Guild Wars 2

Crinn.7864:

Honestly my own research tends to indicate that ArenaNet’s setup is sound, however it’s plausible that there could be bugs.

I would mostly agree.
The only shortcoming I’ve seen is how it expands rating ranges: it keeps it static for several minutes and then grows quickly. A more gradual growth would be more appropriate.

Midi.8359:

So I thought to myself. Glicko / MMR / Elo, are all systems that really are designed to pair 1 person to 1 other person. They work well in that instance. However, since all those systems are based on one or two metrics (“win-rate” and time), they aren’t really good for pairing teams vs teams.

Many people say this. But it’s all based on feelings, not any hard evidence.

Matchmaking is working pretty well, but suffers from three things:

1. Low population at off-hours, especially for higher skill levels. The matchmaker will give players a poor match over no match at all.
2. Players are unable to gauge their own skill level. Many think they’re better than they actually are.
3. HoT builds have a lot of “first order optimal” design which cause players to hit a skill wall. When something a player has done forever no longer works, the typical response is to blame other factors like matchmaking.

Ithilwen.1529:

Forgiving builds are the solution, not the problem. PvP is cruising over the shoal because of matchmaking.

No, overly forgiving builds are most certainly the problem.

Matchmaking is awful because of low population. Low population is mostly due to boring gameplay where picking the right build and facerolling means more than playing well.

Similar to dragonhunter. Very strong for almost no skill investment, but top players can reliably counter its linear nature.

King of the Hill, Deathmatch/Annihilation, Capture the Flag, and other PvP models all share the same fatal flaw:
MMORPG combat doesn’t work well when you concentrate a lot of people in the same area. Someone is either spiked hard and the fight snowballs, or you have a lot of AoE spam.

The beauty of conquest is that it naturally divides teams over many areas with a simple design, while still allowing for larger fights. Stronghold tried to divide across the map as well, but the system broke under its own complexity and lack of tuning.

No.

If you’re a new player, it’s not really fun to get curb-stomped time and again by platinum+ players. Matchmaking keeps people of roughly equal skill together, which makes for better games overall.

G. None of the above.

Two biggest things which hamper GW2 PvP:

• Power Creep Power Creep with HoT specs and the resulting arms race has turned the game into a spam fest with limited builds at almost all skill levels. ANet needs to look at some core abilities and come up with a “power formula” for what an ability should do. Then skills can have similar power values without a handful dominating.
  Note that this is totally different from OP’s option D. The frequency is not a problem; it’s the complete lack of a balancing point and refusal to establish one and nerf skills down to it.

• Ranked Queue up to 5 Players The solo/duo queue is hurting the game way more than helping. “Solo queue mentality” – not caring about the team and playing primarily to kill other players, then blaming everyone else – is running rampant. It breeds toxicity and turns away players who can’t play with multiple friends.
  This is a team-based game; forming and playing as a team should be advantageous. The counter-arguments to this are full of non-sense. For all but the very top, voice comm and pre-made teams give almost no advantage over solo queue players. And top players are going to be top players anyway.

Lordrosicky.5813:

Exedore.6320:

Pretty much. HoT power creep killed any chance GW2 PvP had. What’s more dismaying is that the power creep is still at ridiculous levels nearly 2 years after HoT’s release.

Hot wasnt the main power creep. That was the trait merge patch where they randomly merged all traits.

If the power-creep was mostly pre-HoT, why do we see so few core builds?

They got rid of expereinced designers who understood risk and reward and understood how to create a game fun to play with and a game where other people were fun to fight against. They replaced these people’s power with those with no design experience.

Totally agree. There is no concept of power level for any abilities (“Jedi curve” for Magic: the Gathering fans).

Automated Tournaments are nice, but it’s a just a band-aid on a fundamental flaw, which is that you can’t do ranked arena with a full team. It’s really hard to 5 people to show up at a specific time and guarantee them only one game as a team. If you can’t play together regularly, how do you ever improve as a team?

The issue with using just a game limit for the season is that someone could play all their games in a couple weeks and just sit at the top of the leaderboard for the remainder.

The whole system needs some serious reform. Starting late in the season shouldn’t preclude you from a chance at a top leaderboard spot. If you can only play weekends, you shouldn’t have a decay rollercoaster. Is there any point in adding decay once you’ve reach the bottom of platinum (or if you’re below it)?

I would recommend the following:

• Remove temporary decay entirely.
• Cap the minimum games/season at 60 (4 weeks worth). For the first 3 weeks, it will be 15*(week number)
• In order to show up on the leaderboard and get end-of-season placement rewards, you must satisfy the current minimum season games and also have played 15 games in the last 7 days.

