Guild Wars 2

gw2efficiency.com pulls data from around 24k accounts and shows that as we near the end of Season 5 the average account sits right around the stated average Skill Rating of 1200.

The top 10% of accounts cited are Gold T3 or higher, the top 1% are Platinum T3 or higher. I feel like the rating system in aggregate is working as intended.

Maybe the only people who post their API-keys to gw2efficiency are PvE Carebears. Maybe polling EU/NA accounts at the same time, in the same group affects these numbers. Placement probably still needs some tweaks.

If these numbers are representative of the player-base as a whole, doesn’t that mean the leagues represent -in general terms- the skill level of each player? Does it make sense to believe that the average player sits somewhere in Silver League?

A popular EU Thief Streamer said online the other day that Skill Rating is usually accurate after about 100 games. He also said Ranking shouldn’t apply until after about 200 games.

I agree with him.

Sources cited: gw2efficiency.com, wiki.guildwars2.com

Kreweless.2196:

A popular EU Thief Streamer said online the other day that Skill Rating is usually accurate after about 100 games. He also said Ranking shouldn’t apply until after about 200 games.

That much was said by a Dev himself a while back. It’s why the Placement matches is an utter joke and it’s why things always start to “settle” towards the end of a season.

I’m curious if they will either remove the placement matches entirely, or increase the number of placement games to 50 at minimum.

I can not agree with this totally because of decay.

Do you know how this chart scaled decayed players because it looks like it decayed a player and he played 1 match and said no more for me.

I had a friend who had a rating around mine of 1700 and he took around 2 weeks off. He has 120+ games but he came back played a game and lost and his rating his now 12xx.

For me, there should be no decay system. Just have a min games played. 150 is a great number. Fix the first 10 games, have it so the highest rating you can get is 1600. Now all the PvP giants have to battle it out to get to plat + and Leg.

Then Anet simply needs PvP rewards for the PvP players.

Literally everything about PvP has turned into lets help Pve players with AP, or new skins, or them getting there ascended armor.

Like i love new blood coming into the PvP scene but you want them there to learn to pvp, not to farm pve rewards.

kdaddy.5431:

Do you know how this chart scaled decayed players because it looks like it decayed a player and he played 1 match and said no more for me.

It’s unclear whether gw2efficiency registers decay when they provide the stats; maybe someone with decay can check their account on that site and see?

I’m still not sure how much decay affects the statistics. I would figure the majority of players with higher Skill Rating are active enough to not suffer from it.

kdaddy.5431:

Then Anet simply needs PvP rewards for the PvP players.

I agree with you 100%. The BL Chests started to include Guaranteed Wardrobe Unlock items that thematically agree with the original sPvP rewards which really only unlocked skins and finishers (you know, sPvP stuff?).

I think a return to that reward philosophy would be a good compromise, going forward. All players would be happy to get more skins, but I think most dedicated PvPers aren’t really interested in crafting materials, etc.

Saiyan.1704:

It’s why the Placement matches is an utter joke and it’s why things always start to “settle” towards the end of a season.

You’re misinterpreting the purpose of placement matches and are vastly exaggerating their inaccuracy.

Placement matches basically exist to mask the large rating shifts when your rating is relatively unknown, especially if you haven’t participated at all or in a long time. If players saw rating changes of +200 or -200, they’d freak out. By the time you finish those placement matches, your rating is relatively close to where it should be and shifts will mostly be small. There are outliers, but the majority will be in the ballpark.

It’s important to remember that rating should not be thought of as an absolute number. In Glicko2 there’s actually a second number, rating deviation, which expresses the bounds where you skill actually is. Shifting up or down 50 points from day to day is perfectly reasonable and doesn’t mean the system is inaccurate.

Another thing to note is that when you compact ratings (like the soft reset), it does take a while for players to spread out again, especially at the top and bottom. If you’ve been near the top since the start of the season, you’ve probably noticed your rating go up by 50-100 points since the start. This isn’t an affect of individual games played, but requires all players to play more games.

kdaddy.5431:

For me, there should be no decay system. Just have a min games played. 150 is a great number. Fix the first 10 games, have it so the highest rating you can get is 1600. Now all the PvP giants have to battle it out to get to plat + and Leg.

how would you protect against rating protection though? it would be dumb if someone could play the minimum, quit, and not have their rating decay at all. maybe add a max decay of 200 over 2 weeks? idk. something seems necessary.

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

But if you think that skill rating truly shows your and only your skill level, then explain, how can some players with skill rating belonging to 0,5% top players drop to skill rating of 5% in over 10 games with total games played being 150+ .

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

In each individual game, sure, it does — and your win or loss is determined in the same part by the five random players you face off on the other team. Over many games, it reflects your skill.

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

SlippyCheeze.5483:

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

In each individual game, sure, it does — and your win or loss is determined in the same part by the five random players you face off on the other team. Over many games, it reflects your skill.

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

You kinda forgot that the random 5 players that you are facing can have up to 600 more individual skill rating than you. Which can lead to – If u cant beat someone with 500+ rating, you wont progress.

It can be called skill rating if the matchmaking is perfectly balanced and that will never happen in gw2 due to population problem.

To your, so called classic mistake:
Does it make sense that my skill rating that reflect my skill ( thats what you think i guess) will go drastically down in over a day. We are talking about 250 rating here. If the skill rating perfectly represents your skill that situation should never happen unless my playstyle went really downhill or the matchmaking of the game is bad.

Hmm let me guess which one will it be….

SlippyCheeze.5483:

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

In each individual game, sure, it does — and your win or loss is determined in the same part by the five random players you face off on the other team. Over many games, it reflects your skill.

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

All he is trying to say is that the part should not be equated to the whole, and vice versa, which is a common logical fallacy:

Division Fallacy basically infers that something is true of one or more of the parts from the fact that it is true of the whole. 

P is part of X
X has property L
Therefore, P has property L.

There is another fallacy that many argue as true: Composition Fallacy, which infers that something is true of the whole from the fact that it is true of some part of the whole. 

P is part of X
P has property W
Therefore, X has property W.

Disclaimer: I am not a philosopher, but I believe these are correct – if I get them backwards or wrong, please let me know so I don’t get it wrong in the future.

Its´just the number of games you need to get to your place depends on luck ….You can be there after 10 or 500 games especialy when the syste m"thinks" you are at the right spot the effort gets much higher to climb. It also depends largely on your carry ability and this on class and build and that you handle this type well. So …..

skylink.4972:

SlippyCheeze.5483:

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

In each individual game, sure, it does — and your win or loss is determined in the same part by the five random players you face off on the other team. Over many games, it reflects your skill.

