Guild Wars 2

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

If you put such illogical theory into real life practice, it would not work! Simple example, if you do a team work at class (if you are a student) and your team losses, your personal rating goes down (not because of the contribution you gave but just because of the fact the whole team lost).

Wait but we forget that the team was made up of people who are new to the subject too because their rating is automatically at the same level as those who have studied this subject for many years!! That is again how illogical another aspect of MMR is… but according to ANET. You belong there, u should stay there, and we will add players new to pvp in your category and “oh it should balance out”.

So in a real life scenario, the real brains behind the losing team gets sent down to a worse team because of the reduction in MMR and so this student will keep having to work extra hard to try and make a team of potential new students win the team task (so s/he can improve their personal rating). At one point the student will just think.. W T F! This system will definitely not work in real life, not in sports, not in education and not in ANY competitive environment.

To summarise, the actual concept of MMR that is employed in GW2 PvP was flawed from the beginning. It does not reward the individual for individual’s efforts rather it punishes/rewards the individual for what a team does! And that is what the Dev’s regard as “individual’s skill”.

Only real solution
Best way out is either play in a team on teamspeak/discord (whatever) and have fun doing that or just quit PvP altogether because soloQ in a flawed MMR system is not viable.

edit: added more stuff

lead.7543:

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

Stopped reading there.

After 4 years of playing spvp consistantly 3 or 4 nights a week, I can assure you MMR is based over WHO you win and WHO you lose.

I’ve seen people with 60% win/lose ratio that were in the 90% of the ranked leaderboard MMR (skill based) system.

Nothing about simply grinding wins… Except from 2015 january and up.

You are either too new to know, or not good enough (never been ranked on old leaderboard) to feel for it.

I was able to win every match in the sapphire league last night probably out of sheer luck, Amber and emerald took much longer. I’m afraid of the ruby league now. honesty if its the type of pvp system that’s bothering people and ruining the experience there is always unranked at least until they decide to tweek it. I love it so far although

Low Ruby is the way it is because the other divisions are easier to climb out of due to lack of tier loss. Even bad players who play a lot can get a couple of lucky games to get out of tiers.

Since Ruby can’t be dropped, this causes an effect where a lot of people end up in low Ruby. This wide range of skill is contributing to the frustration.

I also agree with buffing reward tracks and using special titles etc as rewards for climbing. Legendary shouldn’t be the only devision that’s rewarded.

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

Players absolutely should remain in the divison appropriate to their skill level. That is the point of a system that places you in the league according to your skill level!

It seems to me the people who complain about getting stuck at 50% winrate are the ones who want a particular reward from league advancement. That should simply not be the end goal of ranked PVP. Make the rewards available in alternative ways that do not compromise matchmaking. Or just leave it as it is. The notion that every possible reward has to be attainable by every player does not go hand in hand with ‘’competitive PVP’’.

Jourdelune.7456:

lead.7543:

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

Stopped reading there.

After 4 years of playing spvp consistantly 3 or 4 nights a week, I can assure you MMR is based over WHO you win and WHO you lose.

That is where you are wrong! YOUR MMR is based on your team wins and losses not your personal wins. GW2 PvP has no “who YOU win and who YOU lose” at all! There is no 1v1 league so try again… Also the MMR is a rating based on teamwork NOT personal work.

MountainPanda.5831:

I was t6 diamond doing well initially, then the system teamed me back to back with the same guy throwing matches non stop. I got a lose streak and now I cannot recover because I am getting newbies non stop.

You can look at my game history and you will understand how crappy the system is. Last season at least I get different teammates to perform with. This season is same teammates consecutively for multiple matches?

Another issue is I got teamed up with people who has zero pips and they were clearing class achievements and dailies in ranked games. To these people losing is nothing, so they just troll the game with nothing to lose.

At least you got that in diamond, I got that problem back in emerald all the way through to T5 sapphire. I’ve spent over a week in this Tier. To say it’s getting frustrating would be a lie. If some of the people on my team were in the same room as me they would not have left the room with out a permanent limp and a speech impediment.

With this system theres alot of people who wont get their legendary backpack when the year has ended. Some people are litterally stuck in emerald, are they gonna be stuck in emerald everytime this year? Its just a bit unfair to them, make the legendary backpack accessable in other ways in the pve world then as well.

Evan Lesh.3295:

Yasi.9065:

Im sorry Evan, but if this is supposed to be where the majority of players should be, then why are at least 70% of the matches totally frustrating to everybody?

And please, tell me, why did I get to ruby on one account and got stuck there, whereas on another Im now nearly through Diamond?

Where exactly am I supposed to be now… ruby hell or nearly legendary?

Matches should not be volatile when you are in your ‘correct’ division, ideally. I think that is a definite area to improve on. As to why your two accounts are different, I can’t say without digging into your account histories.

I don’t know if you have any data proving what i will say but my impression is that people have in great part a bad pvp experience and you are loosing players. is it only my impression or are there any data that show this ? Problem is not the division but the fact that we too often face bad matches ….

lead.7543:

Jourdelune.7456:

lead.7543:

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

Stopped reading there.

After 4 years of playing spvp consistantly 3 or 4 nights a week, I can assure you MMR is based over WHO you win and WHO you lose.

