Guild Wars 2

QED, you didn’t understand how the system works.

The process is this:

- First, the matchmaker finds 10 players (regardless of their personal MMR) to put in a game, getting them from the +/-15 pips range.
- AFTER the 10 players are taken, they’re put in teams. And this is the step in which the MMR is taken into account, creating two teams which are always stacked: the best 5 MMR players on one team, the worst 5 MMR players on the other.

Already at this point the system is failing to create a fair match based on divisions.

But the consequences are even more dire.

Let’s say you have these MMR among the 10 players: 2000, 1900, 1800, 1700, 1500, 1500, 1400, 1300, 1200, 1000 (absolute values aren’t important, relative ones are). Since the MM puts players in team with “similar MMR”, ofc the 2 players who won’t go for sure in the same team are the 2k and the 1k guys.

After that, the 1.9k guy is the most similar to the 2k guy, so he will go in the 2k team. Same goes for the 1.2k guy, who will be matched in the 1k team. This process go on until you arrive at the two guys with 1.5k MMR.

And what happens now? Now the system isn’t able to chose among them, both teams have a slot left and no way to decide if guy A is better B or viceversa. So they are put in the teams at RANDOM. Let’s say A goes with the low MMR team, and B with the high MMR team.

Due to this randomness, A will lose the game with a higher probability than B. A will lose MMR points, while B will gain those.

What does this mean?

- B got, by pure chance, a more favorable chance of winning than A, even though his MMR is the same as A.
- B, after the match, will have a higher probability of getting put in teams with higher MMR than A, again by pure chance.

This is exactly the opposite of fair competition for all players involved, top, good, average, bad, and bottom of the barrel.

It’s the same problem that we had in S1, except it’s inverted. In S1, the MMR would be averaged out, so a good player would get matched in a team with the worst player, decreasing his chances to win (and here, moreover, all the kittened speeches about “carrying the game” falls, since this was the most whined thing about S1 apart from bunker builds). In S2, the system has just been inverted, giving the burden of getting the worst players to average ones, dooming them to lose a lot more than a good player in S1.

If you, and others, fail to see that this is the problem, then I don’t know really how to explain this to you in other terms.

Ah, ofc this problem is amplified by the small player pool for obvious reasons.

Unranked and Ranked have separate MMR.

AegisFLCL.7623:

Stop making assumptions that the system is: luck based, rng, purposely stacking, doing anything in its’ power to keep you or any of these other players from the progress you’re entitled to.

When you or anyone else makes these statements based on assumption, they are ones ignoring logic and fact. Not a single user has provided player wide data supporting their claim that MM is “broken”. I see plenty of good solo Q players in legendary and diamond, and I’ve seen plenty of bad “primordial legend players” stuck in sapphire and ruby.

I’m not making “assumptions”. I just read and understood what has been stated by Anet in this blog post:

https://www.guildwars2.com/en/news/pvp-league-season-two-changes/

If you want to continue to discuss the argument, please first read the blog post and understand it. There’re no assumptions here, the problem regarding match making is explained perfectly there. If you fail to understand it, it’s your problem, not mine.

AegisFLCL.7623:

It’s a competitive ladder.

Your team is grouped by your MMR, which you share closely with your teammates. Enemy teams are based on pip range and have nothing to do with your MMR directly. Better players have a higher chance of winning, and subsequently should have a better MMR. As the season progresses players will be fleshed out by skill far more accurately than S1, and the ladder as a whole is a better representation of skill.

Stop ignoring the logic behind the system if you want to argue why or why not a competitive game ladder does or doesn’t work.

p.s. +1 to Branny

The one ignoring logic it’s you. Stop being so obtuse by switching cause and effect.

brannigan.9831:

Yes, its totally fair to have all the good players on one side if we want league placement to be meaningful.

This is one of the most stupid argument I’ve ever read. A good player is good because he/she rises to the top without a match maker who favours him by stacking the games.

If he/she can’t “carry” an averaged MMR team, then he/she’s not that good. At least, that’s the “good” players argument.

