Guild Wars 2

Howdy! NA Gold player here- Wait wait come back! I swear I got some decent feedback and hopefully some agreeable points players from all ranks can talk about.

Anyways, I wasn’t going to rush my opinion on the current PvP Season 5 until enough time has past, dozens of games have been played across a few hundred rating that I have been messing around in, etc. Naturally I’ve also been keeping tabs on various PvP related threads. I was the guy who wanted people to not rage within 24 hours of the Season starting…I admit, underestimated the rage in these forums a bit.

But all that aside, I believe I have a very clear idea about how this Season is, compared to the previous Seasons, and what we should maybe focus on down the road. I think we all can understand the positives this season has brought out, I need not go into detail but between the reward split from prestige, actual rating progression with potential to drop across divisions, decay and uniquely interesting titles…

I believe I can say the integrity and authenticity of the current system in place is absurdly better than any previous season. This kind of system should have been in place Day 1 of Season 1, it’s late, but thank you Arenanet for giving us an actual ranking system based on merit.

Unfortunately, because nothing is ever perfect, I believe it is good to go over the problems we have right now. And some of what I want to talk about can get a bit high-level. Let’s start with the simple stuff first, ahem, first problem this season:

Placement Matches were bogus. From just about every single thread posted, to my own experiences, I ultimately have to say the placement matches were bad. There are very likely players that are divisions lower than they should be, unfortunately having the worst time possible and spending dozens of matches trying to climb to an acceptable rating, all because A) They might not have played the previous season and choose to skip thus giving them a ‘rating’ that was not averaged out correctly, or Just the worst kind of synergy in the world, they lost a lot of Placements.

Despite my qualms with the placements, I believe the placements for next season might be better, if they exist for players who played this season. Ideally, Placements would be best used for new players only, returning players shouldn’t have to go through the massive variance in skill ratings to find their approximate spot. But for now, I think the placements were poorly done, and part of it was the new rating system implementation.

Next is this whole chat about Solo Versus Duo Queue. Man I cannot imagine any other competitive game where this dilemma came up…

I am pretty sure League of Legends has the phrase ‘Solo/Duo Queue’ copyrighted because of the sheer number of complaint threads about the subject. And I got some news for you guys who are not aware of how League handled it. …They didn’t.

There’s exactly zero solutions other than wiping out another aspect of grouping up with someone to do Ranked together. Solo/Duo Queue have to remain as is, despite the advantages a duo-queue brought. LoL did try to inflate the duo queue’s average rating a bit, because that would face them against harder opponents, but that wasn’t a perfect solution. And I think for the sanctity of the gamemode, we need Duo Queue to exist with Solo Queue because ultimately SPvP is a team game. You naturally would want to group up with others who you can play well with, form teams and win all the money!

I mean, we are welcome to discuss it some more, but you need only put into the search bar ‘League of Legends Solo versus Duo Queue’ and you will get all the arguments over years about it. Go ahead, get informed! Or experience deja vu, cause some of the arguments already are giving me flashbacks.

(cont)

Lastly, and after some consideration, I believe the current rating gain/loss is still too simple. Our resident SPvP Gameplay Programmer, who likely requires Salt to stay alive, Evan agreed in this thread about the concept of the current rating system being about how you win matches rather than your own skill. The idea that your effort to learn from your mistakes and win more matches than you lose will ultimately make you the better SPvPer.

That’s fine and dandy, but I believe we need to address the fact that since there are unwinnable games for these perfect players, based upon factors that are outside their control, that will still lose rating at the same pace as someone playing poorly.

In other words, I think there needs to be some special cases set in how rating is won or lost, based on the end score and player performance. …I am NOT advocating that the guy who thinks ‘GG’ once the score hits 50-300 and starts playing to get Best Stats across the board, to negate his rating loss all together. That’s a stupid kind of bad.

What I want to see though, is closer matches not punish the best performing losers as much as the worst performing. Right now there’s nothing that stops this, that keeps the extremely skilled player from losing the maximum rating compared to the tempest who had to briefly afk at spawn to answer the door.

We can always improve the rating system, as for what constitutes a ‘close’ game is up in the air, I’ve not decided quite yet. But you know kitten well the games that end 498-500 shouldn’t give the losers full rating loss, give them a rematch!

That’s all I had to say for now! Thanks for reading!