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

You kinda forgot that the random 5 players that you are facing can have up to 600 more individual skill rating than you. Which can lead to – If u cant beat someone with 500+ rating, you wont progress.

It can be called skill rating if the matchmaking is perfectly balanced and that will never happen in gw2 due to population problem.

To your, so called classic mistake:
Does it make sense that my skill rating that reflect my skill ( thats what you think i guess) will go drastically down in over a day. We are talking about 250 rating here. If the skill rating perfectly represents your skill that situation should never happen unless my playstyle went really downhill or the matchmaking of the game is bad.

Hmm let me guess which one will it be….

I didn’t forget that the random five players on the other team can have 600 more MMR than you … just like I didn’t forget that the players on your team can have that 600 more MMR.

The problem with “ELO Hell” or “MMR is inaccurate because team composition” as ideas is that every single thing you say is keeping you down because it can randomly happen to one team or the other … applies equally to both teams.

You are also mistaking “MMR is accurate” for “MMR is perfectly accurate, all the time”, which your argument would break, yes. However, why did your skill go down 250 points in a day?

• you had an off day, and played really badly today
  * you have high volatility because you have not played in a while
  * you have high volatility because, despite randomly matching with players better and worse than you, you still lose when expected to win, and win when expected to lose
  * your MMR is higher than your skill, and it’s still moving toward the correct value (which takes ~ 100 games)
  * you switched your build or class to something else, and suck at it compared to the one you played before (right now, at least.)
  * you duo-queued with someone with a very different MMR, and that made it harder for the MMR algorithm to home in on your skill
  * you played at a super-quiet time, so a broader distribution of MMRs in the games, leading to increased volatility

Ultimately, though, your MMR going down means just one thing: you lost matches you were expected to win, and didn’t win matches you were expected to lose. The people you were randomly matched with played better than you did.

Going down a long way means one of two things: either your personal volatility is high, or you consistently played badly despite it being low. Either way, your MMR was lowered to reflect the quality of your play.

Sorry, I know that’s not nice to hear, but … ultimately, your MMR will wiggle up and down a little around the target point, because it is always an approximation of skill. If it changes dramatically, though, that’s about you, not the game.

Soothsayer.9206:

SlippyCheeze.5483:

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

In each individual game, sure, it does — and your win or loss is determined in the same part by the five random players you face off on the other team. Over many games, it reflects your skill.

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

All he is trying to say is that the part should not be equated to the whole, and vice versa, which is a common logical fallacy:

Division Fallacy basically infers that something is true of one or more of the parts from the fact that it is true of the whole. 

P is part of X
X has property L
Therefore, P has property L.

There is another fallacy that many argue as true: Composition Fallacy, which infers that something is true of the whole from the fact that it is true of some part of the whole. 

P is part of X
P has property W
Therefore, X has property W.

Disclaimer: I am not a philosopher, but I believe these are correct – if I get them backwards or wrong, please let me know so I don’t get it wrong in the future.

Second time im answering to you, what you put sounds good and its logical, the problem its the root (where you start) xd, its not the matchmaking make P have the properties of W, but the matchmaker have enough Xs (X1, X2, X3, etc) and its parts (each team member) Ws to get near the actual value of P- interval- (basically think of it as when you do a linear regression and get an interval of the true value of how much a independent variable influence in a dependent one, example may not be perfect, but its the same type of principle what makes Glicko systems work on games)..

Edit: Im sleepy and its not my main language, so bear my example, anyway its simple xd

SlippyCheeze.5483:

You are also mistaking “MMR is accurate” for “MMR is perfectly accurate, all the time”, which your argument would break, yes. However, why did your skill go down 250 points in a day?

• you had an off day, and played really badly today
  * you have high volatility because you have not played in a while
  * you have high volatility because, despite randomly matching with players better and worse than you, you still lose when expected to win, and win when expected to lose
  * your MMR is higher than your skill, and it’s still moving toward the correct value (which takes ~ 100 games)
  * you switched your build or class to something else, and suck at it compared to the one you played before (right now, at least.)
  * you duo-queued with someone with a very different MMR, and that made it harder for the MMR algorithm to home in on your skill
  * you played at a super-quiet time, so a broader distribution of MMRs in the games, leading to increased volatility

1. When i have an off day i play maybe bad for the first 1 – 2 games not 12
2.+3. Decay was only on the first match. Supposed-to-win-match wouldnt result in 500-50 score. ( that was the score of 10 games out of 12)
4. I climbed to that rating with 120 + /- games played so if my skill rating was 250 rating below that, why was i allowed to progress there.
5. I played only one class entire 150 games + same build.
6. SoloQ only
7. Que was pretty fast.

I dont understand your reasoning with that you have the exact same chance to get that 600+ rating above players with you. If you were little aware you would realize that there are xxxx threads about how high rated players are stacked on one team vs low plat/high gold. You have to realize that the more RNG factors there are the less control you have over your personal skill rating. Its like playing dice-rolling game during queue and in match you have to work with what you got.

Its hard to call something skill rating when skill is not the only thing involved.

Exedore.6320:

Saiyan.1704:

It’s why the Placement matches is an utter joke and it’s why things always start to “settle” towards the end of a season.

You’re misinterpreting the purpose of placement matches and are vastly exaggerating their inaccuracy.

Placement matches basically exist to mask the large rating shifts when your rating is relatively unknown, especially if you haven’t participated at all or in a long time. If players saw rating changes of +200 or -200, they’d freak out. By the time you finish those placement matches, your rating is relatively close to where it should be and shifts will mostly be small. There are outliers, but the majority will be in the ballpark.

It’s important to remember that rating should not be thought of as an absolute number. In Glicko2 there’s actually a second number, rating deviation, which expresses the bounds where you skill actually is. Shifting up or down 50 points from day to day is perfectly reasonable and doesn’t mean the system is inaccurate.

Another thing to note is that when you compact ratings (like the soft reset), it does take a while for players to spread out again, especially at the top and bottom. If you’ve been near the top since the start of the season, you’ve probably noticed your rating go up by 50-100 points since the start. This isn’t an affect of individual games played, but requires all players to play more games.

We don’t need more placement matches necessarily what we need is a hard MMR reset. You can’t go to a basically solo queue league system and rely so heavily on MMR a lot of people built up playing mainly in groups in the past for initial placement. Everyone should get a fresh start when you change things so radically. Each season should be unique if they are so good at catching skill level in a minimum amount of games as you claim they are why not just rely on the placement games totally? This game has been out for like five years time for a MMR reset! They also need to find a way to encourage people to play in a range of games of around 200 total at least during the season to eliminate any of the luck factor and eliminate doing things like starting highly ranked and then playing the bare minimum to try to protect your rating.

