Changing the Rating System

[Sampson.2403]

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Yes i agree but the que times would get super slow and potentially in off peak times it could not be possible to find a game depending on ur rank

[SlippyCheeze.5483]

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

[Faux Play.6104]

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

[Faux Play.6104]

@Sampson.2403 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Yes i agree but the que times would get super slow and potentially in off peak times it could not be possible to find a game depending on ur rank

That would only affect the top players. Allowing AT matches to count towards a seasons rating could eliminate that issue.

[SlippyCheeze.5483]

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place on that team just as often as you place against that team...

[Faux Play.6104]

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

[SlippyCheeze.5483]

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

[Faux Play.6104]

Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@Brother.1504 said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

[Dreddo.9865]

@Bigbosos.2063 said:No because alt accounts.I don't think it's fair that the majority of players should be punished with a bad MM due to a handful of players with alt accounts. And honestly I don't find it healthy someone competing for rank positions with multiple acc. Yes people are allowed to do it but there is no need to promote this type of play.

[SlippyCheeze.5483]

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only ever placed in "win expected" matches?

[Vieux P.1238]

At this point does anyone really cares if they change it or not? Cuz i don't. I'm like playing 1 or 2 matches then get off to another game.

[Vieux P.1238]

At this point does anyone really cares if they change it or not? Cuz i don't. I'm like playing 1 or 2 matches then get off to another game.

[Brother.1504]

@Vieux P.1238 said:At this point does anyone really cares if they change it or not? Cuz i don't. I'm like playing 1 or 2 matches then get off to another game.

I’m not inclined to be a fatalist. Obviously esports as a goal was a failure but the spvp mode can be fun. If the matchmaking felt better I would play a lot more. Anet should make changes that bring more players into the mode. I know elitists are allergic to casuals and beginners but spvp needs the numbers.

[SlippyCheeze.5483]

@Brother.1504 said:

@Vieux P.1238 said:At this point does anyone really cares if they change it or not? Cuz i don't. I'm like playing 1 or 2 matches then get off to another game.

I’m not inclined to be a fatalist. Obviously esports as a goal was a failure but the spvp mode can be fun. If the matchmaking felt better I would play a lot more. Anet should make changes that bring more players into the mode. I know elitists are allergic to casuals and beginners but spvp needs the numbers.

This! I'd like to see PvP improve, too. GW2 is actually one of the more fun games to PvP in, and I want it to keep getting better. If that means changing the MMR system, or the matchmaker, or whatever, so be it. Keeping it fun is goal number one, haha., :)

[SlippyCheeze.5483]

@Brother.1504 said:

@Vieux P.1238 said:At this point does anyone really cares if they change it or not? Cuz i don't. I'm like playing 1 or 2 matches then get off to another game.

I’m not inclined to be a fatalist. Obviously esports as a goal was a failure but the spvp mode can be fun. If the matchmaking felt better I would play a lot more. Anet should make changes that bring more players into the mode. I know elitists are allergic to casuals and beginners but spvp needs the numbers.

This! I'd like to see PvP improve, too. GW2 is actually one of the more fun games to PvP in, and I want it to keep getting better. If that means changing the MMR system, or the matchmaker, or whatever, so be it. Keeping it fun is goal number one, haha., :)

[Math.5123]

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

I already explained this before, but I'll do it again.Your mmr is compared to the enemy teams average. Just like their personal mmr is compared to your teams average. Meaning, you're 1800 and the average enemy / (and ally) teams average is 1600, you're predicted to win every game. The Matchmaker does not take your allies rating into account when calculating win or loss.

This is why legend players queuing of peak often gets +4 as your highest gain, while losing closer to 30 rating. Regardless wether the team mmrs were the same or not.