I agree but what way is there to reward individual skill? The only result at the end is whether or not your team wins or loses. If you include criteria for individual performance, you’ll be incentivizing playing to meet that criteria rather than playing for a team win.

I think the amount of times you die in a match determines skill somewhat. I also think the amount of healing/dps/boons you have in a match determines how much you contributed to the team somewhat as well. Those should be bigger factors for points awarded in a match.

FrownyClown.8402:

I think the amount of times you die in a match determines skill somewhat. I also think the amount of healing/dps/boons you have in a match determines how much you contributed to the team somewhat as well. Those should be bigger factors for points awarded in a match.

So much for saccing yourself to save a cap right?

But I don’t think the system is this complex, it’s just avg rating and win/loss + multiplier.

The amount of damage has very little correlation to how effective you are. A good player doing spike damage will kill a lot of people with 200k damage. Whereas an aoe spamming group fighter may rack up 750k damage and not provide anything more than moderate pressure to the people in the team fight. The simplicity of the current system is what makes it good. You can try to put together metrics, but at the end of the day the only stat that matters is which team has the most points at the end. People that are the best at maximizing their teams chance to win will rise to the top.

The only thing I would like to see is more access to people’s MMR and stats, and me not losing as much xD

I’m liking the system somewhat apart from some minor tweaks here and there. I’ve had way more close matches this season than any other season once people started settling into their placements.

Also bonus points is I get to see some of those pro league/streamers play solo/duo and actually see that they have a trash rating because they got carried by premades and can’t hold their own weight :^)

FrownyClown.8402:

I think the amount of times you die in a match determines skill somewhat. I also think the amount of healing/dps/boons you have in a match determines how much you contributed to the team somewhat as well. Those should be bigger factors for points awarded in a match.

Absolutely. There’s a multitude of factors that needs to be taken into account. The fact that Anet implemented a Defense statistic, “Fighting near a contested or owned objective” means 1) They’re in the right direction. And 2) They have the tech to add such coding.

They just need to add more “smart statistics”, then eventually make a new metric that governs in game performance to an extent. Not a sole determinant, just a suplement i.e +5 rating or -5 rating at match’s end.

@OP
I agree, best Season to date. Just wish it was implemented in S1 or S2.

Playing or not playing in past seasons has no impact on rating. I played one game in S3 (to get stacked rewards) and didn’t play at all in S4. I went 6-4 in placement and ended up in gold tier3 and moved up to platinum.

Aside from the vague ‘How to figure out individual skill’ dilemma we have right now, the other perspective I’ve thought of is, to what extent that skill should affect your rating.

Saiyan brought up something like a flat number, but the rating gains and losses change at the extremes. You get higher in ranking and suddenly you are earning a few points per win, and lose 3x at least if you lose. Not that I think that’s too far off, if you are rank 1 you shouldn’t be able to pull away from those below you that easily…

So, I am going to assume Arenanet can eventually figure out a decent way to measure player skill, and determine the best players on each side of the match in a close game. Neat, so how do you reward the ‘losing player’ for trying his/her hardest for just barely losing by say a tick?

What about having the loss take a percentage hit? You performed the best of the losers in a 450-500 game, so maybe instead of losing 10 rating you lose 7? And that consolation applies to just you? I think a percent-based rating adjust might be best here.

If we can get the rating system to be a bit more complicated, such as promoting players who are losing to play their hardest to win in order to not take the biggest hit to their rating, that might promote better games.

Does anyone disagree? And yes, I do realize I dodged the whole ‘Player Skill’ measure here, with the changes to ‘Best Stats’ I assume Arenanet has concept going forward with this.

Sykper.6583:

Aside from the vague ‘How to figure out individual skill’ dilemma we have right now, the other perspective I’ve thought of is, to what extent that skill should affect your rating.

Saiyan brought up something like a flat number, but the rating gains and losses change at the extremes. You get higher in ranking and suddenly you are earning a few points per win, and lose 3x at least if you lose. Not that I think that’s too far off, if you are rank 1 you shouldn’t be able to pull away from those below you that easily…

So, I am going to assume Arenanet can eventually figure out a decent way to measure player skill, and determine the best players on each side of the match in a close game. Neat, so how do you reward the ‘losing player’ for trying his/her hardest for just barely losing by say a tick?

