Zypp.7921

First off, no class is bad in pve! Both elite specializations are good in their own way and have a very different playstyle.

With mirage you will find ways to generate as much endurance (the fuel for dodging) as possible. Start with a build [such as this one](https://metabattle.com/wiki/Build:Mirage_-_Condi_Mirage). It is a very safe, simple, ranged build because you get to dodge so much and dodging is a dps increase. It deals damage over time. You can freely spice it up by using scepter with either torch or pistol instead of a second staff. If you go for this, [you could eventually go full axe](https://snowcrows.com/raids/builds/mesmer/mirage/condition/) and learn how to play the best confusion based dps builds in the game, which is very good for certain bosses which fire of skills very fast. However, the axe build has a more complicated rotation and is only really used for a couple of raid bosses.

Chronomancer has a more complicated rotation, even for [a simplified open world build](http://https://metabattle.com/wiki/Build:Chronomancer_-_Phantasmal_Duelist). Chronomancer focuses on resetting cooldowns to spawn as many phantasms as possible in a small time window. In my opinion, chronomancer is one of the strongest elite specs for open world because it is very self reliant. You can give yourself powerful boons (quickness and alacrity) that normally require a boon support. This makes it one of the strongest solo dps builds in the game. You can also customize the chronomancer to give quickness to many people, making it a very useful build in open world events (it really speeds them up). Chronomancer is also very useful in end game content because of their dps, boon support and tanking abilities.

A berserker thief can carry the entire game with minimal damage and kills. Something like 200k damage can represent 8 unfavorable duels that the thief won. It could also be the amount of damage a trap dh did in three midfights that were won due to some critical interrupts and stomps by the thief.

You can get a lot of damage/healing by 1v1ing on a point that's owned by the enemy team. Which is not a good thing to do.

As far as I know, kills is just how many people died that you tagged at some point before that. How is a berserker roamer supposed to compete with a support for kills?

This 'feature' will just make people throw more because they believe they are the best on their team and so they will start playing for top stats instead of the win when their team starts losing.

I don't think implementing a system sophisticated enough to properly detect game throwing is worth the resources it will cost to implement. It would probably take a machine learning team a couple of years to do this.

I played a disgusting mesmer phantasm build before the trait rework in 2015 (which was a few months before HoT). At this time phantasms still stick around, attacking over and over until they are killed. They also take up one of your 3 illusion slots. This is where 3 inconspicuous little traits come into play:

1. [Crippling Dissipation](https://wiki.guildwars2.com/wiki/Crippling_Dissipation): Clones cripple nearby foes when they are killed.

2. [Confusing Combatants](https://wiki.guildwars2.com/wiki/Confusing_Combatants): Your illusions cause confusion to foes around them when they are killed.

3. [Debilitating Dissipation](https://wiki.guildwars2.com/wiki/Debilitating_Dissipation): Clones apply a random condition to nearby foes when they are killed. (bleeding, vulnerability or weakness).

The thing is. When you have 2 phantasms and 1 clone and you create a new clone, it always replaces the clone and keeps your 2 phantasms while triggering all 3 traits.

Rest of the build: rampager stats (power, precision, condi damage), blink, mirror images and decoy to stay alive and generate clones. I forgot most of the other traits but I did use deceptive evasion to create a clone on dodge. Staff and sword/pistol. Pistol gives a really good hybrid damage phantasm when traited to bleed on crit and the staff phantasm deals more damage for each unique condition (which is a lot). Staff and sword are great kiting weapons.

It may not seem like much, but imagine playing a core build against this mesmer. They don't need line of sight to attack you, they can hide and kite while phantasms do the work. If you do get close (especially as melee), you will automatically cleave clones and get cripple, confusion, bleeding, vulnerability and weakness. Both f1 and f2 shatters deal a lot of damage with rampager stats and can be used when phantasms are off cooldown or just to finish the fight. It's a very strong build for going far and winning 1v1's and kiting 1v2's for a long time.

After a ranked game I was stuck in battle of kyhlo. When the game was over I tried to leave ASAP. After the lions arch aerodome loading screen I got the battle of kyhlo loading screen again. Turns out everyone was there in battle of kyhlo again and we were all stuck. After logging out for 20 minutes, the daily reset happened and the game update happened (with the the thief and ele bugfixes) the problem was resolved. We also submitted ingame reports.