I like the new system, but I’m not too happy with the divisions. A gold t1 player gets the same emblem as a gold tier 3 despite having a higher rating.

I don’t like the idea that if you’re average, you can get silver. Being slightly better, you’re gold. You have to exceed top 10% just to get platinum, I think the gaps are just too far.

I won’t pretend to have done enough thinking about hard resets and the inner workings of a solid Glicko system to talk about those things, but I would like to see the up-and-down of that number reflect actual performance, rather than just the win or loss.

Let’s say I sweep the top stats for my team in a really bad silver tier 1 (bottom tier) loss. If the system is gathering data on damage, heals, revives, etc., I feel it wouldn’t be unreasonably difficult to devise a system where

1) the game compares my top stats to the mean/median top stats for silver tier 1 matches,

2) determines how close (within x standard deviations, or percentile, or whatever) my stats are to those mean/median silver tier 1 stats, and

3) adjusts my skill rating loss to reflect that I tried pretty hard (and succeeded to some extent) to carry more than my weight.

I don’t think it should reduce the skill rating loss to 0. Also, I understand how things can be relatively accurate in aggregate, even if the components making up that aggregate may not reflect that.

That being said, why not have a slightly higher-resolution way to assign skill rating changes? I think it could only help with matchmaking precision. As a compromise, I don’t think the skill rating adjustment should ever come close to zeroing out the loss – I’m talking going from a full 12-point skill rating loss down to maybe an 7 or 8 at most. I’m not sure if the same calculations would justify adding more skill rating to wins where you sweep 3-5 top stat ratings. I guess it could, since that could accelerate differentiation (if a player is consistently carrying their team in 5 stat categories through wins, time to move them up even faster, no?).

Lucius.2140:

Soothsayer.9206:

SlippyCheeze.5483:

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

In each individual game, sure, it does — and your win or loss is determined in the same part by the five random players you face off on the other team. Over many games, it reflects your skill.

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

All he is trying to say is that the part should not be equated to the whole, and vice versa, which is a common logical fallacy:

Division Fallacy basically infers that something is true of one or more of the parts from the fact that it is true of the whole. 

P is part of X
X has property L
Therefore, P has property L.

There is another fallacy that many argue as true: Composition Fallacy, which infers that something is true of the whole from the fact that it is true of some part of the whole. 

P is part of X
P has property W
Therefore, X has property W.

Disclaimer: I am not a philosopher, but I believe these are correct – if I get them backwards or wrong, please let me know so I don’t get it wrong in the future.

Second time im answering to you, what you put sounds good and its logical, the problem its the root (where you start) xd, its not the matchmaking make P have the properties of W, but the matchmaker have enough Xs (X1, X2, X3, etc) and its parts (each team member) Ws to get near the actual value of P- interval- (basically think of it as when you do a linear regression and get an interval of the true value of how much a independent variable influence in a dependent one, example may not be perfect, but its the same type of principle what makes Glicko systems work on games)..

Edit: Im sleepy and its not my main language, so bear my example, anyway its simple xd

it’s good and logical, because it is true (argument is sound and valid as far as i can tell). plus, you have to remember the matchmaking works within a “range” and on certain “assumptions” (which Anet outlines – as a result, matchmaking cannot/does not account for every sort of variable and can be wrong for some whether that’s advantageous or disadvantageous).
so then, you cannot assume to equate the part with the whole or vice versa. it is because of these gaps with range/assumption that some will fall outside of optimal/exact results (again working with ranges/averages and assumptions/data available not exactness/certainty). i only hope my logic is not off, but if it is, then it’s an opportunity to learn something.
lastly, i would have answered your last post but i was not going to answer any more on the other thread (not that i didn’t want to respond to you personally, so i hope you don’t take it that way), so i apologize for not responding sooner. btw, no problem on your language though, completely understand and won’t make a big deal out of it (no one should, but you know how people can be).

voltaicbore.8012:

I won’t pretend to have done enough thinking about hard resets and the inner workings of a solid Glicko system to talk about those things, but I would like to see the up-and-down of that number reflect actual performance, rather than just the win or loss.

It’s worth noting that it is not just “thinking” that backs Elo and Glicko, it’s actual hard math. Probably a mostly pedantic point, but … it’s worth noting that these have had significant thought and math and refinement applied in the real world.

Anyway, the key issue with your idea is that it builds a metric into your score changes. Once you have a metric, people play to the metric — and you get what you measure, nothing else.

If taking top stats is worth something, you get builds for it: the “top DPS” glass cannon build, the “nothing but stability, and resurrection” build, all of which contribute less to winning the match, but more to the personal score of the player involved.

The only thing that, ultimately, reflects skill without forcing particular playstyles on people is … win/lose. Not how you get there. Which is why these systems pretty much all end up falling back to it.

Also useful to remember: they also result in good matchups, using only that metric, so this is a tested and valid way to measure skill.

Also also useful to remember: no matter which game you point to as having a “better” MMR system, it’s almost certain they use the same basic math that GW2 does.

Kreweless.2196:

kdaddy.5431:

Do you know how this chart scaled decayed players because it looks like it decayed a player and he played 1 match and said no more for me.

It’s unclear whether gw2efficiency registers decay when they provide the stats; maybe someone with decay can check their account on that site and see?

I checked and no, gw2efficiency do not take decay into account. It displays and use your true skill rating, not your effective skill rating.

steelheart.7386:

Exedore.6320:

Saiyan.1704:

It’s why the Placement matches is an utter joke and it’s why things always start to “settle” towards the end of a season.

You’re misinterpreting the purpose of placement matches and are vastly exaggerating their inaccuracy.

Placement matches basically exist to mask the large rating shifts when your rating is relatively unknown, especially if you haven’t participated at all or in a long time. If players saw rating changes of +200 or -200, they’d freak out. By the time you finish those placement matches, your rating is relatively close to where it should be and shifts will mostly be small. There are outliers, but the majority will be in the ballpark.

It’s important to remember that rating should not be thought of as an absolute number. In Glicko2 there’s actually a second number, rating deviation, which expresses the bounds where you skill actually is. Shifting up or down 50 points from day to day is perfectly reasonable and doesn’t mean the system is inaccurate.

Another thing to note is that when you compact ratings (like the soft reset), it does take a while for players to spread out again, especially at the top and bottom. If you’ve been near the top since the start of the season, you’ve probably noticed your rating go up by 50-100 points since the start. This isn’t an affect of individual games played, but requires all players to play more games.