And yet, they whined and protested during S1 for being teamed with lower than average players, but in S2 everything’s magically good, cause they don’t have to carry and/or lose games due to a stacking algorithm.

You’re at Trump-level logic, truly.

You know that you’re getting older than you’d like to admit when you find that your AMA question has been answered, but you didn’t see it directly on reddit but in a copy/paste (nice) work on the official forums. :facepalm:

fishball.7204:

I mean I get Chaith and Caed ON my team sometimes and other times I get total dweebs. Either they expanded the MMR range heaps or something’s weird.

What division are you at and how many pips? Dunno if they both already reached legendary div or they are still in diamond, but I guess you are a diamond yourself or in the last tiers of ruby.

Sirbeaumerdier.3740:

If every player “worth their salt” that you know had a positive MMR with this algorithm they were never fed garbage by the current algorithm. In other words, what does it prove at all regarding the issue ppl have?

Well, it proves exactly the fact that the algorithm is designed to stack teams. It’s funny how these white knights of the “skill play” show the analysis capabilities of a drunken mussel.

Aenesthesia.1697:

mm is not setup to give anyone an advantage. You are playing in the same division other people are, so you play against them all.

Team MMR != Division position.

It’s appalling that some of you haven’t yet understood this simple concept.

Cyanon.1928:

Lol… Will ANET even pay attention with the billion posts about WvW and sPvP problems?

Will they see the Elephant on the room or will they wait and silence until they lose 50% of the players before considering changing their strategy…

There, some more shop sales for you!

Pakkazull.6894:

But where was the person you were duo-queuing with at?

He was at T3.

Hammerguard.9834:

I didn’t think Anet could make a legendary so ugly.

Based on past experience, I wonder how you could say that. Most legendaries are ugly as sin lol.

InsidiousWaffle.7086:

I’m not sure what you mean, the system already finds 10 players within 2 minutes, so I dont think that will be a problem

Yeah, by widening more and more the pips-range from which to choose.

Example, the other day I played a match in duoq, my team was a double duoq+soloq of all sapphires (I was at T4, 15 pips total), enemy team had a duoq (ruby+sapph) and 3 soloq (2 rubies + sapphire). So, at best, 3 of their players were +10 pips from me.

So, in order to shorten the queue times, the system widens the accettable range from which to choose. If you limit this range to just 3 pips (0/+1/-1), queue times will skyrocket, we don’t have enough players.

Have you, moreover, noticed how many times you get the same people over and over in your matches, even for 3-4 consecutive games?

Not, enough, players.

InsidiousWaffle.7086:

Of having a matchmaking system where it takes 10 players within a pip range, and tosses them into a game without any manipulation?

This would be a much more fair system.

However, it’s impossible to apply to GW2: not enough players, by a long shot. And the numbers are going down every day apparently.

Sunshine.5014:

Yes, I have had bad teammates, who after the game (won) I asked, he said that was his 2nd match with Reaper (explained the bad positioning). But overall, it’s better than S1.

Must’ve been a balanced matching then, who would’ve thought.

To answer OP: for the moment I’m just playing the 3 daily matches and that’s it. I’ve better things to do in life than playing vs a stacking, kittened algorithm. Like ramming my head on the wall corner, way more enjoyable.

Personal performance != personal score.

If in a match you’re bunkering on a point because of your and enemy comp, for example, you won’t get the higher score. Conversely, if you only go around and skirmish outside of points instead of defending/decapping/capping them, you could have a high personal score, but your personal performance should be assimilated to “kittenty player”.

Loboling.5293:

we live life to win.

And yet the prize is always the same: death.

Crinn.7864:

5v5 shows the sum total of the skill of the individuals. True skill is represented, it’s just not absolute.

If you actually look at big game changing plays, you’ll notice that those big plays are often executed by single individuals.

You’ve read too much Nietzsche.

Malediktus.9250:

It is strange that Anet said nothing so far. Wish they would communicate wether they change the matchmaking or not.

They demonstrated over and over how obtuse they are when confronted with their (numerous) failings. Why should this instance be any different?