What about having the loss take a percentage hit? You performed the best of the losers in a 450-500 game, so maybe instead of losing 10 rating you lose 7? And that consolation applies to just you? I think a percent-based rating adjust might be best here.

If we can get the rating system to be a bit more complicated, such as promoting players who are losing to play their hardest to win in order to not take the biggest hit to their rating, that might promote better games.

Does anyone disagree? And yes, I do realize I dodged the whole ‘Player Skill’ measure here, with the changes to ‘Best Stats’ I assume Arenanet has concept going forward with this.

I would like to see more stats preferably available through an API. However, I disagree that individual stats should buff people that lose matches. People would just exploit this and run builds that generate the best individual stats. If you want to reward losing teams do it by point differential. Losing 500-495 won’t lose as much rating as 500-50. The better players will on average be able to maintain a tighter point differential than others. That will reward people for playing to help the team vs. playing to inflate their personal stats at the expense of the team.

Faux Play.6104:

Sykper.6583:

Aside from the vague ‘How to figure out individual skill’ dilemma we have right now, the other perspective I’ve thought of is, to what extent that skill should affect your rating.

Saiyan brought up something like a flat number, but the rating gains and losses change at the extremes. You get higher in ranking and suddenly you are earning a few points per win, and lose 3x at least if you lose. Not that I think that’s too far off, if you are rank 1 you shouldn’t be able to pull away from those below you that easily…

So, I am going to assume Arenanet can eventually figure out a decent way to measure player skill, and determine the best players on each side of the match in a close game. Neat, so how do you reward the ‘losing player’ for trying his/her hardest for just barely losing by say a tick?

What about having the loss take a percentage hit? You performed the best of the losers in a 450-500 game, so maybe instead of losing 10 rating you lose 7? And that consolation applies to just you? I think a percent-based rating adjust might be best here.

If we can get the rating system to be a bit more complicated, such as promoting players who are losing to play their hardest to win in order to not take the biggest hit to their rating, that might promote better games.

Does anyone disagree? And yes, I do realize I dodged the whole ‘Player Skill’ measure here, with the changes to ‘Best Stats’ I assume Arenanet has concept going forward with this.

I would like to see more stats preferably available through an API. However, I disagree that individual stats should buff people that lose matches. People would just exploit this and run builds that generate the best individual stats. If you want to reward losing teams do it by point differential. Losing 500-495 won’t lose as much rating as 500-50. The better players will on average be able to maintain a tighter point differential than others. That will reward people for playing to help the team vs. playing to inflate their personal stats at the expense of the team.

You cannot generate the highest stats during a match without using a build that inevitably help your team to the uttermost best that you can possibly achieve in that scenario.

If you want the highest dmg number, you will run a build that create a huge amount of aoe dmg therefore….you will most likely be forced to jump into the most teamfight oriented battles

Same thing if you want to have the highest healing contribution, you will run a bunker build and you’ll be forced to jump in the right situations to reach the best possible outcome

In the end even the most selfish kitten out there will end up helping his/her team more than he/her really wanted to in the first place.

Yes people may try to run troll builds that contribute nothing to the team, most likely the troll is expecting to lose the match but…if no enemy engage the troll, he will end up losing just as much rating as the rest of the team, he will be forced to search for fights and him trolling 1-2 people of the enemy team will already contribute a lot for the team.

If it’s a troll “never die” BS then the enemy team will send either 2 burst or another bunker, its team can focus their effort on another point, they will still lose but, their effort will generate high stats like healing and dmg that will reduce the impact on rating losing.

I fully agree with the OP, the “try hard” should not be punished just as much as the “feeder”, win or lose is part of the game still…give some credit to those who tried regardless of all

This is the best season to date. remove class stacking and duo queue and it’d be golden.

Ithilwen.1529:

This is the best season to date. remove class stacking and duo queue and it’d be golden.

The class stacking is an option that I can see implemented, but only when the SPvP population gets high enough to support. There also is plenty of debate needed for it, my personal experience is that stacking professions to excess never really works out unless it is used for pubstomping. However, there’s a competitive precedent for it from ESL so /shrug.

I can never agree to getting rid of Duo Queue though. If you want I can give you an essay of reasons why Duo Queue with Solo Queue is ultimately healthy for that particular game-mode despite the disadvantage for the solo queue.