That is where you are wrong! YOUR MMR is based on your team wins and losses not your personal wins. GW2 PvP has no “who YOU win and who YOU lose” at all! There is no 1v1 league so try again… Also the MMR is a rating based on teamwork NOT personal work.

Clearly not.

I don’t have a high w/l ratio (because I am a dedicated spvp guild leader who tried out hundreds of players of all skills level) and my MMR is better than many that have higher w/l than mine. It’s way more complex than you think it is.

https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm

At the heart of PvP matchmaking algorithm is the Glicko2 matchmaking rating (MMR). This rating, which is an approximation of your skill level, helps match you with other players with similar skill level. In addition to two core ratings (one for unranked and ranked arena), a rating is also kept for each profession. More than one rating is used in order to encourage players to experiment with other professions they may not play regularly.

https://en.wikipedia.org/wiki/Glicko_rating_system

“The RD measures the accuracy of a player’s rating. For example, a player with a rating of 1500 and an RD of 50 has a real strength between 1400 and 1600 with 95% confidence. Twice the RD is added and subtracted from their rating to calculate this range. After a game, the amount the rating changes depends on the RD: the change is smaller when the player’s RD is low (since their rating is already considered accurate), and also when their opponent’s RD is high (since the opponent’s true rating is not well known, so little information is being gained). The RD itself decreases after playing a game, but it will increase slowly over time of inactivity.”

RD is used for MMR on the Glicko page.

sanctuary.1068:

With this system theres alot of people who wont get their legendary backpack when the year has ended. Some people are litterally stuck in emerald, are they gonna be stuck in emerald everytime this year? Its just a bit unfair to them, make the legendary backpack accessable in other ways in the pve world then as well.

The idea of a PVE or WvW item that can’t be get anywhere else than spvp, is to bring more people into the game mode.

I agree that the GRIND should be different from the SKILL to get it, but that’s just me.

I want Legendary Players to be really that. Top players playing together.

Not grinders vs top… anyway grinders stop when they accomplished their goal.

This bring a second issue, we need more than that BackPack to get players into it… people like me, that play it only to get good matchup are shafted by that.

All you need to do, is separate league rewards from the league.

That way, you have all PvErs and people who are there for rewards gone from ranked. Which will lead to healthier PvP experience.

Reset MMR, scrap the League, make it only MMR based. Profit.

They said that “everybody” would be able to finish wings this year, and they are correct. If you play every season for a bit, you get your wings.

But, tieing wings to league progression wasnt that good of an idea… why not make it into a multi-tiered reward track that only progresses in league matches?

Legendary players get their title and their badge, it is after all a prestige thing. In addition, maybe convert the ticket system into a new currency which offers access to minis, outfits, fluff. Tickets should be earned for crossing prestige league divisions also.
Also, I was kinda “meh” about the prestige level not showing on the badge… would it be possible to put in the number there somewhere… or make it bigger, more glowy the higher the prestige level?

As to rubyhell… I think its safe to say there are two main problems:

1. No division loss
The system is based on the assumption that a winstreak will settle you exactly where you belong. Reality though is, that based on your luck, and on your mmr which can be a direct result of luck instead of skill, you can overshoot that target by a lot. The system then starts downadjusting, which is painful for the one experiencing it (loss streak). It also will result in others being downadjusted with him that actually belong there or a bit higher (w-w-l-w-w-l scenario). It is after all, a team based game.
So, the streak mechanic lets some overshoot, other just simply grind their way into ruby. If you play enough matches, you will make it into ruby. But not beyond. These players cant be downadjusted anymore, and they just congregate there. Which means, instead of settling along the ladder, these players will cause the system to stay volatile. They can still climb, if they are matched with someone that is good but has a bad mmr, or if they team up with good players. Which means, ruby hell can expand well into diamond, my guess is something around diamond 5 atm.
It can be solved by increasing tier size in sapphire to 6 (4×6 pips), and adding “can loose division” starting with ruby.

2. Team-based mmr instead of performance-based mmr
So, you are good, your mmr is bad. For whatever reasons. Doesnt matter. But, you are good enough to slowly climb. That means you are good enough to carry or motivate 4 other players in your team to a win. Now you carried that double-axe warrior just running onto close and instantly dieing there to a win and elevated him thereby. He gains a pip he shouldnt get, his mmr gets upwards adjusted, which shouldnt happen.
His mmr now reflects a level of skill he doesnt have, he gets into a pip-range he doesnt belong. The system will probably settle him down again which, if you are unlucky, will be alongside yourself. So you carried him to a win, maybe even five wins, and as a thank you the system says “he doesnt belong, begone”. Your bad luck is, that you are still in the same mmr range as him and you get dragged down with him.
Why? Because you can only carry to a certain point. The closer you get to your own limit, the less you are able to carry…. now you might think… nobody can be that unlucky… to get the same guy over and over… but the thing is… there are other teams, other players that carry, and even more players that get carried. Winstreaks exaggerate this problem by a lot. It inflates mmr of genuinely bad players.
I dont really have a solution for this problem, tbh. There is no sure indicator as to what makes a player ballast, and what makes him a carrier.
Maybe, not adjusting mmr quite as severly when loosing closely might work, or maybe looking at topstats might work too. But the system is just not suited for soloqueue… for changing teams.