Changing the 3 day temporary decay to a weekly game total requirement gives more flexibility to when players participate. And it only affects people in leaderboard range. Lowering the season cap allows someone to start mid-season and still have a reasonable shot at the leaderboard. At the same time, it still forces top players to participate regularly.

OP is wrong. The problem of GW2 is not one of too many skills; it’s a problem of a handful of skills doing too much.

Elite specs are the biggest offender. Many of them were poorly designed mechanically, so their traits and abilities were over-buffed to compensate. When you just steamroll stuff with a few buttons, it doesn’t matter how well or poorly different pieces of it interact.

ANet also fell into the trap of creating an arms race to counter their over-buffed elite specs. Elite specs spam boons too easily, so ANet increased boon removal to ridiculous levels. Too many auto-defenses? Just increase damage on commonly used abilities so you can kill people between auto-defenses!

@ Acandis.3250

Careful when you say “broken”. Based on those ratings and what I recall from ANet Dev posts explaining the algorithm for team selection, the algorithm is working as intended.

What you really mean is that you don’t like the algorithm itself. But what ANet has chosen is the best in a bad situation. Think about it. You have two conflicting ideals:

1. Group players of similar rating on the same team. This way, you don’t have one really bad player dragging the others down. However in low population cases, especially at the top end, this leads to likely blowout matches.
2. Make each team have about the same average rating. In theory this should make the match more fair. But by offsetting a strong player with a weak player, it frustrates the higher rated player who has to make up for dead weight.

ANet’s solution was somewhere in the middle where they’d favor grouping players of similar rating together, but wouldn’t allow the average ratings of each team to get too far apart.

Now they can sell victory poses/emotes.

The second video is far more applicable than the first. The “moment 37” vs. Street Fighter 5 example is great analogy to what HoT did to GW2; with HoT elite specs, players can use skills poorly, i.e. spam, and still succeed. Similarly, using skills well is far less meaningful because of the shear number of defenses on top of automatic defenses.

The first video’s “buff, don’t nerf” mantra is misleading. What it’s really talking about is using added variety to address problems of numeric balance and fixed tactics. It’s what Extra Credits calls an “incomparable”.

What GW2 really needs to do is take a look at risk vs. reward on skills. If something is easy to do or widely applicable, it shouldn’t be that powerful. In comparison, something which needs setup, preparation, or timing needs to be more powerful. All too often, the designers just added more stuff to abilities which weren’t that good mechanically until they became way too strong with shear numbers.

sigh…

Frequency of balance is not the problem; the problem is scope of the changes.

HoT added a ton of power creep. ANet’s approach has been to make tiny changes to cooldowns or durations instead of re-working entire skills. And in many cases, they added more power creep to compete with the original HoT power creep. The result has been stale gameplay where you can spam and not be punished for it.

The writing was on the wall since HoT was released. Veteran PvPers immediately realized how brain-dead overpowered elite specs were from the start. It just took some a while to finally give up.

The dramatic drop in population from S5 to S6 is mainly a result of a dramatic spike in population from S4 to S5. For S5 they finally implemented a proper ranking system and that, in addition to easy ascended items, drew a lot of people back to PvP. But when the balance patch for S6 hit and it showed absolutely no progress in resolving the awful state of gameplay, those people just gave up again on PvP. If you draw a line from S4 to S6, it would be the same downward trend as the rest of HoT.

GW2 PvP is dead until the profession design team is “re-assigned”. They’ve shown that they’re incapable of making or unwilling to make the major changes that need to be made in order to save the game. I’m certain it’s the former, as we’ve had a serious of “what were you thinking” changes like daze on thief shortbow. There is absolutely no concept of risk vs. reward or trade-offs.

Hydromancy was removed because chill is a somewhat unique condition and it should be linked to profession abilities. Not only is it a snare, but it also affects ability refresh. If you still want CC, they put in an AoE cripple sigil.

However, they should remove Sigil of Doom since poison should also be linked to profession abilities and not given to everyone.

No.

HoT elite specs were so detrimental to GW2 (not just PvP) because they have overly strong damage and defense, and that defense covered for a lot of player mistakes. It used to be that you had to avoid burst or big attacks to survive, and people who didn’t were punished for it. Now everything does a lot of damage, so you have to have constant spam defenses. And because of those constant spam defenses, it’s harder to land those powerful attacks, and many are instantly marginalized by automatic defense. To top it off, the bonuses of many HoT elite specs are direct upgrades – no trade-offs at all.