Fivedawgs.4267:

So what exactly are you trying to prove with this one?? There is no way to tell someone MMR based on this form. The fact are average/ high skilled player with low MMR where paired with noob at the start of the season while others (with high MMR) where place together; and each group had to compete against one another. Of course a good player will always catch up. However, ANET favoring one side over the other was wrong.

So there us no basis to collect data, i.e. a noob with high MMR who got carried all the way to ruby; who is now having trouble reaching diamond or experiencing loss streak in ruby.

So what would be your interpretation of the data?? That this dude belong to RUBY??

You see where I am going?? There is no need to collect data because :

1- the independent and dependent variables are not the same for everyone.
2- ergo, we are dealing with the luck based system.

Honestly, I’d just like to know if there’re other people who are experiencing the same that happened to me (read my previous post), i.e. going from uber high win rate to absymal kittenty win rate just stepping into a new tier of the same division.

That is already an indicator of how bad the matchmaking is, a good matchmaker should give you very high win rate at start until you approach your theoretical skill cap, the more you reach that cap the more you start to lose games, up to the point (cap) where you are around a 50% win/rate. But it should be a lot more continuous (like a math function) and not so steep.

Sorry I don’t have the screenshot, however the last game I played last night had this composition:

- My team: two duoq, one soloq, all sapphires.
- Enemy team: one duoq (ruby + sapphire), three soloq (2 rubies + 1 sapphire).

Ofc it ended very badly, 502-62.

On top of that, one of the duoqueuers refused to switch class since “I’m already on a loss streak, so I’ll play randomly with no regard to team comp, I don’t really care”.

Meanwhile, the population is constantly shrinking more and more due to this system, nice job Anet.

Edit: at that moment, I was at 0 T4 sapphire, so exactly 15 pips into the division.

1) Started playing sunday 28th
2) ATM Sapphire Tier 4
3) Honestly can’t compute this, I didn’t write down all the game played in this season
4) I’m stuck at Tier 4, progressed very quickly from Amber to Tier 3 sapph
5) I’m experiencing the very, very bad matchmaking – i.e.: I had roughly an 80-90% win ratio from amber up to sapphire T3, as soon as I hit the sapphire T3 my win rate went down the toilet, around 30%: this shows how much the use of MMR for composing team is flawed, you can’t go from very good to very bad just when you go from T2 to T3, it’s not like you jump from that tier to, say, diamond div.
6) Probably I will a little, just to hit ruby, after that I’ll just do my 3 dailies to complete part IV of the achies and then quit altogether, there’s no fun in playing stacked games.

Oh, and btw, since we are stating the obvious (at least, obvious for half-brained people with an experience of at least 5 games):

when you’re fighting large battles (3v3+, usually at mid at start), you don’t need to be ALL at ALL times on the point, you just need one of the team at a time to prevent enemy capping – stop camping the point all together and eating all the enemy’s aoes.
Conversely, don’t go all together chasing that one enemy who is off point – you’ll give the point to the enemies.

Let’s add one thing: if you’re down and there’s your scrapper teammate around, try not to constantly use downed #1 when he uses the stealth gyro to avoid you being stomped.

Removing leap from Bound would lower the build diversity: with D/P we’d be pidgeon-holed in using Dash.

Bound gives us more ini/dmg/stealth uptime by giving away a lot of mobility given by Dash. If you remove the leap finisher from it, we would lose the higher ini/stealth, remaining with just a bit more damage: we then you’d ever want to go with it when using D/P instead of Dash?

TL;DR:

• Bound: more ini/stealth/dmg, slow as kitten.
• Dash: less ini/stealth/dmg, great mobility.

They’re ok as they are. As opposed to the utter useless SA/Acro traitlines. :P

Straegen.2938:

A DH of equal skill on a decent build will generally wreck a thief at least in 1v1.

So much this.

Anyone claiming that he can beat regularly DHs in 1v1 with any thief build, has faced only really bad DHs.

Staff, definitely.

It all depends on what kind of role you want to lean towards, decapper or +1.

For the decapper, ofc Dash is the way to go, permaswiftness and a longer dodge animation let us rotate a lot faster.
For the +1, Bound increases the dmg output so less TTK opponents.