We don’t need more placement matches necessarily what we need is a hard MMR reset. You can’t go to a basically solo queue league system and rely so heavily on MMR a lot of people built up playing mainly in groups in the past for initial placement. Everyone should get a fresh start when you change things so radically. Each season should be unique if they are so good at catching skill level in a minimum amount of games as you claim they are why not just rely on the placement games totally? This game has been out for like five years time for a MMR reset! They also need to find a way to encourage people to play in a range of games of around 200 total at least during the season to eliminate any of the luck factor and eliminate doing things like starting highly ranked and then playing the bare minimum to try to protect your rating.

BIG NO TO MMR RESET!!!!!!!!!!!!!!!!

NOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO!

NO!
NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!NO!

After a soft reset, it took me 150 games or more before I started playing with allies and enemies of similar skill. Now my games are usually 3/5 allies know what to do, not me and maybe 1 other if I was lucky.

Don’t get me wrong, I would love to reset my MMR after this season. The math says I was 1600+ before placement games which means my MMR has dropped nearly 400 and it’s been a hell of a time trying to get back to the “good” matches because people in this tier, the matches in this tier are really random. (I’m OCX, may not be like this during Prime)

Evan Lesh.3295:

We’ve done a full reset before and it was awful. Keeping everyone’s relative position means matchmaking isn’t completely random at the start of the season.

Soothsayer.9206:

it’s good and logical, because it is true (argument is sound and valid as far as i can tell). plus, you have to remember the matchmaking works within a “range” and on certain “assumptions” (which Anet outlines – as a result, matchmaking cannot/does not account for every sort of variable and can be wrong for some whether that’s advantageous or disadvantageous).
so then, you cannot assume to equate the part with the whole or vice versa. it is because of these gaps with range/assumption that some will fall outside of optimal/exact results (again working with ranges/averages and assumptions/data available not exactness/certainty). i only hope my logic is not off, but if it is, then it’s an opportunity to learn something.
lastly, i would have answered your last post but i was not going to answer any more on the other thread (not that i didn’t want to respond to you personally, so i hope you don’t take it that way), so i apologize for not responding sooner. btw, no problem on your language though, completely understand and won’t make a big deal out of it (no one should, but you know how people can be).

Separated it in two (3) parts:

• 1: I agree its a range (thats why the interval part), the system has a deviance for it that depends of three factors: time from last data, number of iterations and how much stable its the performance (volatility- its accounted better since glicko 2 and added to the deviance, the first two were the original glicko 1 improvement over Elo).

Deviance are used in the entire process of determining the skill rating, both the user one and the others deviances.

The deviance itself represent the errors and as such the unaccounted variables you refered. Its normal to not have all the variables but it have the more significatives: participants suposed skill rate and deviances.

Glicko 2 itself its an alghoritm, so it will change the skill rate and deviances making them more accurate each time (100 matches tend to be a very good place) except deviances have a minimum, because if not the system will not allow you to raise ever (without a time out) after a certain number of stabilized iteractions xd.

• 2: To correlate mathematically your propositions to the Glicko 2 system i made this small model (simplificated but i think will be enough):

Each player capability to influence the match will be refered as “a”, a1 for player 1, a2 for player 2, etc. And will be its average capability.

Now: P = player 1 and its real atribbute its a1.
X its the team itself composed for players 1 to 5.

X total performance its a1+a2+a3+a4+a5, and the average performance of members of X its that divided by 5.

This average performance its the atribute Wr and can be called awr, where r denotes the skill rating, so Wx its the mean performance for a player in said skill rating.
.

P is part of X
X -> awr
P -> awr, However as you put, saying a1 = awr isnt necesarily true and such its not a valid conclusion.

In fact a situation where it is would be very improbable if its done from nothing.
Of course this a1, etc have deviances, and they include the propieties of the deviances i told you before in point 1.

Now, Glicko actually has 1 more set per match of your proposition for the enemy team. Lets say this match (match 1) have the player team as an X1 and the enemy team as X2.

The awr of X2 will be denoted as awr2 and the one in X1, awr1, the chances of winning for X1 are: (awr1/(awr1+awr2))x100, tnow matchmaker tries to make teams of equal skills, so what it tries to do its: awr1=awr2 and as such if a player has a1=awr1, his win rate its 50%.

Well what the Glicko system do its this: It feeds on an incredible big ammount of this equations and the actual winners and loser and adjust the values of the rating of each player to try to make their “a” to be equal to the awr (or basically try to put you in a skill rate where your win rate since you entered that skill rate will be 50% – with ups and downs in the skill ratte- if it doesnt, it just keep moving). This process its repetead a lot of times.

So what it does in itself its to create this new propositions in the group you alredy stated:

P -> a1
a1= awr

So it ends as:

P-> a1
a1= awr (Wr)
P->X
X-> awr (Wr)
P-> Wr (awr)

That its sound as valid, and works as long as you feed the system well and a good group of players did, the system works better over iterations.
All of this with the smaller deviation posible for the system (its minimum).

Edit: For this to happens it also try an eventually gets: the mean of a determined awr for all matches its: awr (mean), that its equal between teams (in means) and equal to a1.

Basically work with the win rate for an skill rate and put peoplet with a similar performance together and against until they are as much as similiar as posible.

• 3 Lol np xd ^^.

Spoichiche.1290:

Kreweless.2196:

kdaddy.5431:

Do you know how this chart scaled decayed players because it looks like it decayed a player and he played 1 match and said no more for me.

It’s unclear whether gw2efficiency registers decay when they provide the stats; maybe someone with decay can check their account on that site and see?

I checked and no, gw2efficiency do not take decay into account. It displays and use your true skill rating, not your effective skill rating.

thank you,

but that still doesnt answer the come back factor.

Example- I have a 1300 rating, i go on vacation and am gone for 2 weeks. I receive -700 from decay.

I now have 600 rating. I play 2 games and lose both. I receive +200 from the decay and lose 60. I decide i dont want to play anymore.

Is my new effective rating 740? Since i came back to play to get my new rating, ive seen a few people on my friends list who had X rating, took a break and play some games and there new rating is much lower then the old rating before the decay.

kdaddy.5431:

Example- I have a 1300 rating, i go on vacation and am gone for 2 weeks. I receive -700 from decay.

If you’re willing to take a two week vacation during the season, then you’re not really serious about PvP. So, the Bronze SR accurately reflects your level.
#AirTightLogic

Also, sitting at 1300 – you’re pretty much an average player.

I think the biggest complaint with the current (and in my opinion way better) badging system is that only a few players get the shiniest badges.