Sword/pistol staff viper mirage. Using dueling/chaos with the interrupt trait that immobilizes, grants might and a random boon and Condi. Blink, daze mantra and arcane thievery to strip stab. It's very high damage, high mobility and can lock down/kite quite easily.

A good chunk of DPS comes from your axe clones using the ambush skill when you dodge (the grandmaster mirage trait infinite horizon gives your clones mirage cloak when you get it, thus giving them ambush skills).

Generally we don't shatter unless we really need to breakbar with f3 or need f4 to survive a big hit. You can shatter to finish off trash mobs but it's not worth it for DPS to shatter during long fights.

It's better to give a few rules for doing your rotation than to give an actual complete repeatable rotation. It would be too hard to follow because endurance and dodges are part of your DPS rotation.

1. Try to weaponswap on cooldown to proc sigil of energy to gain 50% endurance (so try to also be below 50% endurance). By the way, you can equip just the mainhand in your first weaponset and just the offhand in your second weapon set. This allows you to weaponswap even though you only have an axe and a torch equipped.

2. Dodge and ambush as much as possible. Don't let any endurance go to waste if possible at all.

3. Try to finish your autoattack chains. It's hard to time this exactly so just try to autoattack for a few seconds at a time to at least get a few chains off.

4. The ambush skill has a short cooldown which stops you from ambushing twice in a row, it would cost some DPS to do this. Ideally you would autoattack for a few seconds until your endurance is full, dodge, ambush, axe 3, dodge, ambush and do a few seconds of autoattacking again.

You can take deceptive evasion in dueling instead of superiority complex (to create a clone every time you dodge). The DPS increase of superiority complex is very minimal. In my opinion, unless you know you won't have to take new targets you can always run deceptive evasion.

During the spirit phase you want to save your axe 3 and move on the next spirit just before everyone else and change the target of your clones with axe 3. That way you won't have to resummon them.

I believe the system is doing a great job making balanced teams, **based on the known ratings of players**. I also believe there are other flaws in the system:

Here's what I noticed when playing together with a friend who has a lower rating:

* When we win, the player with the highest rating gains less than the other player

* When we lose, the player with the highest rating loses more than the other player

Both options converge the rating of both players. This means that if I have a rating of 1600 and play exclusively with a friend who has a rating of 1000 we will converge to the same rating sooner or later. For example we might both be rated at a solid (high confidence) 1300 after playing together for a while even though I'm still a 1600 player and my friend is still a 1000 player. We part ways and now I'm set up for a winning streak back to 1600 and my friend is set up for a painful losing streak back to 1000.

Another thing is playing on a build or character the player is not used to. It seems the system does nothing to compensate for this. I.E. when I'm playing on a new class I get a bad losing streak. What happened to that? I though I read at some point the system uses a different rating for each class you play. If that were the case I would expect something close to a 50% winrate because I'm playing with lower rating players but I would be awarded a lot less points for winning while losing more for losing. There are a lot of players trying to do the 'year of the ascension X' achievements where you have to win multiple **ranked** games with multiple different classes. It's great that players are learning how to play different classes because I think it's making everyone better at playing their main class as well, but the system is not currently handling it very well when players are not on their main.

I know it's a balance team issue but I'm posting it here anyway. The TIMING of balance patches is hurting PvP. Please pass this on to the balance team or find another way to restructure the teams or the way pvp is balanced. The meta is not balanced often enough and at unfortunate moments. Last season you dropped an expansion mid-season. This season you dropped a balance patch at day 1 of the season. Whenever there is a _significant_ balance change during a season, some players will adapt faster than others. This means that the skill rating will be incorrect for some players and it will create winning or losing streaks before the system updates the skill rating of every player. It's also frustrating, I want to use the downtime between seasons to test builds and determine which build(s) I will main during the next season. I can't do this when the balance patch is dropped on day 1 of the season, this makes the downtime between seasons extra boring. Here's what I envision for balance patches for pvp:

* Keep the current schedule of balance patches, but move them up a few days. They should be at the end of the season or in the middle of the time inbetween seasons. This is a very easy and obvious change, I really don't understand why this hasn't happened a long time ago.