Anomaly.7612:

Really great combo but is it worth using the dodge offensively when a heart seeker works just fine?

I suppose it’s a good backup when you know you can be more offensive and sacrifice a dodge.

Nope, for several reasons:

• HS can be interrupted, Bound can’t
• For targets with 50%+ HPs, HS has a coeff. of 1, Bound of 1.38
• After using Bound, you have a +10% dmg buff for 4s, thus increasing the dmg output of Backstab and one full chain of AA
• You don’t use 3 ini to complete the combo, just 50 endurance

Ghorman.1368:

Is there an add-on like a DPS meter? how are you guys measuring it?

http://www.gw2dmg-meter.de/

Evan Lesh.3295:

The change wouldn’t have actually made class stacking better unfortunately. What was fixed was it breaking other matchmaking rules.

Can you elaborate on this point?

Before and now we get these same class-stacking results, so exactly what was changed and why?

Because that’s why it’s programmed since beta days. ^^

omgdracula.6345:

So if I am correct. Black Power -> Dodge -> Steal right after stealth takes effect.

No, as I said, you need to use steal WHILE dodging, so Mug can proc too before the dodge animation is over and you inflict both Mug and Bound dmg and end the combo in stealth, rdy to use Backstab.

Edit: and if you can, try to use Black Powder with the target already picked, so if he has Aegis you rupt it with the BP skill, allowing Mug+Bound+Backstab to land (if not, Mug will be blocked by Aegis).

omgdracula.6345:

Would it be cool if I added you? I did not know you could do some of these combos. Or if you could explain them quickly that would be cool too.

If you’re asking about the opening combo, you just need to use Black Powder, then dodge (with Bounding Dodger GM trait in DD line) and while dodging use Steal to the target, so you land the damage of Bound while going in stealth on top of your target, and after that use Backstab.

Changed matchmaking to not favor rosters with unique professions and instead disfavor rosters with duplicate professions. This prevents other scoring factors from being obscured.

Same old crappy results.

Is that hard to put 1 necro + 1 ranger + 1 thief per team and the other remaining players at random?

Sorry, but apparently you have no idea how to boost your DPS while raiding. :>

Staff is the way to go, but NOT with just autos, the proper rotation using AA+WC+Bound+Vault is needed.

Before the last patch I was easily reaching around 28-29k DPS on the VG (with the proper group/buffs ofc). There’s nothing like staff in pve raids to boost DPS.

Sir Vincent III.1286:

Azure Prower.8701:

For what situations to use Staff or D/P in PVE.

Staff all around in PvE. This is not even a question.

This. There’s no contest vs nobrainer mobs.

Evilek.5690:

Ofc no data lol

Today bug fix
Staff Autoattack—Staff Strike, Staff Bash, Punishing Strikes: Increased the damage by 36%, 15%, and 33% respectively.

Maybe first time learn your class and make test before you start write someting

You have no idea how to conduct comparing tests, this pic proves it.

I played a bunch of matches in unranked (so anything should be taken with a grain of salt, the unranked fauna is pretty heterogeneous, you can get good enemies/teammates and complete clueless ones too), trying mostly d/p and staff.

AA buff aside, it felt like the core problem of thief is still there unchanged: we’re paper thin and can’t really sustain mere seconds in aoe spammed points without disengaging, and the same is when we’re focused even by 2 enemies. Conditions still are a huge problem imo, but I haven’t tried to equip Acro with the new anti-condi traits (still not sure that the new Pain Response can be good and I can’t decide in case what to drop in order to get Acro).

If you tried some matches, what are your impressions? New patch, same old same old?

While I was reading the patch notes, I thought the same as you, I guess I’ll experiment somewhat with Bandit’s Defense.

Impact.2780:

I’d forgotten about that change to hard to catch. In this case it can’t be known because it would depend on whether the game engine combines the modifiers with addition or multiplication, with or before applying them to, the base rate.

