Guild Wars 2

There is no denying it, this new and improved PvP system makes people salty. There’s three main issues that keep popping up: placement, matchmaking and rank changes. Placement is OKish in my opinion, it’s not perfect but it has to be done somehow. Matchmaking leaves a lot to be desired, but I suppose that’s unavoidable with low population. Onto rank…

Having your rank gain/loss be dependent on personal rather than team MMR is ridiculous. You’re 20% of a team and 10% of the match, yet for some reason your personal MMR relative to the enemy team’s is the only one that matters? Are you fighting 1v5? Team MMR baseline should be the first step. Second step would be to incorporate some form of individual performance monitor. We have stats: damage, defense, deaths, healing, etc. How about having each person on the team ranked one through five on these statistic and having their personal rank’s win/loss be dependent on their performance. So say the team’s MMR baseline is 20 and you ranked 1st on each statistic, you get the full 20 on a win and deduced 10 on a loss, whereas someone who ranked in the bottom gets 10 for a win and deduced 20 for a loss.

Being 20% of a team and 10% of a match makes winning or losing a match a bit too random. In order to have skill appropriate rankings, there needs to be some form of individual performance gauge. Without it, I am convinced this game mode will continue to lose population because there is no incentive to play and get better and you aren’t rewarded for your efforts even if you outperform.

Anet can not accurately account for individual performance in a team game. There are just too many factors to consider. However, gains/losses should be calculated based on your team’s average rating vs enemy team average rating. This is such a simple concept it is astounding anet hasn’t done this yet. It needs to happen.

Shaogin.2679:

Anet can not accurately account for individual performance in a team game.

It’s actually entirely possible to do so. Both the win condition for a match and the combat mechanics are mathematically deterministic. I spent some time back in s5 figuring out how to do a system to accurately evaluate contribution towards win condition and how to integrate it with glicko. It’s entirely doable.

However I realized there are two problems with such systems.
1) such a system is biased towards “playmaking” classes like thief. Now such a bias is accurate in that playmakers like thieves do have a higher impact on the match outcome compared to a teamfighting class. However having such a bias in the rating system would not go over well with the community.
2) such a system is fairly complex conceptually, computationally it’s not bad, but most players who aren’t big on theorycraft would have a hard time understanding the system and would likely consider the system witchcraft and write 989423 threads about “why in the name of kittens did I get X rating from Y situation”

Maybe just an exhaustive sheet (archivable) with detailed stats, just to know who did what and how. I know it would help me know where I failed and where I could have been better. I know some dps meter give a lot more useful information than ingame match recap.
For the rest, rank based on team mmr vs ennemy team mmr would be a good start. Still, something should be done to avoid alt accounts faking the results.

they can account for individual performance with top stats, but they said in a previous thread that they wont cuz they don’t want ppl chasing stats rather then the win. it makes sense for some stats, but for others it seems kinda hard to not have to chase the stat and the win at the same time.

Contribution can not be measured in any meaningful way. If for example you have a ranger, that constantly kites 2-3 people who are dumb enough to chase him over the map (because they can’t 1v1 him) he is effectively carrying the game without gaining any “normal” stats like Kills, Defense, Caps etc.

Shaogin.2679:

Anet can not accurately account for individual performance in a team game. There are just too many factors to consider. However, gains/losses should be calculated based on your team’s average rating vs enemy team average rating. This is such a simple concept it is astounding anet hasn’t done this yet. It needs to happen.

They would have to eliminate duo queue then. Otherwise people would just smurf their team.

Faux Play.6104:

Shaogin.2679:

Anet can not accurately account for individual performance in a team game. There are just too many factors to consider. However, gains/losses should be calculated based on your team’s average rating vs enemy team average rating. This is such a simple concept it is astounding anet hasn’t done this yet. It needs to happen.

They would have to eliminate duo queue then. Otherwise people would just smurf their team.

As apposed to losing excessive rating or gaining minimal rating while you had to carry the hell out of your team? Also, removing the final aspect of a team game from PvP, no thanks. If anything it should be strictly 5 man queue. This solo queue crap is garbage.

Agree with OP’s initial point, not so much with the top stats modifier ( there are too many situations where someone is contributing but not directly to a top stat, like with kiting enemies. Also classes that are hybrid would lose out, because while they are performing in multiple top stat categories, they aren’t getting the TOP stat in that category.)

But seriously Anet, listen to the community here. Gains/losses need to be based on TEAM MMR. NOT personal MMR.

Preaching about unfair rank gain/losses in combination with wonky MM for 2 seasons now. I wonder if it will ever be fixed or will pvp completely die before that happens =.=

While I think that using team-MMR as a baseline for rank gain/loss wouldn’t necessarily be a bad idea for soloQ, it just leaves too much room for exploitation when considering duoQ.

As for the whole player-performance thing….just NO. Im 100% with Shaogin on that one. Top-Player stats mean absolutely nothing when it comes to player impact.
You can have an allied druid fight with an enemy on THEIR node for 10 minutes straight….he will get Top player offense, but that doesn’t do anything for the team at all. Top stats in general don’t mean anything.

Stop asking for an individual performance based rating system. It is conceptually near impossible to create, and even it they managed to make an abomination that vaguely resemble such a thing, it would be an absolute nightmare to keep it up to date with the meta.