* Introduce a mid-season pvp only mini balance patch. This patch tweaks skill and trait numbers for pvp only and the pvp team should be in charge of this patch entirely. Players know exactly when and what to expect: for example in this season I would expect slight changes to scourge, firebrand and maybe mirage. This would motivate me a lot more to play in the end of the season and try to place well on the leaderboards. Currently I'm just not playing at all until the next balance patch because I don't want to slug through scourges supported by firebrands until the end of the season :(

There need to be improvements to team compositions with matchmaking. Currently the system limits each team to 2 of each profession. But it is still possible to get a team with too many roamers or bunkers. I'm very unhappy when my team has 2 roamers and the enemy has 1, or the same for bunkers. This is a tricky problem to solve, but a step in the right direction would be to mark certain skills with a mobility value or a support value. For example, thief's shadowstep and steal moves have a high mobility value. The matchmaking system can now make a better effort of balancing the total combined mobility and support values of both teams.

Finally, I suspect match maniplation is still a big issue in < 1600 rating games . I base this purely on the amount of disconnects I see. As a platinum 2 player I barely get any disconnects anymore, but some of my friends who are trying to break into platinum are still experiencing a lot of disconnects, especially on the losing team. There is no reason to do this unless you're working together with someone you know on the winning team.

With all of these issues in play, way too many matches are onesided even though the matchmaking system can correctly balance the skill ratings of both teams.

By the way, I love this game and gamemode to bits. I've been playing since the beta and I have more games of pvp than hours played in this game. Please don't take my comment for hate, I'm just trying to make pvp a slightly better place. I'm very glad to see these threads on the pvp forum and I hope we can have a civilized discussion on this topic. Thank you for everything!

Edit: Do you have any statistics on the average score of the losing team? I think something like "95% of games end with a score of at least 400 for the losing team" is a lot more meaningful than "95% of games have well balanced skill rating on both teams". I think this metric is a lot closer to perceived quality of the matchmaking system because balancing the skill rating is just one part of a good matchmaking system.

> @"Benjamin Arnold.3457" said:

> There's a lot of bickering and people slinging things as facts that aren't, so let me clarify something. Internally we want T2 fractals to be similar to core tyria dungeons in difficulty. T4 fractals should be the hardest 5 man PvE content in the game other than CMs, but notably easier than raids. CMs should be very close to raid difficulty, but not quite there (this is tough to accomplish). Saying T4 Amala is as hard as a raid is simply not true.

I did my first raid this week, just Gorseval. As a DPS class it was so easy to me: the boss has a very predictable pattern. Just read up on a guide and follow the steps. The difficulty of the raid lies in getting a group with a good composition together, assigning roles and attempting the fight until nobody makes any serious mistakes that mess up that attempt. The difficulty in a lot of t4 fractal lies in dancing around aoe's being spammed at me that will take over 50% of my life if I don't respond instantly, while also trying to avoid clumsy teammates trying to kill me with social awkwardnes.

They are different types of difficulty and will be experienced differently by different players. To me the easiest (and only) raid I did was far easier than the hardest fractals (among which the new twilight oasis fractal).

I'll take almost any offhand as long as it's good for pvp. Right now there's just torch and everything else is kind of meh because phantasms are so irrelevant. If I would drop staff (I rarely do) I would probably just use two torches.

> @"Wandering Mist.2973" said:

>your basic illusions don't add too much damage compared to phantasms.

The build relies on the mirage trait that gives your clones mirage cloak when you gain it. This also makes your clones use the ambush skill! This is the main power of the build and the reason you should never spawn a phantasm or use a shatter (unless you really need to distort).

Of course you can experiment with shatters. You could throw in an f2 if you're confident you can make 3 clones again before your endurance is full and weapon swap comes off cooldown. If you can shatter and get back to 3 clones without 'wasting' endurance it might be a small dps increase. Endurance=dps for this build, keep that in mind.

>@"Wandering Mist.2973" said:

>On a side note I'm not too sure about some of the chosen traits. For example, taking "duellist's discipline" seems kinda silly when you don't even use a pistol, and taking >"blinding dissipation" when there are no shatter skills in the rotation. Can anyone explain the logic in this?

The build takes dueling because it gives you the highest dps output, but the first trait choice doesn't matter as none of those traits affect dps. Feel free to take either the stealth or the phantasm fury trait as they have some situational use, while the pistol trait is completely useless for the build.

I and many other players are in favor of this, but it will never happen because:

* More work to keep it balanced

* Expansions are an incentive for competitive players to spend money

* This would spread the competitive population over multiple queues

It's too bad, because vanilla would be such an interesting meta. Less spammy builds, easier to see what's going on when there's not 4 scourges spamming aoes and sending conditions back and forth. Some of the most fun builds I played where from before HoT but that will probably never return.