Either way, how much the final modifier exceeds the base regeneration rate (or how much the final value exceeds 2*Base of the base regeneration rate) is of little importance if it gets chopped down to a modifier of 100% (or a value increase of 2*Base) before being applied.

Read my post before yours. :P

Sir Vincent III.1286:

Yes I get that, but let’s assume that I’m under the effect of Hard to Catch where I already have +100% endurance regen. Does Vigor makes it +150% regen? Does Vigor + ES makes it +175% regen?

I just tested Hard to Catch vs Svanir, it works exactly as Signet of Agility: when you suffer a Hard-CC, you get +100 endurance (not 100% end regen).

It doesn’t matter if you have Vigor or not active.

babazhook.6805:

Well one can also trait rrevealed using staff and s/d. S/d has a stealth built in via CND , the same as dagger/dagger.Staff is much harder to be sure meaning one has to go SA or take stealth type utilities. Hidden thief with knockdown followed by revealed as you leave stealth hits hard and that AOE off vault will hit everyone with that bonus from revealed and the Staff mastery that in that area of effect (steal to target with hidden thief and then vault. Few block that as few know it coming)

Satff with dd/da/sa is very doable.

Well, yes, I agree with you. The problem is linked with sword and our traitlines, even after the AA buffs, are pretty mandatory. DD/Trickery/DA are still a must, at least 2 out of all 3, I can’t see how we can drop one of them to bring Acro and get the Swindler’s Equilibrium (which gives only a +5% on sword dmg).

I didn’t check in depth the new changes to Acro thou, so I may be mistaken, but at superficial glance it still seem way subpar to the other traitlines.

babazhook.6805:

When comparing sword damage to dagger one can not just ignore the 5 percent buff in the acro line. While this of course has to be traited and some might not take acro the fact remains dagger has no such extra damage.

Staff does of course have the buff in the DD line again if mastery traited.

Well, it can be argued that when going D/P, you can use Revealed Training, which is a +200 power buff while revealed. And with dagger you get in and out of stealth a lot of times, so the reveal uptime in conjuction with RT must be considered similar to Swindler’s or Staff Mastery.

reikken.4961:

Galandil.9641:

I checked the DPS twice for each weapon, stopping when I reached 1M dmg overall. The results are:

Staff – 3100 DPS (consider the constant 13-16 Vuln stacks)
Sword – 2800 DPS
Dagger – 3400 (3-4 constant Poison stacks)

I have no idea how you’re getting so much lower numbers for sword than dagger, when I’m getting almost the exact same damage. 3200 for sword and 3350 for dagger (I turned off poison damage in the log. but it’s only ~100 dps anyway)

This is odd. Did you check if you had NO dmg buffing equip/traits?

I used Acro/SA/DD with no major traits equipped (only the minors, they don’t give any dmg buff), and I used Sigil of Bloodlust + Benevolence + Rune of Altruism (so no dmg buffs from these too).

The difference thou can be caused by the variance of crit. chance and wpn dmg (sword has a wider range, 950-1050, while dagger a smaller range, 970-1030).

However, I did again the sword DPS test, this time I got 2950 (explained by variance), while with Swindler’s Equilibrium I achieved a 3200 DPS (way higher than a +5% increment in DPS).

If you can all do some more tests, it’ll be very useful, but still apparently sword has a lower DPS than dagger nonetheless.

Sir Vincent III.1286:

Galandil.9641:

Ok, I’ll add something to the math calculated DPS of all the main weapons.

I did the test using the jaxnX DPS meter on the immortal golem in the lobby.

I removed ALL damage buffing rune/sigils/traits, with Marauder amulet, and avoided (by asking people not to) any Might/Fury boons during the test. The stats are these:

Power: 2050
Crit. chance: 62.57% (Signet of Agility equipped)
Crit. dmg: 187.3%

I checked the DPS twice for each weapon, stopping when I reached 1M dmg overall. The results are:

Staff – 3100 DPS (consider the constant 13-16 Vuln stacks)
Sword – 2800 DPS
Dagger – 3400 (3-4 constant Poison stacks)

I’m baffled by my results compared to the math done before in the thread, since from that apparently Staff should have a lot higher DPS than Dagger AA. Must be cause of the aftercasts in game opposed to theoretical ones.