ANet has two options to fix the problem:

1. Big nerfs to HoT elite specs. However, since many HoT elites still have glaring gameplay issues, this may make them too weak. Remember that many were bad during BWEs and had numbers (damage, duration, cooldowns) buffed immensely at launch to cover up lackluster gameplay mechanics.

2. Make at least one core line into an elite spec – probably the one that deals with F# skills like Virtues, Arcane, Illusions, Discipline, etc. This would be accompanied by small nerfs to HoT elite specs and some core traits and potential buffs to that new core elite line. In making that core line mutually exclusive with future elites, ANet has less to worry about with negative interactions, especially with elite spec focus often being on the F# skills.

If they didn’t mention anything, it’ll probably be 2-4 weeks from Tuesday.

The fact that they don’t even do a preview of what problems their addressing speaks volumes about the problems of the balance process.

Amaimon.7823:

or.. or.. hear me out?
or.. you make a build to counter the elite specs, or make an elite spec your own.
“Never whine about an enemy having a weapon, instead, overcome it with one of your own”

You’re missing the point. I’m pretty sure the people calling for removal of elite specs play them. Elite specs kill the fun and skillful play that the game used to have. The risk/reward balance on elite specs is terrible; it favors spamming and gives you too many ways to completely and easily cover for repeated mistakes.

HoT made skills more powerful and more accessible. PvP is pretty much dead now. There might be some causation there…

Because most professions in GW2 do not have a resource system for skills, the only limiter is cooldowns. If the cooldowns are too short, you no longer need to choose between using a skill or saving it for a better moment later.

The ESL decision has ZERO impact. Most ESL-level players quit in the months after HoT. The ones that won money stuck around until the money dried up months ago.

The terrible state of PvP is due largely to HoT power creep. It killed normal PvP and killed ESL-level PvP.

@OP:
Your proposed system is a bad option for evaluating rating.

Given three ways to evaluate rating changes:

1. Composite team vs. Composite team (what you’re suggesting)
2. Individual vs. opposing composite team
3. Individual vs. each opposing individual

The composite team vs. composite team option provides the least rating stabilization. Someone wrote a research paper on it.

The correct way to fix this is to prevent these huge rating differentials from ever happening with the matchmaker. Also, a lot of this silver vs. platinum reporting is because you’re seeing an opponent’s decayed rating.

Here’s how the matchmaker works in summary:

Note: A roster is who you queued with. It could be just yourself.

1. Take the roster at the front of the queue. This roster is called the “target” and is the base for selecting people for this match. The target roster will be in the match no matter what.

2. Starting at the front of the queue, create a list of “potential” rosters for the match which must all be within a rating range of the target. The range appears to be ±50 and starts increasing after 5min, but the rate of increase is unclear. If not enough potentials could be found, put the target roster at the back of the queue and start again at #1.

3. Start building teams. The target roster forms the basis of the first team. Continue the following until both teams are full:
a. Within the potential list (created in #2), find the best single roster to add to one of the two teams.
b. The “best” is determined by calculating the score of each remaining potential roster based primarily on time in queue, max roster size on each team, and how well adding the roster to either team keeps the ratings of the teams equal to each other. One point of rating differential is worth five seconds in queue (configurable weights).
c. Add the best roster to the team on which it scored best

One thing I noticed in the February 2017 edit of the wiki page (between season 5 and season 6), is that the rating screening for potential matches was changed. It used to double-pad and now it doesn’t. If that was a real change and not just clean-up of the wiki, it explains a lot.

Right now, from how I read the wiki, potential matches start out as ±25 from your rating. Before the edit it would pad both sides: (your rating – 25) vs. (potential + 25) and (your rating + 25) vs. (potential – 25). Finding a potential match within a range of 100 is a lot more likely than a range of 50. And probably just as reasonable.

Because the range is so tight, nothing happens for people near the top for 5min. Then at 5min, the floodgates open and the rating ranges expand dramatically. ANet didn’t list the Filter/Rating/@padding value, so I can’t be sure, but it looks like rating expands at 3/sec (so an additional ±90 rating by 5min30sec).

So if there aren’t 20 rosters in queue with 50 rating between highest and lowest, when you hit 5min you’ll be dumped into just about any match within a range of 200 rating. This seems to be the problem the top players are experiencing.

I’d like ANet to double-check the numbers and pseudo-code, particularly the missing Filter/Rating/padding value, but it seems like they need tweaked. I would probably set Filter/Rating/start to around 2min30sec, since the average queue time is just under that and this will only begin to affect the fringes. I would also increase the Filter/Rating/min value, but make Filter/Rating/padding more gradual.