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

I love how leagues became more about skills but there is still huge flaw.
Make matchmaking based on divisions/pips. Add placements matches at begin of the season to more “better” players to upper divisions. Right now matchmaking just causes a lot of lopsided matches, even in diamond and legend. Those are hardly fun for either team.

No matter what divisions you are in, you will be always either on totally winning or totally losing team. Nothing in between.

Jourdelune.7456:

https://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm

Yes, you are not the only one that has studied the PVP MM Algorithm and there is nothing in this algorithm that takes your personal actions within a match into consideration. Maybe I am speaking a totally different language to you or you are not listening. This MMR is not based on your individual actions, death, heals, damage dealt, kills, and more in a match per profession. It is not based on that at all. SO looking at the PvP algorithm please explain how a player can actually show he is skilled?

SOME factors that show skill include:
Damage Dealt/Received
Condition Damage Dealt/Received
Conditions Applied
Boons Removed
Boons Applied to Self/Allies
Conditions Removed from Self/Allies
Healing to Self/Allies

Where is this in the PvP MM algorithm?

I challenge you to show how the PVP MM Algorithm shows you are better at 1v1 using a profession than another profession or what you are k/d (or damage output vs damage received or even heal output to damage received) ratio per match is factored in to the algorithm to ascertain your skill. Show how the algorithm calculated for tactics or strategy.

edit
p.s. If you look at https://wiki.guildwars2.com/wiki/Talk:PvP_Matchmaking_Algorithm you will see me discussing this algorithm with Justin ODell in 2014 under the section penalty system
so… yes you are not the only one that has studied this algorithm.

lead.7543:

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

so do you think it would work to change the mmr to be based on match stats like kills, resses, deaths, damage taken/mitigated, condis removed, outgoing healing, damage done etc.? maybe have a box check option (for tank, dps, support) in the build panel window so the mmr will judge your game outcomes appropriately.
imo it doesnt make any sense to have an mmr based on wins, losses, and d/c’s.

edit
oh i just read your above post it looks like thats what youre leaning towards. lol how did anet think their current scoring system would work?

Stand The Wall.6987:

lead.7543:

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

so do you think it would work to change the mmr to be based on match stats like kills, resses, deaths, damage taken/mitigated, condis removed, outgoing healing, damage done etc.? maybe have a box check option (for tank, dps, support) in the build panel window so the mmr will judge your game outcomes appropriately.
imo it doesnt make any sense to have an mmr based on wins, losses, and d/c’s.

edit
oh i just read your above post it looks like thats what youre leaning towards. lol how did anet think their current scoring system would work?

MMR should not be based on just wins or losses (or just kill/death). GW2 PvP is more complicated than that and as such MMR does not reflect this at all. It needs to be more intricate and precise to really judge a player’s skill.

First I wondered why have the devs not written an algorithm to analyse true skill… is it difficult? But the data they have from matches allows for “big data analytics” to be employed to get a suitable MMR system however it has never been done. Maybe it is due to lack of resources in-house? Whatever the case, MMR needs to be totally updated as it does not in any way reflect any player’s true skill.

edit: typo

lead.7543:

SOME factors that show skill include:
Damage Dealt/Received
Condition Damage Dealt/Received
Conditions Applied
Boons Removed
Boons Applied to Self/Allies
Conditions Removed from Self/Allies
Healing to Self/Allies

I don’t think this would be good idea. You can have person spam conditions all match but have no impact on a game. And you can have someone who is cleverly using their skills, like perfect timed fear to prevent stomp or perfect time heal etc…
Basically what I’m saying is you cant show statistically someones skill in this way, like if you spammed more of stuff heals boons etc, but who is using it more smartly, who is rotating better, and this you cant catch with your statistics.

lead.7543:

The real elephant in the room is that MMR is based on wins, losses and disconnects. It is not based on what the player actually does in a match on a specific profession. MMR is flawed so what Evan says about players staying in a division based on MMR is extremely flawed.

If you put such illogical theory into real life practice, it would not work! Simple example, if you do a team work at class (if you are a student) and your team losses, your personal rating goes down (not because of the contribution you gave but just because of the fact the whole team lost).

Wait but we forget that the team was made up of people who are new to the subject too because their rating is automatically at the same level as those who have studied this subject for many years!! That is again how illogical another aspect of MMR is… but according to ANET. You belong there, u should stay there, and we will add players new to pvp in your category and “oh it should balance out”.

So in a real life scenario, the real brains behind the losing team gets sent down to a worse team because of the reduction in MMR and so this student will keep having to work extra hard to try and make a team of potential new students win the team task (so s/he can improve their personal rating). At one point the student will just think.. W T F! This system will definitely not work in real life, not in sports, not in education and not in ANY competitive environment.

To summarise, the actual concept of MMR that is employed in GW2 PvP was flawed from the beginning. It does not reward the individual for individual’s efforts rather it punishes/rewards the individual for what a team does! And that is what the Dev’s regard as “individual’s skill”.

