Guild Wars 2

This has already been said but I’d like to emphasize the point. Glicko was developped for rating chess games. That is an entirely different situation from a spvp match, where the outcome is correlated not only your individual skill but also to the skill of 4 other players. Also, in chess, the outcome is not dependent on a score while team score is the meat and bread of an spvp win.

Kudos for getting rid of the old personal score sheet. Anet needs to keep at it and realize that match score is the best possible estimate of skill, as the high score IS the winning criteria. My earlier suggestion for personal rank still stands.

Keep up the great work, I do enjoy the gameplay.

https://forum-en.gw2archive.eu/forum/game/pvp/What-I-would-change-in-the-ranking-system/first#post6103769

The result should influence the rating lost. The closer you are to 500 the less you should loose.

Wolfric.9380:

The result should influence the rating lost. The closer you are to 500 the less you should loose.

Same for the winning team, if the opponent is closer to 500, les points gained

That way it won’t increase the total rating of all players

Giving more points to blowouts then close matches for the winner seems wrong. I am against giving a boon to players for stomping a team maybe due to AFK/trowing a match after bad start or bad composition. The intention is to give a reason for fighting hard and not to drop a player in rating to much because he just lost 5 games by 1-5 points

An AFK player is something on it’s own, that problem needs a own solution – but when one team is able to stomp the other team (assuming class stacking isn’t an issue anymore) the skill difference seems higher – and triyng to measure skill is the thing we all want, no matter how low/high our skill is

Tiefsee.3647:

Wolfric.9380:

The result should influence the rating lost. The closer you are to 500 the less you should loose.

Same for the winning team, if the opponent is closer to 500, les points gained

That way it won’t increase the total rating of all players

You’re suggesting that players get rewarded more for stomping noobs than winning a good match against decent oponents. That is a horrible suggestion.

Honest question, since i don’t want to ruin someone fiction or brag about difficulty level interms of class/profession used ingame but did they dump pip system at all or did they combined them the new and old for better innovative system? i mean, if i/we scored 350 (speaking of teamscore) back then i gain nothing other than count as a lose and remain division right? is it still the same?

Shaogin.2679:

Tiefsee.3647:

Wolfric.9380:

The result should influence the rating lost. The closer you are to 500 the less you should loose.

Same for the winning team, if the opponent is closer to 500, les points gained

That way it won’t increase the total rating of all players

You’re suggesting that players get rewarded more for stomping noobs than winning a good match against decent oponents. That is a horrible suggestion.

Yes

Because that speeds up the movement into the right direction, the “noobs” go down and for fighting other noobs and the players who are placed against noobs get a higher placement, far away from noobs

If players can stomps, there is an other problem – AFK players (this should make an excuse to this rule), class staking (hopefully it will be solved) or skill difference (and this would be solved, if the good ones move faster up and the unexperienced ones move faster down)

the noobs go down anyway. You don´t need to speed up the stompers. And it will push up players regardless their skill for matches thes even don´t need to fight hard.
Human factor will always decide a large number of matches and handing out bonus points for RL factors on the oposing team is bullkitten.

Wolfric.9380:

the noobs go down anyway. You don´t need to speed up the stompers. And it will push up players regardless their skill for matches thes even don´t need to fight hard.
Human factor will always decide a large number of matches and handing out bonus points for RL factors on the oposing team is bullkitten.

after all those storys i heard about platinum players here in the forum, i am not sure if the movement is fast enough (somethimes they sound worse as bronce)

If you have a match that’s close. This means the enemy teams skill and your teams skill
are similar.

Thus winning team should get only a tiny increase and losing team should get tiny decrease. Players below the average rating should get an increase and players above the average rating should get a decrease.

Build a sum and there you go.

That’d be common sense.

Title is nonsense. Glicko2 may not be optimal, but it’s better than any system yet devised.

Score differential isn’t a great indicator of how close a match is. Winning a team fight can cause a shift of 100 points, but this can swing back in another direction later (just look at some ESL matches). Strong secondary mechanics like on Legacy of the Foefire and Temple of the Silent Storm further muddy the waters.

@OP, it’s almost like sPvP was designed solely for premade 5 v5 teams and not Solo Queue pugs and the Match maker reflects that, where it’s a single team bs team, players were able to choose who their teammates were.