However, as far as AA goes, Dagger wins by a huge margin.

Which traits did you take? Did you take Swindler’s Equilibrium (This trait now also grants 5% bonus damage to sword and spear attacks) for the sword?

Edit: posted too soon
Can you also test if this 5% buff actually makes a difference?

I guess I can, I still have 5 minutes before dinner, I’ll try it now and post the result.

Check here: https://forum-en.gw2archive.eu/forum/professions/thief/Patch-notes/5947028

the test I did in the lobby. Sword AA DPS is still way lower than Staff and Dagger AA DPS.

Ok, I’ll add something to the math calculated DPS of all the main weapons.

I did the test using the jaxnX DPS meter on the immortal golem in the lobby.

I removed ALL damage buffing rune/sigils/traits, with Marauder amulet, and avoided (by asking people not to) any Might/Fury boons during the test. The stats are these:

Power: 2050
Crit. chance: 62.57% (Signet of Agility equipped)
Crit. dmg: 187.3%

I checked the DPS twice for each weapon, stopping when I reached 1M dmg overall. The results are:

Staff – 3100 DPS (consider the constant 13-16 Vuln stacks)
Sword – 2800 DPS
Dagger – 3400 (3-4 constant Poison stacks)

I’m baffled by my results compared to the math done before in the thread, since from that apparently Staff should have a lot higher DPS than Dagger AA. Must be cause of the aftercasts in game opposed to theoretical ones.

However, as far as AA goes, Dagger wins by a huge margin.

Thief

Withdraw: The healing values of this skill have been increased by 10%.
Double Strike: This skill’s damage has been increased by 42%.
Lotus Strike: This skill’s damage has been increased by 29%. The attack point has been increased by 0.06 seconds and the aftercast delay has been reduced by 0.35 seconds.
Shadow Shot: This skill will now trigger Assassin’s Reward correctly, even if it misses.
Shadow Strike: This skill will now trigger Assassin’s Reward correctly, even if it misses.
Slice: The attack point of this skill has been increased by 0.16 seconds.
Slash: Reduced the aftercast delay of this skill by 0.1 seconds.
Crippling Strike: Increased the attack point of this skill by 0.08 seconds. Reduced the aftercast delay of this skill by 0.48 seconds.
Basilisk Venom: While under the effects of this venom, your attacks are unblockable.
Hidden Thief: This trait now grants 50% bonus movement speed while in stealth.
Fleet Shadow: This trait has been renamed to Instant Reflexes. This trait no longer grants bonus movement speed while in stealth and will instead grant the thief evasion for 2 seconds when struck while under 50% health, with an internal cooldown of 40 seconds.
Feline Grace: This trait’s vigor duration has been increased from 4 seconds to 5 seconds.
Pain Response: This trait now also removes torment in addition to its current removals.
Guarded Initiation: This trait now removes confusion in addition to its current removals. The health threshold has been reduced from 90% to 75%.
Swindler’s Equilibrium: This trait now also grants 5% bonus damage to sword and spear attacks.
Assassin’s Reward: This trait’s base healing per level has been increased by about 50% and its healing power contribution has been increased by 70%.
Upper Hand: This trait now grants regeneration for 6 seconds when the initiative regain occurs.

Daredevil:

Driven Fortitude: The healing from this trait will now appear in the combat log.
Staff Strike: This skill’s damage has been increased by 36%.
Staff Bash: This skill’s damage has been increased by 16%.
Punishing Strikes: This skill’s damage has been increased by 44%.
Debilitating Arc: Fixed a bug that prevented this skill from functioning with Assassin’s Reward.
Bandit’s Defense: This skill no longer delivers a kick when you first block an attack. After blocking an attack, you will now gain access to a second ability, Reflexive Strike, for 6 seconds.
Reflexive Strike: Deliver a massive kick that knocks enemies down.

Just… wow? The +44% dmg to Punishing Strike is SO huge.

Exedore.6320:

Waiting for a competent balance team.

This will take a while.