Only real solution
Best way out is either play in a team on teamspeak/discord (whatever) and have fun doing that or just quit PvP altogether because soloQ in a flawed MMR system is not viable.

edit: added more stuff

By your logic all team based competitive games do not work properly then. Seriously DOTA2, LoL, CSGO are all TEAM games and yet they use MMR just fine to match people of equal skill. I can tell you 100% that your MMR in DOTA2 has ZERO bearing on the specifics of any given match, all that matters is the W or L at the end of it.

People in this game suffer from some serious dunning-krugar effects where all negative outcomes are the result of either their team or the system and all positive outcomes are forgotten or attributed to themselves. I read this thread and see responses about how losing a few matches in a row makes them never want to PvP again and can only think… you shouldn’t be PvPing at all if you cant take a loss. This isnt PvE where you can memorize a pattern and then just robot each fight after that and win. If you want your feel goods then PvP is NOT for you.

The discussion on whether the rewards should be grindable and decoupled from the league divisions are a good start and will ameliorate the hurt feelings of some players who are only here to get a reward. This would make it exactly like another reward track, however, and so there is some discussion around the duplicate functionality.

Duncanmix.5238:

lead.7543:

SOME factors that show skill include:
Damage Dealt/Received
Condition Damage Dealt/Received
Conditions Applied
Boons Removed
Boons Applied to Self/Allies
Conditions Removed from Self/Allies
Healing to Self/Allies

I don’t think this would be good idea. You can have person spam conditions all match but have no impact on a game. And you can have someone who is cleverly using their skills, like perfect timed fear to prevent stomp or perfect time heal etc…
Basically what I’m saying is you cant show statistically someones skill in this way, like if you spammed more of stuff heals boons etc, but who is using it more smartly, who is rotating better, and this you cant catch with your statistics.

That is why I highlighted “SOME” and the whole of the stats derived should be looked. If you spam condi etc and cant make a difference, the stats will show such information and an algorithm can be easily developed to see the impact of the spam and therefore how useful it was to the team and the match.

I like how all these people with no real evidence other than their own experiences or the experience of a friend of a friend are telling someone who has access to actual stats and metrics that their system isn’t working the way they intended. It is pretty obvious that average players are only supposed to make it to ruby, if you are stuck there its probably where you belong and no amount of grinding should allow you to escape it unless you are playing better.

It should be a requirement to post a video of your game play in every whine thread, a lot of people would probably be exposed as the reason they are losing so much or stuck in “MMR hell”.

Also a lot of people complain about the afkers or match throwers and that has nothing to do with the match making. People are going to be kittens whether you like it or not, if they end up on your team that sucks but it is more a problem with the reporting feature.

lead.7543:

That is why I highlighted “SOME” and the whole of the stats derived should be looked. If you spam condi etc and cant make a difference, the stats will show such information and an algorithm can be easily developed to see the impact of the spam and therefore how useful it was to the team and the match.

An algorithm like this cant be easily developed. If it was easy, other games would do it by now. All giant competitive games use w/l ratio in their core and yes this is challenge for great minds and is def not an easy developed thing as u assume.

First, thank you Jourdelune for the explanation of what mmr is. Now the important question is: Do you think it works as planned?

If I go back to last season where both teams were in my mmr range (if I understood it correctly). A game on the forest map started with one player running to close and 4 to the forest creature. After the forest creature was killed, both teams battled for mid. However, when I entered ruby (division 4), the dominant strategy changed entirely. Now 1 player went close and 4 players went mid. Once close was captured, the player from close killed the forest creature.

I dont want to start a discussion about the optimal strategy. Let us for now assume that it is the division 4 strategy. Let us furthermore see my annecdote as a signal for better players in division 4, worse beyond. Last season we got ranked against players in our mmr and in our division (roughly). What happens now with my mmr in this situation? Since I play against better opponents and since there are some players improving strongly and enter diamond, I start losing more than I win. Hence, my mmr drops (as I said: paired against similar mmr). Let us now compare my mmr before season 1 with after season 1. Am I higher, lower or equal?

Now the question again: Is mmr a fair representation of your skill? Do you think, I am comparable before and after season 1?

@Evan Lesh:
You need to keep two things in the back of your mind (and this will be hard now, sry):
1) Guild wars is a product on a demand market. The players can easily switch to e.g. Aion, Blade&Soul (yes, I products of ncsoft). In other words: The customer is ALWAYS right.
2) A product is only as good as it is perceived (not as it actually is). That is something you learn in marketing pretty early. In other words: You may believe you created an ideal product but as long as the customers do not recognize it as such, you do NOT have an ideal product. If you believe it is the customer’s fault, please read 1) again.

If you do not believe I am right, maybe the following conversation of Yoon Jae-soo (a leading person of ncsoft) and an analyst about the HoT introduction will help:
Question:" Yes, I’d like to ask you a question about Guild Wars 2; it seems that the performance of the expansion pack is a bit weaker than what you had expected. So in terms of the business model, do you have any strategies, for example, to beef up items, in-game item sales, or any strategy to do that?"
Yoon Jae-soo:“(…)if you look at the actual expansion pack versus what our expectations were at the launching time, versus what the response has been from our users, I think that there is a gap that we need to address.”

Duncanmix.5238:

lead.7543:

That is why I highlighted “SOME” and the whole of the stats derived should be looked. If you spam condi etc and cant make a difference, the stats will show such information and an algorithm can be easily developed to see the impact of the spam and therefore how useful it was to the team and the match.