I know that it is counter-intuitive to give more importance to close matches than to matches with large score differentials, but it is the logical thing to do. You have to keep in mind that matchmaking and personal rank are two different things. The goal of matchmaking is to offer players the best opportunity to make a difference in the outcome of a match and that outcome is a win or loss regardless of points. From a matchmaking standpoint, large score differentials are less informative and kind of a failure. The goal of a personal rank — besides bragging rights — is to give a stable foundation for a matchmaking algorithm. If you adjust personal rank substantially for uneven matches, there is no way to know what part of the score difference comes from skill and what part from the circumstances of matchmaking. Two things that should be determined independently are messing each other up and don’t really do their job.

@BlaqueFyre, you’re absolutely right. If team composition was invariable it would be an entirely different situation. In that case, an algorithm like glicko-2 could be argued to apply.

MoVitae.2417:

I know that it is counter-intuitive to give more importance to close matches than to matches with large score differentials, but it is the logical thing to do. You have to keep in mind that matchmaking and personal rank are two different things. The goal of matchmaking is to offer players the best opportunity to make a difference in the outcome of a match and that outcome is a win or loss regardless of points. From a matchmaking standpoint, large score differentials are less informative and kind of a failure. The goal of a personal rank — besides bragging rights — is to give a stable foundation for a matchmaking algorithm. If you adjust personal rank substantially for uneven matches, there is no way to know what part of the score difference comes from skill and what part from the circumstances of matchmaking. Two things that should be determined independently are messing each other up and don’t really do their job.

@BlaqueFyre, you’re absolutely right. If team composition was invariable it would be an entirely different situation. In that case, an algorithm like glicko-2 could be argued to apply.

Yes and that’s why it’s used since that is what the gamemode was designed around, but due to Solo Queuer complaints and lack of population Solo Queue was shoe horned into it. And as it is right now iirc most/all games that have a PvP gamemode use a variation of Elo/Glicko 2 that have PuG match making.

If they all use it, it must be good, right? Saying it out loud feels off somehow. :P

Well there is no better alternative so there is that…

There is one in my opinion. This alternative or a variation thereof would have the advantage of being well-founded in statistical theory. Sadly the use of elo/glicko variants for spvp is not.

https://forum-en.gw2archive.eu/forum/game/pvp/What-I-would-change-in-the-ranking-system/first#post6103769

Less than 100 points. -20% of lost rating. Obviously the game wasn’t balanced but you still lost.
200-400 as it is now.
400-500 -35% from lost rating. Yeah people would throw for friends and let them hit 400 but wouldn’t be too big a deal. Some problems with this? Yes. But it would forgive matches that shouldn’t happen from decking your rating whilst ensuring that equally skilled teams stay at the same level.

You know what’s the best way to fix Matchmaking in Gw2 PvP? You remove Solo queue and have it as it was designed around 5v5 Team Queue, since that what the entire system was designed around. It would take very minimal effort on Anets part to implement and would cull all the QQ about horrible Matchmaking, now tie that to completely visible Skill Rating after matches and boom all problems solved. The big flaw with player created “variations” is they have no basis to run off of besides what they feel would work or benefit them the most. While most Games use elo/Glicko variants for Matchmaking and guess what so do most professional sports leagues. Want to know what all of those have in common? They are based upon static/semi static groups for their actual raked/competitive leagues. But yeah it has no basis that it works..

everything is better than the current system

BlaqueFyre.5678:

You know what’s the best way to fix Matchmaking in Gw2 PvP? You remove Solo queue and have it as it was designed around 5v5 Team Queue, since that what the entire system was designed around. It would take very minimal effort on Anets part to implement and would cull all the QQ about horrible Matchmaking, now tie that to completely visible Skill Rating after matches and boom all problems solved. The big flaw with player created “variations” is they have no basis to run off of besides what they feel would work or benefit them the most. While most Games use elo/Glicko variants for Matchmaking and guess what so do most professional sports leagues. Want to know what all of those have in common? They are based upon static/semi static groups for their actual raked/competitive leagues. But yeah it has no basis that it works..

And you are left with only a few players in the system, who only team que at their uptime, never meeting most other teams. Most players wont bother to form a duo que much less a 5 man team. People will still qq about their team mates. And 5v5s will continually get the same matchups vs the same teams since there will be so few teams playing at their skill level. Gold would be empty and only bronze or plat would be worth playing.