Chase.8415:

I don’t like the idea that if you’re average, you can get silver. Being slightly better, you’re gold. You have to exceed top 10% just to get platinum, I think the gaps are just too far.

“It’s not fair that only the top players get the top badges.”

C’mon guys, do you really all want participation trophies? Doesn’t that completely remove the prestige of earning the top badges? Maybe there are ways to game the system, but let’s tackle those instead of taking steps back to S1-4 where everyone could grind to Legendary.

It’s not really Legendary if anyone can get it.

“Skill rating” is accurate??? lol

It’s 5v5 and you can’t queue with a team but you are rated only on team performance. That’s 80% of the effort and decision making on your team that is entirely out of your control. That’s 80% of your rating that is not derived from your actions.

Accurate he says… ROFL

SlippyCheeze.5483:

You seem to be making one of the classic mistakes, which is to look at a single event, then complaining that something that measures skill over many games is inaccurate.

could be wrong but I think what some people are saying is that they get bad team mates or match throwers often.

voltaicbore.8012:

snipples

this is great man. if it could be done seems like it would help. I would think that the same rules apply to wins as well.

hackks.3687:

It’s 5v5 and you can’t queue with a team but you are rated only on team performance. That’s 80% of the effort and decision making on your team that is entirely out of your control. That’s 80% of your rating that is not derived from your actions.

This is the argument that just about everyone (myself included at times in the past) deploys when they’re dissatisfied with their Skill Rating.

It takes an exceptional player to consistently support their team through wins. Only those players (as stated, the top 10% of thousands of players) rise out of the majority of the player-base into Platinum/Legendary.

These statistics and ratings are derived from aggregations of experiences. You can’t get a few bad matches and cry, “Foul! ArenaNet isn’t supporting my belief that I’m better than everyone!”

I think the average player is way more willing to poopoo on their teammates for mistakes and cry “MMR is broken, I didn’t get good teammates again!” than they are to examine their gameplay and get better.

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating. I think the numbers prove that. I think a lot of the stories of climbs through the rank support that proof.

Kreweless.2196:

hackks.3687:

It’s 5v5 and you can’t queue with a team but you are rated only on team performance. That’s 80% of the effort and decision making on your team that is entirely out of your control. That’s 80% of your rating that is not derived from your actions.

This is the argument that just about everyone (myself included at times in the past) deploys when they’re dissatisfied with their Skill Rating.

It takes an exceptional player to consistently support their team through wins. Only those players (as stated, the top 10% of thousands of players) rise out of the majority of the player-base into Platinum/Legendary.

These statistics and ratings are derived from aggregations of experiences. You can’t get a few bad matches and cry, “Foul! ArenaNet isn’t supporting my belief that I’m better than everyone!”

I think the average player is way more willing to poopoo on their teammates for mistakes and cry “MMR is broken, I didn’t get good teammates again!” than they are to examine their gameplay and get better.

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating. I think the numbers prove that. I think a lot of the stories of climbs through the rank support that proof.

But 1 player can and will throw the game if he see’s fit. It happens a lot in “elo hell”. I’m not too proud to say this, but I threw a game with a 250 point lead just annoy some kid who was abusing everyone and I actually had more fun watching him rage than actually winning the game. I even jumped on his corpse as he was down and getting stomped. (I could have easily saved him too)

Was it fair to the other 3 players on my team? Hell no, but according to you, they should have carried us 2 crap heads for the win, and if match making was awesome and put similar levels of skill together, we probably should have won that regardless with such an extensive lead.

That was one game at my doing, but I’ve had plenty where I was one of those 3 allies just tryin to play the game for the win. Completely out of my control and completely unwinnable and THAT is unfair to a personal skill rating for anyone regardless of division or skill.

(don’t feel sorry for that guy either, over 3 games he afk the first one, raged the second and 3rd one before I’d had enough)

steelheart.7386:

We don’t need more placement matches necessarily what we need is a hard MMR reset.

Is this the new excuse when you aren’t as good as you think you are?

You can’t go to a basically solo queue league system and rely so heavily on MMR a lot of people built up playing mainly in groups in the past for initial placement.

Full resets have much more volatility. It’s harder to make good adjustments up and down when you know nothing about anyone. If you leave rating intact (or a soft reset), then any outliers who were boosted in previous seasons can easily and quickly be adjusted. Your claim is based on a an awful assumption that everyone’s rating is incredibly wrong.

hackks.3687:

“Skill rating” is accurate??? lol

It’s 5v5 and you can’t queue with a team but you are rated only on team performance. That’s 80% of the effort and decision making on your team that is entirely out of your control. That’s 80% of your rating that is not derived from your actions.

Accurate he says… ROFL

Researchers have run simulations on rating systems, including ones with random teams. For the most part, players are properly distributed based on rating, even with random teams. There’s room for improvement, but it’s not like an average player ended up near the top or bottom.

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

But if you think that skill rating truly shows your and only your skill level, then explain, how can some players with skill rating belonging to 0,5% top players drop to skill rating of 5% in over 10 games with total games played being 150+ .

actually, only 10% is you while 40% is your team and finally 50% is the opposing team.

the actual player 10% may not be very good, just average, but due to 40% good team mates and 50% average players, the average player 10% can simply be carried by the 40% good team mates versus 50% average players.

@ Lucius.2140 – just went on my 1st losing streak – 5 games, had to stop; needed to stop to keep my frustration level down and sanity – just a bad spvp night. in each match of 5 matches, I had someone who dc’d (1 match), someone who threw the match(3 matches – yes had to use in game reporting as did others on my team), and the last match was 3 guards/1 warrior and 1 necro on enemy team, while on my team we had 2 thieves, druid, Mesmer and warrior (there were at least 2 who called it a match mid way through). clearly the match making is accounting for all of that as it considers my next team and opponents.
statistics are only as good as it’s control group(s) and sometimes it’s not a reflection on the individual and sometimes it is that’s why I don’t like to discount for what others may genuinely be experiencing. simply put, the math doesn’t/can’t account for everything no matter how many matches you play, and even if it could, it cannot genuinely be a reflection of your skill level (whether good or bad) because you are always dependent upon others for your outcome.
after tonight, I cannot see how the system is working properly for all. this is the 1st and last time I will complain about it and only because it helps me make my point.

skylink.4972:

Please dont call it skill rating its so inaccurate. It could be called skill rating if it represented YOUR SKILL but it doesnt. 20% of the so called skill rating is you while 80% of it is random 4 players you get.

But if you think that skill rating truly shows your and only your skill level, then explain, how can some players with skill rating belonging to 0,5% top players drop to skill rating of 5% in over 10 games with total games played being 150+ .