It will do nothing but increase queue times initially.

Then people will realize they can queue on an unpopular class, get in, and switch to a popular one immediately.

In summary: a waste of time.

Ithilwen.1529:

That may be true or not. Regardless, the system punishes playing. Every game risks rating. The leader boards is solely based on rating.

It doesn’t punish playing at all. Fear of playing ranked is completely psychological. If you’re in the neighborhood of your true rating, then there’s nothing to worry about.

We should go back to the old system of number of wins on the season. That’s a measure of skill AND of diligence and it encourages people to play.

No, a system based on number of wins is a complete joke, as evidenced by ANet trying it in the past. All you do is tank your rating hard, then win back up to it. Then repeat. When you can just continually win in bronze/silver, it shows NOTHING. Further, because there are fewer players at the top, the matchmaker takes longer to find matches, so a win-based system hurts the good players even more.

The only way for a win-based leaderboard system to work is completely eliminate skill rating as a matchmaking factor. And if you remember s2 and s3, that sucked hard because high skill players would absolutely destroy people at bronze/silver skill levels.

1. Brain-dead HoT Power Creep
This is by far the biggest problem. HoT elite specs are far too forgiving of mistakes with excessive defense/healing and their damage is usually over-the-top in addition to that. There’s no more skill in pvp; it’s all spam. When you know your opponent is just spamming, but can’t punish them for it because of a pile of defenses, many of which are automatic, it’s just plain disheartening. On top of that, it’s killed variety in builds and team compositions.

2. Friends have mostly left
PvP is a team game, but when you don’t have anyone to play with, you lose some motivation. Many gave up because of specifically because of #1.

3. Inability to PvP with friends
Can’t queue as a team. In a team game. GG. Ranked is a bunch of solo queue heroes, which discourages healthy teamplay and breeds toxicity.

4. Poor Match Quality
This has little to do with the matchmaker and is mostly because of low population from #1-3.

5. No Hope for Improvement
HoT power creep is still going strong over a year after launch. The last “balance” patch was laughable and did almost nothing to address it. When it looks like the team can’t address the elephant in the room or even acknowledge it, there’s no reason to stick around and wait for it to get better.

crunchyraisin.6054:

You’re right that isn’t not an end-all fix. The only way to really engage high-end players is to have a competitive 5v5 tournament system with good/exclusive rewards.

Again, I don’t think that’s the biggest issue. For me, the biggest reason not to play is the ridiculous HoT power creep which makes the game simply not that fun. Too much spam and over-the-top damage and defense.

Not being able to play a team game with multiple friends is also a detractor, but not the biggest one.

You don’t need special bonuses. Just nerf mistake-friendly elite specs and then good players can carry on their own ability. The current problem is that elite specs are too forgiving of spam (bad skill timing) and have a lot of auto defenses or just plain spammable defenses.

It’s hard to carry when you get someone down to 25% and they heal to full in a couple seconds while having a damage amulet.

@crunchyraisin.6054

I don’t think that would fix the problem. PvP is also desolate at higher ratings, and those are usually people who PvP because they like it – rewards are secondary.

Conquest is far from a new concept. It exists in many other MMO games and works well.

How about we let players queue up to 5 players at once so they can remove the random element?

Kills people in low and mid skill levels. kitten annoying, but ineffective at higher skill levels.

If your rating is much lower than that of your opponents, you shouldn’t lose a lot for a loss.

However, ANet has never specified how they handle rating adjustments, but it’s presumably your rating vs. average rating of enemy team. Since the average enemy team rating is much closer to yours (probably 3-4 low plat / high gold), that’s why you’re still losing quite a bit.

There’s not much ANet can do. The population is just too small.

Cynz.9437:

Well gold vs legends does happen more frequent than it should. You wait longer than 5 min, such match is basically guaranteed.

So players in legend should sit in queue for hours, maybe indefinitely?

For the top end of the population, it’s not a question of finding a quick match – it becomes a question of finding any match at all.

Cynz.9437:

I don’t think you understand my point. I do have issue with silver vs plats WHICH is the thing already and is caused by attempt of getting quick queues atm.

No, it’s not. The rating differential is relatively small and only increases with time. So unless you think 5+ minutes is “quick”, you’re being silly.

Xillllix.3485:

Also they have a public testing realm with constant balance tweaks. Anet could certainly learn a thing or 2 from other games.