[Crab Fear.1624]

who cares about rating? we seen God of PvP fall to nobodies

[Faux Play.6104]

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

If you are rated near the top of the rating pool, the average of the team you are playing will always be less than your rating. If I put 1 1700 rating person with 4 1400 players, and had them fight a team of all 1600 players, the 1700 rating person should lose, but since the team they are up against is 100 pts lower, they will lose more points than they can win.

[SlippyCheeze.5483]

@"AngelLovesFredrik.6741" said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

I already explained this before, but I'll do it again.Your mmr is compared to the enemy teams average. Just like their personal mmr is compared to your teams average. Meaning, you're 1800 and the average enemy / (and ally) teams average is 1600, you're predicted to win every game. The Matchmaker does

not

take your allies rating into account when calculating win or loss.

Yes it does. To quote from the ANet maintained PVP algorithm page in the wiki:

Match Prediction (Deprecated)As of 12/13/2016, this feature is no longer used.The system attempts to predict the outcome of a match with the same metrics used in matchmaking, though can be configured separately.

So the same source of data is used to predict match outcome as is used to produce the teams. That also confirms that while they can be configured separately, at this time they are not.

This is why legend players queuing of peak often gets +4 as your highest gain, while losing closer to 30 rating. Regardless wether the team mmrs were the same or not.

I can't confirm or deny that statement of outcomes, but the outcome prediction is definitely not what you thought, and so, conclusions based on that are also in error.

[Faux Play.6104]

@SlippyCheeze.5483 said:

@"AngelLovesFredrik.6741" said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

I already explained this before, but I'll do it again.Your mmr is compared to the enemy teams average. Just like their personal mmr is compared to your teams average. Meaning, you're 1800 and the average enemy / (and ally) teams average is 1600, you're predicted to win every game. The Matchmaker does

not

take your allies rating into account when calculating win or loss.

Yes it does. To quote from the

ANet maintained PVP algorithm page in the wiki

:

Match Prediction (Deprecated)As of 12/13/2016, this feature is no longer used.The system attempts to predict the outcome of a match with the same metrics used in matchmaking, though can be configured separately.

So the same source of data is used to predict match outcome as is used to produce the teams. That also confirms that while they

can

be configured separately, at this time they are not.

This is why legend players queuing of peak often gets +4 as your highest gain, while losing closer to 30 rating. Regardless wether the team mmrs were the same or not.

I can't confirm or deny that statement of outcomes, but the outcome prediction is definitely

not

what you thought, and so, conclusions based on that are also in error.

What are you trying to prove by quoting a system that clearly states it is no longer used?

[SlippyCheeze.5483]

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

If you are rated near the top of the rating pool, the average of the team you are playing will always be less than your rating. If I put 1 1700 rating person with 4 1400 players, and had them fight a team of all 1600 players, the 1700 rating person should lose, but since the team they are up against is 100 pts lower, they will lose more points than they can win.

A farcical hypothetical situation will give farcical results, yes. Do you have any example that actually obeys the rules of the game, such as "5v5"?

[Faux Play.6104]

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

If you are rated near the top of the rating pool, the average of the team you are playing will always be less than your rating. If I put 1 1700 rating person with 4 1400 players, and had them fight a team of all 1600 players, the 1700 rating person should lose, but since the team they are up against is 100 pts lower, they will lose more points than they can win.

A farcical hypothetical situation will give farcical results, yes. Do you have any example that actually obeys the rules of the game, such as "5v5"?

I posted facts. The system takes YOUR INDIVIDUAL RATING vs. THE OTHER TEAMS AVERAGE RATING to calculate rating change at the end of a match. If you are at the edge of the rating curve, you will always be playing teams that are rated lower than you. Therefore you will always lose more points for a loss than you can get for a win. You asked for examples, they were provided. To top it off you are bring in rating systems that haven't been used for over a year like they are part of the current system. What exactly are you trying to add to this?