Lets get real.

There is nothing in the world that can accurately represent your “skill,” at anything or else we’d just use it to put scores and values on human beings from birth.

Kreweless.2196:

hackks.3687:

It’s 5v5 and you can’t queue with a team but you are rated only on team performance. That’s 80% of the effort and decision making on your team that is entirely out of your control. That’s 80% of your rating that is not derived from your actions.

This is the argument that just about everyone (myself included at times in the past) deploys when they’re dissatisfied with their Skill Rating.

It takes an exceptional player to consistently support their team through wins. Only those players (as stated, the top 10% of thousands of players) rise out of the majority of the player-base into Platinum/Legendary.

These statistics and ratings are derived from aggregations of experiences. You can’t get a few bad matches and cry, “Foul! ArenaNet isn’t supporting my belief that I’m better than everyone!”

I think the average player is way more willing to poopoo on their teammates for mistakes and cry “MMR is broken, I didn’t get good teammates again!” than they are to examine their gameplay and get better.

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating. I think the numbers prove that. I think a lot of the stories of climbs through the rank support that proof.

I’ve already risen, and fallen, and risen, and fallen and rise and fallen and I don’t need to do it again. I’ve played more than 16k matches in this game, including against the Pro’s. I have nothing to prove to you or anyone.

I think the standard pleb is more likely to come here and try to disagree and kitten on people like they know what their talking about just to reinforce their own flimsy validations.

sephiroth.4217:

That was one game at my doing, but I’ve had plenty where I was one of those 3 allies just tryin to play the game for the win. Completely out of my control and completely unwinnable and THAT is unfair to a personal skill rating for anyone regardless of division or skill.

Bro, you stated a complaint about bad players and admitted to being one on purpose for fun, in the same breath.

I think your Skill Rating is accurate.

It’s not a reflection of your best intentions, when you feel like trying. It’s an estimate of future performance, based on past performance.

In your case, inconsistent at best (or average, being generous). I hardly think that was your first, or last, afk.

Kreweless.2196:

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating.

I guess you just have to ask yourself why you’re so emotionally invested in a PvP Badge from an MMO. Is getting that Platinum or Legendary badge really so super-duper important that you’re willing to purposefully set out to ruin a group of people’s experience?

SlippyCheeze.5483:

voltaicbore.8012:

I won’t pretend to have done enough thinking about hard resets and the inner workings of a solid Glicko system to talk about those things, but I would like to see the up-and-down of that number reflect actual performance, rather than just the win or loss.

It’s worth noting that it is not just “thinking” that backs Elo and Glicko, it’s actual hard math. Probably a mostly pedantic point, but … it’s worth noting that these have had significant thought and math and refinement applied in the real world.

Heh not sure you read carefully – I clearly noted that I am the one who didn’t think hard about those things, in no way was I trying imply that elo/glicko wasn’t refined/tested/applied math.

SlippyCheeze.5483:

Anyway, the key issue with your idea is that it builds a metric into your score changes. Once you have a metric, people play to the metric — and you get what you measure, nothing else.

If taking top stats is worth something, you get builds for it: the “top DPS” glass cannon build, the “nothing but stability, and resurrection” build, all of which contribute less to winning the match, but more to the personal score of the player involved.

The only thing that, ultimately, reflects skill without forcing particular playstyles on people is … win/lose. Not how you get there. Which is why these systems pretty much all end up falling back to it.

I’m not sure how the current system doesn’t similarly encourage particular playstyles already. The meta builds enjoy their status for a reason, in that the playerbase has discovered them to be highly effective across the vast majority of situations. In other words, meta builds help many people contribute better to wins, and wins are currently the way to gain skill rating. I don’t see how reducing skill rating loss to the extent that I suggested would really shift the already meta-heavy spirit any further.

SlippyCheeze.5483:

Also also useful to remember: no matter which game you point to as having a “better” MMR system, it’s almost certain they use the same basic math that GW2 does.

Totally agreed. Again, not sure how anything I said had anything to do with this point though. Did I miss something? Or are you just saying stuff? Genuine question.

In the end, I think I’ll ultimately have to surrender to the accuracy of the aggregate; while individual data points (spvp matches) might show fairly large swings away from where they are supposed to be, the aggregation (your hopefully 4-digit skill rating) of those points will land on target. I think I have legitimate mathematical, economic, and legal objections to applying that faith in aggregation to as many things as we do, but this is probably not the place for it.

Exedore.6320:

Saiyan.1704:

It’s why the Placement matches is an utter joke and it’s why things always start to “settle” towards the end of a season.

You’re misinterpreting the purpose of placement matches and are vastly exaggerating their inaccuracy.

Placement matches basically exist to mask the large rating shifts when your rating is relatively unknown, especially if you haven’t participated at all or in a long time. If players saw rating changes of +200 or -200, they’d freak out. By the time you finish those placement matches, your rating is relatively close to where it should be and shifts will mostly be small. There are outliers, but the majority will be in the ballpark.

It’s important to remember that rating should not be thought of as an absolute number. In Glicko2 there’s actually a second number, rating deviation, which expresses the bounds where you skill actually is. Shifting up or down 50 points from day to day is perfectly reasonable and doesn’t mean the system is inaccurate.

Another thing to note is that when you compact ratings (like the soft reset), it does take a while for players to spread out again, especially at the top and bottom. If you’ve been near the top since the start of the season, you’ve probably noticed your rating go up by 50-100 points since the start. This isn’t an affect of individual games played, but requires all players to play more games.

…I understand the function of Placement matches so I don’t need a lecture. What I meant was, Placements are utterly broken in doing what they’re suppose to do.

I was placed in Silver and made it to Plat in a month. I stopped getting 30 rating per win in silver, getting 17-20 in Silver T3. Players who did better in placements received 2x more rating per win than me at the same tier. For example, I grinded to Gold Tier 2 getting 15 to17 per win. Players who got placed in Gold were getting 30 per win just because they got a win streak in placements. Placements placed me no where near where I was suppose to be (Plat Division) and this has been the case for many, though, Devs would have those statistics.

As for not participating at all or “in a long time”.. I hit Legendary in S2 through S4… so again, other than explaining the point of Placement Matches which wasn’t even a question here… there’s not a single item you said that was marginally correct.

Saiyan.1704:

Placements placed me no where near where I was suppose to be (Plat Division)

But you got there. After about 100 matches, your Skill Rating more closely aligns with your Matchmaking Rating.

Saiyan.1704:

I hit Legendary in S2 through S4

Translation: I played a lot of games. I don’t think anyone agrees that’s a smart measure of your skill.