Jeff says otherwise:
https://www.reddit.com/r/Overwatch/comments/5yj2zk/ama_request_jeff_kaplan/dew569w/

Saiyan.1704:

Exedore.6320:

You realize the overall algorithm is nearly identical.

It absolutely is not.

I’m not going to explain as to why because I just need to point out to the #1 (some what major) flaw Class Specific Builds.

So which do you want? More factors or less?

ANet at one point set the “profession uniqueness” factor to zero. But it seems many people here want it jacked up really high.

Saiyan.1704:

Regardless, I read up on [Overwatch]’s Matchmaking and I want GW2 to be on that level of algorithms.

You realize the overall algorithm is nearly identical.

You could’ve just said “I want profession based MMR” and it would’ve been the same pile of nonsense.

PvP is in a death spiral because:

• HoT elite specs are way too powerful and too forgiving. They took decision making, i.e. skill, out of the game. It’s just not fun to play when everyone has a ton of auto-defenses which allow them to attack continually with no consequence.
• Solo/Duo queue was one of the worst decisions ANet made. It chased away players who like to play with friends and stabbed healthy team-based gameplay in the heart.

Fashion Mage.3712:

I think you mean selectively discussing it with people. Several people in this thread have brought up legitimate concerns numerous times and haven’t gotten a single response.

Just because they don’t reply to your post does not mean they didn’t read it.

Why does this seem to happen only with thieves? Why not mesmer, necromancer, elementalist, engineer, ranger, or even revenant? Does every stubborn person have a thief alt “cuz stelth is cool”?

You mean a matchmaking system? Like the one that exists in ranked?

@SOFYAORLENA.5849

Game looks fine, score was really close. Top stats was spread out.

There wasn’t any logic.

ANet just tried to gate it behind more stuff so it was harder to get. But most of those achievements don’t make it harder – just more frustrating.

LouWolfskin.3492:

But the community voted for a no on Profession mmr because it would be bad.

And it would be bad. Been explained multiple times in other threads.

The problem with class stacking is balance! Make DH less brain-dead to play and you won’t see so many of them. No need for obtuse rules to bog down the matchmaker with longer queues which don’t even solve the problem.

1. Design by poll is a terrible idea. They should do away with it altogether. The WvW polls are less problematic because they’re mostly “which do you want first”. The PvP ones have been of the form “one or the other”.

2. No reason for this. It’s too much work to get right for no benefit. If you’re at your correct rating, you’ll bounce up and down a bit, which perfectly fine. Players have to understand that rating is not an absolute; it’s better viewed as a range.

3. A way to teach players would help, but a long, in-depth tutorial won’t work. There used to be a mini tutorial (cap a point, stomp an NPC), but I think they disabled. You can’t force map rotations and teamwork.

4. Disagree. Number of tokens should be unlimited, but require continual play with bonuses for higher skill players. Like it is now. The problem is that the cost right now is too high. If you cap the major reward, players stop as soon as they reach the cap.

5. Class-based MMR will do nothing. People who play something they aren’t proficient with will drop. That’s fine. The argument that they drag down other players is bunk. There aren’t that many, and your opponents have just as much of a chance of getting them as you did. In the end it averages out and isn’t consequential.

6. Now that stronghold is relegated to the dust-bin, we can bring back conquest-specific dailies for a better rotation. Make ranked/unranked game a permanent major reward daily. That fixes the profession daily issue.

Saiyan.1704:

Can we have another on-hit condi sigil plez?

Pretty sure those are intentionally omitted. It’s hard to balance condition damage skills when they get access to more conditions through sigils and runes.

Just like power builds had burst damage sigils removed in favor of smoothing it out with +%damage or might and vulnerability stacking, condition damage builds will have to use +duration or might and vulnerability stacking.

Vanilla necromancer can work, but you need to play it well. I believe someone on EU got it to legendary last season. I think this is the build:
https://metabattle.com/wiki/Build:Necromancer_-_Vanilla_Condi_Necro

Evan Lesh.3295:

I’ve updated the original post with some updates. Please take a look!

Sigil of Enhancement
Unless you plan to greatly reduce boon spam with profession changes, this is essentially a straight 5% damage sigil.

Sigil of Ruthlessness
Please no on-interrupt sigils. Interrupt/CC spam from certain abilities is already a pain and they could stack this sigil insanely quick. An ele could wade through a group with shocking aura and trigger this sigil multiple times just from incidental cleave. Pulmonary Impact spam thieves will do even more damage.

Sigil of Absorbtion
Doubly bad. It has on-interrupt design and is boon steal, something which should be left to profession abilities.