[SlippyCheeze.5483]

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@"AngelLovesFredrik.6741" said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

I already explained this before, but I'll do it again.Your mmr is compared to the enemy teams average. Just like their personal mmr is compared to your teams average. Meaning, you're 1800 and the average enemy / (and ally) teams average is 1600, you're predicted to win every game. The Matchmaker does

not

take your allies rating into account when calculating win or loss.

Yes it does. To quote from the

ANet maintained PVP algorithm page in the wiki

:

Match Prediction (Deprecated)As of 12/13/2016, this feature is no longer used.The system attempts to predict the outcome of a match with the same metrics used in matchmaking, though can be configured separately.

So the same source of data is used to predict match outcome as is used to produce the teams. That also confirms that while they

can

be configured separately, at this time they are not.

This is why legend players queuing of peak often gets +4 as your highest gain, while losing closer to 30 rating. Regardless wether the team mmrs were the same or not.

I can't confirm or deny that statement of outcomes, but the outcome prediction is definitely

not

what you thought, and so, conclusions based on that are also in error.

What are you trying to prove by quoting a system that clearly states it is no longer used?

I really don't know how to explain this to you. You described the matchmaking system as working in the way it USED TO WORK, as documented in the wiki, to which I linked. That also describes how the win/loss prediction is now calculated, which is using the same data and scores as the matchmarking system, which explicitly do calculate it based on team vs team expectations.

If that is not sufficiently clear, I have nothing left, and can only wish you luck in your efforts to get GW2 changed to behave in exactly the way GW2 currently behaves.

To your additional comment, in which you say:

I posted facts. The system takes YOUR INDIVIDUAL RATING vs. THE OTHER TEAMS AVERAGE RATING to calculate rating change at the end of a match.

... I can only direct your attention, once more, to the documentation written by the ANet PvP developers which explains why you are, in fact, wrong.

[Faux Play.6104]

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@"AngelLovesFredrik.6741" said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:Bec> @SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@SlippyCheeze.5483 said:

@Faux Play.6104 said:

@Sampson.2403 said:

@"Brother.1504" said:Is anyone satisfied with gw2 spvp matchmaking?

What alternative is there? For the MM system to only create games where all 5 players are super close to each other in MMR? That would create insanely long que times. This would be like Lebron James refusing to play another basketball game unless everyone else on his team is as good as he is. It's just simply unrealistic.

I do think that the points won / lost could be revisited though. No matter how you slice it, having no choice but to que up solo into random team games where you can potentially lose 5 times more points than you can win isn't very healthy.

You can have top 10 people on the same team as gold 1. That is like James waiting 2 minutes to get queued with a random 7th grader vs waiting 10 minutes to queue with someone at D1 college level. The reason top people lose so many pts for a loss is the average team ratings are so low.

Do you mean the other team has a much lower average MMR, or that the team they are on has a much lower average MMR?

Your rating is compared with the other team's average rating. So if you are 1700 rating you may have a couple with that rating on the other team, but most will be lower. Therefore, you lose more points than you can gain.

Ah, the other team is lower, thus you are predicted to win, and when you don't the weaker team receive a large MMR gain, and you receive a large MMR loss.

Of course, on average you should place

on

that team just as often as you place

against

that team...

The problem is both teams are similar. So it doesn't balance out.

Why wouldn't it balance out for you, over the course of multiple matches, when they are similar? At that close you should be getting a similarly close win or lose prediction, essentially hinged the very slight randomness of who gets the marginally better MMR.

I'm sure you have some theory, or mental model of why, over the course of multiple games, that doesn't happen. Can you explain what that is?

If you are gaining 6-8 points for a win and losing 18-20 you have to have a 3-1 win ratio to stay even.

Wrong question. Why wouldn't the number of matches where you place on the "win expected" team and the number where you place on the "loss expected" team balance out? Why wouldn't the distribution of the "extra good" players balance out?

Not "why are the number different", because we all know that: an expected win and an unexpected loss give those numbers cited. What is your theory about why you, if those are all you ever see, are only

ever

placed in "win expected" matches?