I did my placements on power gs/sw+sd mesmer (mistake) and am literally stuck in (rn)gold.

Multiple times I have seen:
- people stomping mesmer clones
- people repeatably crossing guard traps
- people full bursting endure pains
- people solo pushing mid v 3 people
- people rerolling off ele in a zero ele team
etc.

If you’re saying that gold is meant to represent the top 10% of the playbase that is actually super depressing.

#buildwars2

Kreweless.2196:

kdaddy.5431:

Example- I have a 1300 rating, i go on vacation and am gone for 2 weeks. I receive -700 from decay.

If you’re willing to take a two week vacation during the season, then you’re not really serious about PvP. So, the Bronze SR accurately reflects your level.
#AirTightLogic

Also, sitting at 1300 – you’re pretty much an average player.

I think the biggest complaint with the current (and in my opinion way better) badging system is that only a few players get the shiniest badges.

Chase.8415:

I don’t like the idea that if you’re average, you can get silver. Being slightly better, you’re gold. You have to exceed top 10% just to get platinum, I think the gaps are just too far.

“It’s not fair that only the top players get the top badges.”

C’mon guys, do you really all want participation trophies? Doesn’t that completely remove the prestige of earning the top badges? Maybe there are ways to game the system, but let’s tackle those instead of taking steps back to S1-4 where everyone could grind to Legendary.

It’s not really Legendary if anyone can get it.

I agree with the legendary symbol status. Its suppose to mean something.

Your other point is a bit of a joke. What if marvin quit playing today because of work or family life. Dude has more games played in NA then anyone and he is constantly in the top 100. All of a sudden he cant play for 2 weeks and he falls from 1900 to 1200 and its his fault, please stop it.

Same thing as tarcis. I saw him early in the season playing DH, he got as high as top 40 ranked player. He played 100+ games early int he season. You cant find him on the Leader Board now.

Your right the guy 11-2 in the top 50 deserves it and not tarcis a guy former ranked number 1 in solo que, former ESL competitor. By your statement, not a serious PvP player. Even though he has 100+ games more then players who are not getting decayed.

We need to have logical debates and quit making up personal stats to be applied to a season.

Kreweless.2196:

Saiyan.1704:

Placements placed me no where near where I was suppose to be (Plat Division)

But you got there. After about 100 matches, your Skill Rating more closely aligns with your Matchmaking Rating.

I think you skipped over the point of our (Exodore and I) conversation… He said Placements is a whole lot more accurate than I make it out to be. I was proving him wrong with facts.

Kreweless.2196:

Saiyan.1704:

I hit Legendary in S2 through S4

Translation: I played a lot of games. I don’t think anyone agrees that’s a smart measure of your skill.

Again, not the point but to be technical I actually hit Diamond from S2 to S4 with a 60% win rate… i’m a Diamond/Plat player. Never said I was a Legendary player I just added that in. Placements got me no where near where i’m suppose to be.

I suggest you refrain from replying if you can’t follow the conversation we were having.

Saiyan.1704:

What I meant was, Placements are utterly broken in doing what they’re suppose to do.

I was placed in Silver and made it to Plat in a month.

So you’re one out of how many? Rating systems aren’t magic and there are sometimes outliers like your example. But I’m skeptical that there wasn’t some other factor within your control causing you to end up that low.

Exedore.6320:

Saiyan.1704:

What I meant was, Placements are utterly broken in doing what they’re suppose to do.

I was placed in Silver and made it to Plat in a month.

So you’re one out of how many? Rating systems aren’t magic and there are sometimes outliers like your example.

I’m curious where your faith in these test-run Placement matches came from because so many people got screwed over because the system “wasn’t perfect.”

A DC player meant an instant loss. Some unlucky blokes have gotten 3x DC in a row. My 60% winrate means, statistically, i’ll win 6 out of my 10 games. Well I got a DC and another 1 unlucky game where some one dc’d but came back before 2m.

Meanwhile you have those certain few who duo on alt account where one person DC’s to prevent a loss… cheating the system completely. You’re right it’s not perfect, it’s down right broken.

Kreweless.2196:

sephiroth.4217:

That was one game at my doing, but I’ve had plenty where I was one of those 3 allies just tryin to play the game for the win. Completely out of my control and completely unwinnable and THAT is unfair to a personal skill rating for anyone regardless of division or skill.

Bro, you stated a complaint about bad players and admitted to being one on purpose for fun, in the same breath.

I think your Skill Rating is accurate.

It’s not a reflection of your best intentions, when you feel like trying. It’s an estimate of future performance, based on past performance.

In your case, inconsistent at best (or average, being generous). I hardly think that was your first, or last, afk.

Kreweless.2196:

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating.

I guess you just have to ask yourself why you’re so emotionally invested in a PvP Badge from an MMO. Is getting that Platinum or Legendary badge really so super-duper important that you’re willing to purposefully set out to ruin a group of people’s experience?

I didn’t start a complaint. If anything I was trying to enlighten.

I have a picture of a game that I just did then and I thought it might be relevant, you’ll notice the account name in the team chat, the leaderboards, my “accurate” skill rating combined with the match making. It automatically says there’s a problem. a BIG problem when in regards to player skill and accurate skill levels.

I’m sure you’re smart enough to see this for yourself? It’s a 600 point skill rating difference but we are on the same team, I also posted my skill rating gain, is +20 a game I’m meant to lose? On top of that is my total win loss ratio too so I can show you that matchmaking will always make me win some, then lose some, meaning I’m never going anywhere after my placements….. I’ll just leave it there.

sephiroth.4217:

Kreweless.2196:

sephiroth.4217:

That was one game at my doing, but I’ve had plenty where I was one of those 3 allies just tryin to play the game for the win. Completely out of my control and completely unwinnable and THAT is unfair to a personal skill rating for anyone regardless of division or skill.

Bro, you stated a complaint about bad players and admitted to being one on purpose for fun, in the same breath.

I think your Skill Rating is accurate.

It’s not a reflection of your best intentions, when you feel like trying. It’s an estimate of future performance, based on past performance.

In your case, inconsistent at best (or average, being generous). I hardly think that was your first, or last, afk.

Kreweless.2196:

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating.

I guess you just have to ask yourself why you’re so emotionally invested in a PvP Badge from an MMO. Is getting that Platinum or Legendary badge really so super-duper important that you’re willing to purposefully set out to ruin a group of people’s experience?

I didn’t start a complaint. If anything I was trying to enlighten.

I have a picture of a game that I just did then and I thought it might be relevant, you’ll notice the account name in the team chat, the leaderboards, my “accurate” skill rating combined with the match making. It automatically says there’s a problem. a BIG problem when in regards to player skill and accurate skill levels.

