

Saweth You Him.9047:

In many matches the server that is first in a given tier often can determine who comes in second or third. They may place the weaker server in second and the stronger server in third. This becomes an issue in the top tier. This may tend to happen in any tier as stronger servers want to push strong servers away and hold onto weaker servers to move up.

In other tiers than the top this is useless to the winning server, as it will move up a tier anyway, so it does not matter much to them who stays in the old tier and who drops.

The kind of “griefing” you describe is possible with ratings too. If a server has nothing better to do than grief another server, I do not think you can prevent that at all. In fact with ratings, it might be generally more useful, because it may determine which servers you face the next week.

I am in favor of retaining a ratings formula for prestige; to use it for matchups exclusively may lead to stale matchups and ratings gaps across which no comparison takes place any more.

To stay with the chess example: the world champion is not determined by ELO rating, and he/she does not play the handful of players with the next-best ratings exclusively.

I would like for the “winner up, loser down” (WULD) rule to be applied for future matchups. Here are a few reasons in favour:

• Everybody gets different opponents every week, providing for variety.

• Your result against your current opponents becomes most important; there is no “remote competition” where you compete against the score of a server in a different tier. This may result in more intense matches.

• It avoids uncrossable ratings gaps. (Note that the ratings system essentially redistributes rating points between the competitors in a tier; if a gap is big; it cannot be crossed any more. For example, we had this problem in Europe T9 with two rather inactive servers at ~700 rating points; the third server in that tier (rated ~1000) had little chance to get out of it by good performance despite of blowout victories for weeks.) Such a gap can occur anywhere, and as a result there is no more comparison between servers on different sides of the gap.

To counter the criticism that it might produce unbalanced matches:

• New opponents too hard? Learn from them.

• New opponents too easy? Don’t lose in your preferred tier.

• At most, a server experiences 50% unbalanced matches. With uncrossable ratings gaps, a server’s matchups can become perpetually unbalanced.

The reset is a good point to start using this system, because the ratings will be unreliable for few weeks anyway.

• The anticipated silly matchups next week between winners or losers of vastly different tiers will be avoided.

After a few weeks, we can then review the matchups and decide whether or not to revert to the ratings-based system. (Note that a rating formula can be applied regardless of how the matchups are determined; just like in chess, where the ELO system is used in this way).

ColinJohanson.2394:

We’ve completed verifying every update from the november release, and there were zero changes to anything what so ever that would have negatively affected loot in any way/shape/form.

With 25+ years of experience teaching and practicing software engineering, I have found this (“who changed what when”) to be a decidedly suboptimal (albeit common) way of approaching a problem.

The code is a logical construct that misbehaves now, and it is a matter of logic to find out why it is behaving as it is. For example, veterans and champs give hardly any loot. Using breakpoints or traces, it should be possible to determine which code paths are being followed and which are not. Compare this with what should happen, and you will zoom in on the cause of the bug; that is, the place in the code where it deviates from what it should do.

For example, you might find that the DR mechanism had collected just enough data to kick in for many people around 16 Nov; which has nothing to do with code changes introduced on that date.

We can officially confirm this as an X-files level conspiracy at this point.

This is a rather offensive ad hominem in my opinion.

Here is another compromise solution: move all ratings r towards the center M of the ratings scale by a certain factor p; that is, set the new rating to be
M + p * (r – M)
With p = 1, this means no change, with p = 0, this means a complete reset to the mid-point M. So a compromise could be p = 0.5, meaning that all servers move half way towards the center of the ratings scale.

This way, the past is devalued by half; giving more weight to the future performance, but preventing the silly matchups a complete reset would bring the week after next.

My suggestion would be a week of 24-hour matches starting from the current rating, but using the “winner up / loser down” system for matching. With these matches, you can calculate the new rating, starting from 0.

During this week, you can also watch the “winner up / loser down” matching rule in operation. I think it provides for more flexible matchups, avoiding eternal battles and rating gaps that the servers cannot cross (example: Europe T9).

The wall east of the north gate of Stonemist has the same problem. Sometimes the stairs remain standing, blocking entrance; sometimes they break with the wall. I do not know whether further trebbing of the stairs would break them eventually.

Why not simply make the fractal dungeon x% harder each level, without requiring new gear? Then the elite can boast how high they can go, nobody needs to grind for gear, and no gear is devalued.

I do not see what this takes away from the current proposal, which seems to be to make it y% harder each level (y>x), but allow this to be mitigated by collecting and infusing ascended armor along the way. Instead of the skill-based progress described in the previous paragraph, you get gear-based progress, which seem far less attractive and impressive.

Since the increaed stats apply outside of the fractal dungeon, especially in PVP, the current plan introduces gear grind into the game, which kills the motivation of a large percentage of the current players, me included. Giving in to the content locusts is ultimately futile and will turn off everybody else.

My wish is that Anet delay the introduction of ascended armour, to give themselves some thinking time. There is plenty of new content to look forward to this weekend without the game becoming spoiled to feed the content locusts.

Thanks for implementing a straightforward way of removing the saved credit card info: In the trading window, press “Buy Gems”. You get into the window with payment options. Next to your saved credit card, there is now a button to delete it.