An algorithm like this cant be easily developed. If it was easy, other games would do it by now. All giant competitive games use w/l ratio in their core and yes this is challenge for great minds and is def not an easy developed thing as u assume.

Fortunately I have worked with teams that develop similar algorithms, not for games but real world data and that is why I wrote in a response above:

lead.7543:

Maybe it is due to lack of resources in-house?

It can be done but maybe the cost benefit of employing resources to develop such algorithm may not be the best business case for a company such as ArenaNet. However on the flip side, if they do make such an algorithm, and it works they can get their ROI by licensing such algorithm to other game developers. This is a totally different topic.

lead.7543:

Fortunately I have worked with teams that develop similar algorithms, not for games but real world data and that is why I wrote in a response above:

Ok great then enlighten me, how would u go about this problem specifically in gw2? Lets try to break it to an easy problem:
How would your algorithm realize which stun or daze for example had greater effect for the outcome of the match? When there is infinite amount of different stuns or dazes that can happen on different players, different places, different times. Now once you solve this “easy” problem try to work when there is second player in game who cc your target so your will easily chain stun him. Is now your stun less valuable then someone who stunned without help? What if your stun was not needed how will your algorithm recognize this?

displacedTitan.6897:

(…)
By your logic all team based competitive games do not work properly then. Seriously DOTA2, LoL, CSGO are all TEAM games and yet they use MMR just fine to match people of equal skill. I can tell you 100% that your MMR in DOTA2 has ZERO bearing on the specifics of any given match, all that matters is the W or L at the end of it.
(…)

Lets look at this statement a bit closer. Dota, CSGO cant be compared to gw2 simply because professions dont matter there. No skill diversity either. A win says a lot more about the pure skill of a person there since it doesnt get diluted by skill and profession balance. Its like comparing dragon ball arena to spvp. You should see that yourself.

LoL has softresets and promo games to solve their issues with mmr getting inaccurate the longer a system is running. Also, they dont allow players too far apart to team up and they dont start new players on an average mmr.

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

This post seriously seems like an inflammatory “git gud” post. Mechanical skill doesn’t matter in a game that’s more about how well you can outrotate the other team, and what comp your team is running. Face it, Evan. Your team has made a system that caters to ESL level players that are premade on TS.

There are SO MANY builds and classes that are just completely irrelevant this season. How does it show skill that I can get on a Necro (that I rarely play) and faceroll a mid fight and win just because Necro’s are just OP teamfighters? This game is really lacking in the “skill” and rewards department and it’s severely pushing vets (like me) away from y(our) game.

Fix the matchmaking. Base the enemy team on MMR as well. Make crossing divisions require a certain MMR. Increase MMR if you win. Decrease it if you lose. Divisions can be lost. Increase MMR more if the team you’re facing has a much bigger MMR and your team should statistically lose. Don’t decrease the MMR if your team was statistically supposed to lose.

I really don’t understand why your team doesn’t get this (not even) novel concept. It’s like you guys have never played ranked games before.

The system you’re trying to put on the majority of players caters to the minority. The system would work if there were more people playing this game, but thanks to your team not being able to balance to save your careers or add a decently rewarding system, or making a PvP lobby worth staying in, you DON’T have the population to make this system work.

Sorry not sorry for seeming angry. I just feel like y’all are so kitten out of touch with your playerbase. I hope for Anet’s sake that things change, and change quick, because there are other games out there, and more coming soon, that will REALLY hurt this one if things aren’t made better. (See: Crowfall, Black Desert Online, Over Watch, Paragon)

Yasi.9065:

displacedTitan.6897:

(…)
By your logic all team based competitive games do not work properly then. Seriously DOTA2, LoL, CSGO are all TEAM games and yet they use MMR just fine to match people of equal skill. I can tell you 100% that your MMR in DOTA2 has ZERO bearing on the specifics of any given match, all that matters is the W or L at the end of it.
(…)

Lets look at this statement a bit closer. Dota, CSGO cant be compared to gw2 simply because professions dont matter there. No skill diversity either. A win says a lot more about the pure skill of a person there since it doesnt get diluted by skill and profession balance. Its like comparing dragon ball arena to spvp. You should see that yourself.

LoL has softresets and promo games to solve their issues with mmr getting inaccurate the longer a system is running. Also, they dont allow players too far apart to team up and they dont start new players on an average mmr.

So you are saying in DOTA2 that different heroes are actually the same? That a hard carry is the same as a support? Are you saying that item builds don’t matter to the outcome? Do you think there are no overpowered heroes (especially in ranked non-ban games)?

I think you might be the one who thinks this game is a special snowflake and doesn’t realize that the comparisons are very valid.

Let me start by saying that I have always hated PvP because it was so frustrating for MANY reasons. Reason 1, I am not very good at it and getting owned both 1 v 1 and 1 v many is just the opposite of fun. Reason 2, the players in PvP can be real tools (nothing can be helped about that except to close the chat window). Reason 3, team members often have no clue what the proper ‘tactics’ should be as far as capping points, etc…

The net is PvP is only FUN if you have a CHANCE to win.