I’m sure you’re smart enough to see this for yourself? It’s a 600 point skill rating difference but we are on the same team, I also posted my skill rating gain, is +20 a game I’m meant to lose? On top of that is my total win loss ratio too so I can show you that matchmaking will always make me win some, then lose some, meaning I’m never going anywhere after my placements….. I’ll just leave it there.

WOW DUDE!!!! How dare you not to beat someone with 600+ rating? I mean if u cant beat someone who is 1800 rated you deserve to be 1200. And if you think that you are 1300 rating worthy, you have to beat that 1800 rated guy.
It all makes sense now, the skill rating accuracy its just amazing. Its all logic and math it must makes sense. 1200=1800 is the new guild wars 2 logic generation. You must let go your sanity and just accept the beauty of the logical matchmaking system.

Kreweless.2196:

kdaddy.5431:

Example- I have a 1300 rating, i go on vacation and am gone for 2 weeks. I receive -700 from decay.

If you’re willing to take a two week vacation during the season, then you’re not really serious about PvP. So, the Bronze SR accurately reflects your level.
#AirTightLogic

Also, sitting at 1300 – you’re pretty much an average player.

I think the biggest complaint with the current (and in my opinion way better) badging system is that only a few players get the shiniest badges.

Chase.8415:

I don’t like the idea that if you’re average, you can get silver. Being slightly better, you’re gold. You have to exceed top 10% just to get platinum, I think the gaps are just too far.

“It’s not fair that only the top players get the top badges.”

C’mon guys, do you really all want participation trophies? Doesn’t that completely remove the prestige of earning the top badges? Maybe there are ways to game the system, but let’s tackle those instead of taking steps back to S1-4 where everyone could grind to Legendary.

It’s not really Legendary if anyone can get it.

I only said the gaps are too far. Feel free to make legendary only accessible to top tier players, but going from “above average” to top 10% for next division is extreme.

I much prefer Overwatchs division split, platinum is average, diamond is top 25%, masters is top 10%, grandmaster is top 5%.

It used to be worse similar to gw2 where 99% of the game was gold or higher, and you needed to be top 10% just to surpass platinum. The gaps were too large, add more divisions or make different icons for each tier in good and silver, and so on.

Skill Rating is Accurate —-——-> Build Rating is Accurate.

Chase.8415:

Kreweless.2196:

kdaddy.5431:

Example- I have a 1300 rating, i go on vacation and am gone for 2 weeks. I receive -700 from decay.

If you’re willing to take a two week vacation during the season, then you’re not really serious about PvP. So, the Bronze SR accurately reflects your level.
#AirTightLogic

Also, sitting at 1300 – you’re pretty much an average player.

I think the biggest complaint with the current (and in my opinion way better) badging system is that only a few players get the shiniest badges.

Chase.8415:

I don’t like the idea that if you’re average, you can get silver. Being slightly better, you’re gold. You have to exceed top 10% just to get platinum, I think the gaps are just too far.

“It’s not fair that only the top players get the top badges.”

C’mon guys, do you really all want participation trophies? Doesn’t that completely remove the prestige of earning the top badges? Maybe there are ways to game the system, but let’s tackle those instead of taking steps back to S1-4 where everyone could grind to Legendary.

It’s not really Legendary if anyone can get it.

I only said the gaps are too far. Feel free to make legendary only accessible to top tier players, but going from “above average” to top 10% for next division is extreme.

I much prefer Overwatchs division split, platinum is average, diamond is top 25%, masters is top 10%, grandmaster is top 5%.

It used to be worse similar to gw2 where 99% of the game was gold or higher, and you needed to be top 10% just to surpass platinum. The gaps were too large, add more divisions or make different icons for each tier in good and silver, and so on.

I can agree,

I dont know why Anet changed it so much.

Amber-Emerald-Sapphire-Ruby-Diamond-Legendary

Bronze-silver-Gold-Platinum-Legendary

They took out a whole division

skylink.4972:

sephiroth.4217:

Kreweless.2196:

sephiroth.4217:

That was one game at my doing, but I’ve had plenty where I was one of those 3 allies just tryin to play the game for the win. Completely out of my control and completely unwinnable and THAT is unfair to a personal skill rating for anyone regardless of division or skill.

Bro, you stated a complaint about bad players and admitted to being one on purpose for fun, in the same breath.

I think your Skill Rating is accurate.

It’s not a reflection of your best intentions, when you feel like trying. It’s an estimate of future performance, based on past performance.

In your case, inconsistent at best (or average, being generous). I hardly think that was your first, or last, afk.

Kreweless.2196:

If you consistently support your teammates more than you don’t, you’ll rise in Skill Rating.

I guess you just have to ask yourself why you’re so emotionally invested in a PvP Badge from an MMO. Is getting that Platinum or Legendary badge really so super-duper important that you’re willing to purposefully set out to ruin a group of people’s experience?

I didn’t start a complaint. If anything I was trying to enlighten.

I have a picture of a game that I just did then and I thought it might be relevant, you’ll notice the account name in the team chat, the leaderboards, my “accurate” skill rating combined with the match making. It automatically says there’s a problem. a BIG problem when in regards to player skill and accurate skill levels.

I’m sure you’re smart enough to see this for yourself? It’s a 600 point skill rating difference but we are on the same team, I also posted my skill rating gain, is +20 a game I’m meant to lose? On top of that is my total win loss ratio too so I can show you that matchmaking will always make me win some, then lose some, meaning I’m never going anywhere after my placements….. I’ll just leave it there.

WOW DUDE!!!! How dare you not to beat someone with 600+ rating? I mean if u cant beat someone who is 1800 rated you deserve to be 1200. And if you think that you are 1300 rating worthy, you have to beat that 1800 rated guy.
It all makes sense now, the skill rating accuracy its just amazing. Its all logic and math it must makes sense. 1200=1800 is the new guild wars 2 logic generation. You must let go your sanity and just accept the beauty of the logical matchmaking system.

He was on my team and I still gained +20 so I assume the other team had even better players. It was some really good games, the sort of games that I used to play BEFORE season 5 mmr reset combined with what we called placement matches. Might have missed the mark on your sarcasm though

Bleh. Skill rating and matchmaking are horribly designed. Most everyone is stuffed into gold in crap matches.

This season was the most boring since the game launched.

Total fail.

Archon.6480:

Bleh. Skill rating and matchmaking are horribly designed. Most everyone is stuffed into gold in crap matches.

This season was the most boring since the game launched. I lost interest several weeks ago.

Total fail.