I already explained this before, but I'll do it again.Your mmr is compared to the enemy teams average. Just like their personal mmr is compared to your teams average. Meaning, you're 1800 and the average enemy / (and ally) teams average is 1600, you're predicted to win every game. The Matchmaker does

not

take your allies rating into account when calculating win or loss.

Yes it does. To quote from the

ANet maintained PVP algorithm page in the wiki

:

Match Prediction (Deprecated)As of 12/13/2016, this feature is no longer used.The system attempts to predict the outcome of a match with the same metrics used in matchmaking, though can be configured separately.

So the same source of data is used to predict match outcome as is used to produce the teams. That also confirms that while they

can

be configured separately, at this time they are not.

This is why legend players queuing of peak often gets +4 as your highest gain, while losing closer to 30 rating. Regardless wether the team mmrs were the same or not.

I can't confirm or deny that statement of outcomes, but the outcome prediction is definitely

not

what you thought, and so, conclusions based on that are also in error.

What are you trying to prove by quoting a system that clearly states it is no longer used?

I really don't know how to explain this to you. You described the matchmaking system as working in the way it

USED TO WORK

, as documented in the wiki, to which I linked. That also describes how the win/loss prediction is now calculated, which is

using the same data and scores as the matchmarking system

, which explicitly do calculate it based on team vs team expectations.

Odd you keep bring up match making when I'm talking about how rating is calculated after a match. You appear to be confusing the matchmaking algorithm that assigns people to teams to how rating after the match is calculated. While the wiki uses the term "scoring" when it talks about the match maker, they mean how good of a fit a player or group of players is to form a team or an opponent. It has nothing to do with how many points you gain or lose at the end of the match.

If that is not sufficiently clear, I have nothing left, and can only wish you luck in your efforts to get GW2 changed to behave in exactly the way GW2 currently behaves.

To your additional comment, in which you say:

I posted facts. The system takes YOUR INDIVIDUAL RATING vs. THE OTHER TEAMS AVERAGE RATING to calculate rating change at the end of a match.

... I can only direct your attention, once more, to the documentation written by the ANet PvP developers which explains why you are, in fact, wrong.

Please direct me to where it says your RATING change at the end of a match is based on YOUR TEAM AVERAGE. Here is my evidence that it uses your individual rating vs. the average of the opposing team: https://en-forum.guildwars2.com/discussion/comment/331645/#Comment_331645

[Bandlero.6312]

Where the failure begins is that Glicko (and subsequent derivations as well as predecessors such as ELO) was originally developed to rate skill (performance) of a single player in single-player versus single-player competition - NOT miscellany skilled players versus other miscellany skilled players in group play. You wouldn't use Pythagorean to calculate the area of a circle - yes, you could - but it isn't the best or right method to accomplish that. The same holds for Glicko and measuring skill of single players engaged in group play.

One fix would be that personal rating is used only, solely for matchmaking. Once a team has been formed, a team aggregate rating should then be calculated - as an average of all 5 players' ratings, confidences, etc. - which is then subsequently used to determine the rating gain/loss at the end of a match. Your rating gain/loss each match would then reflect team performance and be a more accurate measurement of skill in group play.

Unfortunately, even with a corrected implementation of a rating system such as Glicko, these systems still fail and are completely unable to take into account human-chaos factors inherent to group play. Factors such as win-trading, match manipulation, AFKing, power-leveling/carebearing, bad team compositions (personality, etc); and ignores and is unable to correct for factors such as class-swapping; class imbalance, and class-spec imbalances. A design could be implemented that would take into account, more, personal performances and skill; however, that would further contribute to the factors listed above as people would game the system for their own personal gain.

A personal rating should, ideally, never be used for group play; instead a group rating should be used. Personal rating should only be used for 1v1 matches. 5v5 random matchups should have its own rating. Team-queue matchups should likewise have its own rating.