My friend, who is much better than I am suggested I try ranked play this season because I would not have to face all of the ‘pros’ and would experience a much higher percentage of fair matches. I decided to give it a try and in very short order I was in Sapphire and getting there was pretty easy. Since then however it has been two weeks of total frustration with my high water mark of four pips in tier 1. I have done some solo but mostly teamed with my friend and maybe 10% of the games I have played where balanced. About 20% where won easily and the other 70% there was just no chance of winning. Every player on the other team was about the same skill level or MUCH better, and no matter where I tried to rotate to, the other team always seemed to have more players there. Feeling completely outclassed is NOT FUN.

This matching system has just about convinced me to give up for this season. Balancing the teams within tier/division should be the goal NOT pushing better players up and out because all the losses, through no real fault of the player, send your ranking artificially down making balanced match making even that much harder. As an entry level Sapphire I should NOT be facing top level Sapphires trying to get into Ruby.

Duncanmix.5238:

lead.7543:

Fortunately I have worked with teams that develop similar algorithms, not for games but real world data and that is why I wrote in a response above:

Ok great then enlighten me, how would u go about this problem specifically in gw2? Lets try to break it to an easy problem:
How would your algorithm realize which stun or daze for example had greater effect for the outcome of the match? When there is infinite amount of different stuns or dazes that can happen on different players, different places, different times. Now once you solve this “easy” problem try to work when there is second player in game who cc your target so your will easily chain stun him. Is now your stun less valuable then someone who stunned without help? What if your stun was not needed how will your algorithm recognize this?

Actually, I am writing something along the lines of combat stats analysis as a tool for MMR. But I asked you a question first and you are yet to answer it. First answer it and I will take some time in returning the favour.

Thank you Evan for finally telling these casuals that the reason they are in losing streaks in emerald is because they need to IMPROVE SKILL. You can only climb so much to the point where you need to fix your own gameplay instead of whining. I consider myself to be above average as I almost made it into season pro league, and am legendary rank which sort of supports what I say next. Legendary division feels a lot better this season because the games are closer, the people are better, and unfortunately the queues are longer but its all worth it because I know the people in legendary didn’t have to go on a grind fest. Most of the people I know in legendary didn’t lose a single match until they hit ruby. It’s very simple to climb you just need to IMPROVE your gameplay. /caps typing from phone sorry for grammar

displacedTitan.6897:

Yasi.9065:

displacedTitan.6897:

(…)
By your logic all team based competitive games do not work properly then. Seriously DOTA2, LoL, CSGO are all TEAM games and yet they use MMR just fine to match people of equal skill. I can tell you 100% that your MMR in DOTA2 has ZERO bearing on the specifics of any given match, all that matters is the W or L at the end of it.
(…)

Lets look at this statement a bit closer. Dota, CSGO cant be compared to gw2 simply because professions dont matter there. No skill diversity either. A win says a lot more about the pure skill of a person there since it doesnt get diluted by skill and profession balance. Its like comparing dragon ball arena to spvp. You should see that yourself.

LoL has softresets and promo games to solve their issues with mmr getting inaccurate the longer a system is running. Also, they dont allow players too far apart to team up and they dont start new players on an average mmr.

So you are saying in DOTA2 that different heroes are actually the same? That a hard carry is the same as a support? Are you saying that item builds don’t matter to the outcome? Do you think there are no overpowered heroes (especially in ranked non-ban games)?

I think you might be the one who thinks this game is a special snowflake and doesn’t realize that the comparisons are very valid.

Ok tell me, in gw 2 can you get fed to become unbalanced and 1v5?
Can you have such aim to 1v5?

Mmo ll never be competitive. Period.

Is this chess where you 1v1?

Is it even team q only ?

Jelzouki.4128:

Thank you Evan for finally telling these casuals that the reason they are in losing streaks in emerald is because they need to IMPROVE SKILL. You can only climb so much to the point where you need to fix your own gameplay instead of whining. I consider myself to be above average as I almost made it into season pro league, and am legendary rank which sort of supports what I say next. Legendary division feels a lot better this season because the games are closer, the people are better, and unfortunately the queues are longer but its all worth it because I know the people in legendary didn’t have to go on a grind fest. Most of the people I know in legendary didn’t lose a single match until they hit ruby. It’s very simple to climb you just need to IMPROVE your gameplay. /caps typing from phone sorry for grammar

Wait, wait ,wait… you are an above average player that has legendary rank and you see no problem with that knowing legendary is supposed to be were your top percentile is supposed to end-up in a purely skill based system? As in, not “above average ppl” should be there. You sure the system works if legendary include above average players like you instead of putting them in ruby where above average actually would put someone using a normal curve?

Jelzouki.4128:

Thank you Evan for finally telling these casuals that the reason they are in losing streaks in emerald is because they need to IMPROVE SKILL. You can only climb so much to the point where you need to fix your own gameplay instead of whining.

You will fix nothing when your team is bad. And if you lose because one leaves or someone is trolling or just because the other team in the same division wins with 500:50 there is nothing you can do. You can improve your gameplay, but you still we be stuck with players who lost their last matches.

How about making sapphire able to lose tiers too..