I’d like to see that essay against solo queue only. Probably for another thread though.

Maybe, I’ll get around to it at some point. It’s actually a pretty interesting subject honestly, and with all the ‘test’ cases from other competitive games in a similar vein to GW2 SPvP, I believe it’s a worthwhile discourse.

But it’s late for me, holidays, stuff. I’ll mock up something eventually.

Ok, so, i’m gonna go and give my contribution and opinion of the current Season.

I’m a rifle Engineer, and i have been only playing Rifle Engi the whole week, not touched scrapper at all (even tought there was some moments were rolling scrapper was so tempting to do).

These Winrates have been sooo balanced, im really sitting in 50/50 results and, neither doing duo or solo has changed that, the winratio was quite crazy when doing duo, but at the end, it was pretty much the same story, today i had a little win push of 3 wins, and that’s it, i’m pretty much stuck in gold 3, and i was 1 match away from platinum…. unfortunately platinum got away from my hands :/, maybe another day.

I’m not gonna go deep about how my matchups have been against specific classes but,
Guardians are defenitely great this season, it’s undeniable, it’s pretty much like Pirate Warrior (Hearthstone stuff), it’s not the most strongest build, but holy kitten it’s eveywhere because of the efficiency the build has if it’s weakness it’s not broken at the right moment.

I think it would require a miracle to have a match with 0 guardian presence, with that being said, i’ll agree with OP, the system is way more better than the last season, no comparison here, i totally feel my current mmr it’s reflected right with these results.

I still wonder if people who haven’t placed by mid season will get some severe decay preventing them from jumping into the top slots late.

The matches are still bad. I’ve played 90 matches until today and the majority of them are either plain losses or plain wins. My last games look like this:

575:117, 500:302, 502:484, 503:149, 500:202, 301:503, 500:263, 250:500

That’s really still as bad as all the seasons before.

The scoring.

I win a match and get 12 points. I lose a match and I lose 14 points. Why? This indicates I always lose to teams with lower mmr and win against those with lower mmr? How is this possible when my matches are 500:100?

I’m playing most my games 450-500 in t3 gold-t1 plat. Excepting one game that was obviously stacked 66-500 which I only lost -1 or -2 for losing.

I’ll afk my next 100 games and see what happens err I mean I collect data what exactly happens with my mmr!

fishball.7204:

I’m liking the system somewhat apart from some minor tweaks here and there. I’ve had way more close matches this season than any other season once people started settling into their placements.

Also bonus points is I get to see some of those pro league/streamers play solo/duo and actually see that they have a trash rating because they got carried by premades and can’t hold their own weight :^)

LOL which ones?

I just had Edison on my team yesterday and dude was crushing.

I really enjoyed the Xkratos video so just wondering who, maybe worth a watch

IMHO rating loss/gain for each game should consider the difference between the two teams sums of ratings. So that a win against a much higher rated team would increase individual ratings by a larger amount compared to a win against a team with a lower ratings sum.

wow.. 6 straight crush losses Did anet suddenly change the algorythm again?
Well, I guess I’m back to playing a minimum.

- class stacking is big issue

- lopsided matches are still a thing and matchmaking is still forcing loses on higher rated solo players (i checked rating from my team and enemy team and my teammates are always somehow 100 pts lower than my enemy, how it calculates is beyond me)

- rating gains/losses are extremely punishing (especially in off-hours) because the gains/loss is based on your personal rating and enemy rating but it doesn’t take into account how terrible your teammates were

- duoq is fine, stacking duoqs on one team vs only solo on other is wrong

The problem I see is many people are still settling into their MMR placements.. seen several people land into Platinum (high) and lose all the way down to silver over a few days. On the other hand some people are rising nicely from bronze and still climbing.

Class stacking is something I really would advocate removing from the matchup.
Would like to see profession locking – you que on the class you are going to play on and using a profession MMR basis.

Would also like to see the algorithm calculate a profession based team comp % likelihood of win/lose rather than just the straight MMR- if we are not going to eliminate class stacking. E.g. 4 thief line up vs mixed team. The thief lineup would naturally be favoured in most cases to lose. It’s things like 3 thief’s, 2 eles vs war, guard, guard, ele, nec – that you can almost guarantee the second team will win or be favoured to win.