Far too many people thinking that just because there is a reward, THEY ARE ENTITLED TO MAKING DIAMOND+ just for playing a metabuild. <ESP when those ppl hardly or didn’t at all pvp before seasons -.-

Yes thank you Evan. We all feel much better now that you have told us the truth. Funny how it is a completely different truth than in the previous MMR based ladder. I guess the truth is a flexible concept.

Obviously you can’t please all of the people all of the time. But at least you have managed to please a tiny minority some of the time. So three cheers for that.

lead.7543:

Actually, I am writing something along the lines of combat stats analysis as a tool for MMR. But I asked you a question first and you are yet to answer it. First answer it and I will take some time in returning the favour.

I didn’t notice any question, do u mind pointing it out for me?

CrAzY.3275:

How about making sapphire able to lose tiers too..

Far too many people thinking that just because there is a reward, THEY ARE ENTITLED TO MAKING DIAMOND+ just for playing a metabuild. <ESP when those ppl hardly or didn’t at all pvp before seasons -.-

I m up for that, but as well they should, reset MMR for everyone, separate ques solo/team, remove rewards which are tied to the leagues, create qualify matches in between devisions. Divisions should reflect MMR ranges. Matchmaking – You are getting teamed up with players within your MMR range, opponents in similar range MMR. Make the season last half of the year, or just make it permanent.

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

I dont believe the MMR part one bit. Currently 75% of my games already start with a disadvantaged team composition. Like 2-3 thieves, guards and warriors on team.
Today I even had a game with 2 thieves and 1 relogged thief so we had 3 thieves in total. Needless to say we lost hard. Outrageous matchmaking for diamond division.

Malediktus.9250:

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

I dont believe the MMR part one bit. Currently 75% of my games already start with a disadvantaged team composition. Like 2-3 thieves, guards and warriors on team.
Today I even had a game with 2 thieves and 1 relogged thief so we had 3 thieves in total. Needless to say we lost hard. Outrageous matchmaking for diamond division.

it was thief daily. same kitten happen to me. the player said his main is not thief he just wants daily. so we lost

In this season, I notice that you no longer gain 2 pip when winning against stronger players. Why is that so?

While I was climbing from sapphire to break into ruby, I had couple of matches where my team full of sapphire winning teams full of ruby. A team of sapphire beating a team of ruby, shouldn’t they be given greater rewards for that? People are really kittened when they beat supposedly higher ranked opponents but they are stuck at the lower rank themselves. There should be incentive for beating stronger team and punishment for losing to way weaker team.

Malediktus.9250:

I dont believe the MMR part one bit. Currently 75% of my games already start with a disadvantaged team composition. Like 2-3 thieves, guards and warriors on team.
Today I even had a game with 2 thieves and 1 relogged thief so we had 3 thieves in total. Needless to say we lost hard. Outrageous matchmaking for diamond division.

This is pretty much what happened to me all day. I played around 20 matches, lost 15 or 16 because we had 3 ranger, 1 warrior or 4 necros or groups of three to four people on my team that fought 4v2 on the far node until the match ended – while telling me how to play.
Of that 20 matches today there may have been three which didnt end up a 500:100 or 500:40 loss (or win).

If the matchmaking is about “We put you in a group full of idiots when you enter ruby” it’s working fine.

messiah.1908:

Malediktus.9250:

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

I dont believe the MMR part one bit. Currently 75% of my games already start with a disadvantaged team composition. Like 2-3 thieves, guards and warriors on team.
Today I even had a game with 2 thieves and 1 relogged thief so we had 3 thieves in total. Needless to say we lost hard. Outrageous matchmaking for diamond division.

it was thief daily. same kitten happen to me. the player said his main is not thief he just wants daily. so we lost

then those people should do the relog trick and relog at 450 points or something when winning

Malediktus.9250:

Evan Lesh.3295:

Getting ‘stuck’ in ruby is a very important topic regarding how best league should operate. We changed matchmaking to make the ladder more prestigious and so far it has done that. Divisions are aligning with skill level much better than season 1. The product of this change is that the bulk of players who hover around the middle of the mmr curve should stay in sapphire and ruby. Win rate will level out at 50% because there are plentiful players of similar mmr in the same division.

Should average players be able to climb out of the middle divisions by grinding? This removes prestige and can make matchmaking worse within a division, but always allows a sense of progression.

Should average players stay in the middle divisions because that’s where they belong at their skill level? This retains prestige and only improving skill will allow climbing, but rewards and progression stop at some point through the season.

Is there an incentive that could be added to league games for players who have reached their appropriate division and can no longer climb?

I dont believe the MMR part one bit. Currently 75% of my games already start with a disadvantaged team composition. Like 2-3 thieves, guards and warriors on team.
Today I even had a game with 2 thieves and 1 relogged thief so we had 3 thieves in total. Needless to say we lost hard. Outrageous matchmaking for diamond division.

You don’t have to.. all you have to do is look at the algorithm and what he is saying, there is no corrolation.

[/quote]
It’s amazing how your argumentation regarding the system not being a significant problem at all is rather “light”…[/quote]

Ok, really have a problem in the system and this problem can be solved, allow it to fall division, an example from the emerald can return to amber.

It also leaves the 5 pips to spend division and then we’ll see less people complaining of the players for high division as ruby (since they may regress to other divisions below), and with 5 pips to spend tier complicates a little more and proof who really should evolve division or not.

This thread and this league season embody the contradiction between setting up a competitive ranking system vs. being a private company that needs to please its customers.

On the one hand, a competitive system seeks to show who is the best and thus requires that many have their egos shattered, but a private business model requires an organization to placate customers to keep them satisfied and this requires feeding your customer’s ego for revenue.

It’s interesting to watch really.

Kuya.6495:

This thread and this league season embody the contradiction between setting up a competitive ranking system vs. being a private company that needs to please its customers.

On the one hand, a competitive system seeks to show who is the best and thus requires that many have their egos shattered, but a private business model requires an organization to placate customers to keep them satisfied and this requires feeding your customer’s ego for revenue.

It’s interesting to watch really.

Very true, especially when this competitive system groups all good players together and match them against bad players. A sight to be seen indeed.

Fivedawgs.4267:

Kuya.6495:

This thread and this league season embody the contradiction between setting up a competitive ranking system vs. being a private company that needs to please its customers.

On the one hand, a competitive system seeks to show who is the best and thus requires that many have their egos shattered, but a private business model requires an organization to placate customers to keep them satisfied and this requires feeding your customer’s ego for revenue.

It’s interesting to watch really.

Very true, especially when this competitive system groups all good players together and match them against bad players. A sight to be seen indeed.

Of course. Professional sports teams only choose the best in order to compete with other teams composed of the best players.

A model that mixes and matches players of varied skills levels, sets them against teams of similar aggregate skill level and artificially seeks to create a 50% win rate is not meant to be competitive, but to create a false sense of competency on the part of the customer and make them feel satisfied.

In theory, a 50% win rate should be reached eventually as every player reaches the division that matches their skill level, the question is, if anet wants to keep leagues competitive, whether players are reaching divisions that reflect their skill level fast enough. IE: where they don’t lost every match they play for weeks straight. Or, how long should a average or low skill player be expected to suffer through losing streaks until he finally reaches the division he is meant to be in according to skill level where his win rate approaches 50%?

Of course, if anet were to find the answer to that question, it still wouldn’t stop players who reach the division that matches their skill level from complaining that they don’t feel their skill level deserves such a division. You would still get ruby players saying they shouldn’t stay in ruby.

Fivedawgs.4267:

Kuya.6495:

This thread and this league season embody the contradiction between setting up a competitive ranking system vs. being a private company that needs to please its customers.

On the one hand, a competitive system seeks to show who is the best and thus requires that many have their egos shattered, but a private business model requires an organization to placate customers to keep them satisfied and this requires feeding your customer’s ego for revenue.

It’s interesting to watch really.

Very true, especially when this competitive system groups all good players together and match them against bad players. A sight to be seen indeed.

Kuya has totally missunderstood this league. It wouldnt even have half the players if they didnt hide a legendary backpack behind it. Me for instance, i stopped basically playing pvp for many months bc its totally infuriating in a way ive never experienced in another pvp game before. I came back to play pvp because i want the backpack. I dont care at all about the rank rly. You do realise about 10% of the playerbase is hardcore pvp’ers. Without the casuals there would almost not be any pvp alive anymore and this league with a hidden legendary backpack was what would attract the pve’ers to play pvp (because its dying). Now when anet has switched the matchmaking system to a kittened level, where many pve’ers now never will see a glimpse of the pvp legendary backpack…do you rly think they will continue next seasons? Who will suffer the most? The pvp player base.

Kuya.6495:

Fivedawgs.4267:

Kuya.6495:

This thread and this league season embody the contradiction between setting up a competitive ranking system vs. being a private company that needs to please its customers.

On the one hand, a competitive system seeks to show who is the best and thus requires that many have their egos shattered, but a private business model requires an organization to placate customers to keep them satisfied and this requires feeding your customer’s ego for revenue.

It’s interesting to watch really.

Very true, especially when this competitive system groups all good players together and match them against bad players. A sight to be seen indeed.

Of course. Professional sports teams only choose the best in order to compete with other teams composed of the best players.

A model that mixes and matches players of varied skills levels, sets them against teams of similar aggregate skill level and artificially seeks to create a 50% win rate is not meant to be competitive, but to create a false sense of competency on the part of the customer and make them feel satisfied.

In theory, a 50% win rate should be reached eventually as every player reaches the division that matches their skill level, the question is, if anet wants to keep leagues competitive, whether players are reaching divisions that reflect their skill level fast enough. IE: where they don’t lost every match they play for weeks straight. Or, how long should a average or low skill player be expected to suffer through losing streaks until he finally reaches the division he is meant to be in according to skill level where his win rate approaches 50%?

Of course, if anet were to find the answer to that question, it still wouldn’t stop players who reach the division that matches their skill level from complaining that they don’t feel their skill level deserves such a division. You would still get ruby players saying they shouldn’t stay in ruby.

I agree that why I have always wondered why folks call games such as lol, Dota or CS:GO competitve. I can’t belive they still use the 50/50 formula, tbh that’s NOT competitve. You only get competitve games when 7 is match against a 3. 5 vs 5 is just ridiculous. Hopefully, those games will learn from us.