Sometimes players can overshoot their ‘true’ rating by chance, but it will not be by far and not last very long. They certainly won’t reach the top 250.
I don’t mean to pick on you specifically as the “dreaded 50% matchmaking” conspiracy has a lot of the community in its grips. Ratings obviously can’t be known unless matches are played to generate data. Do you want to play random matches and have a chaotic experience with an inaccurate rating? Or have balanced, competitive matches with an accurate rating?
Sure, we could have zero matchmaking, but two things will happen: First, your teammates will have random ratings and your opponents will have random ratings. Did you lose because you are less skilled than your opponents, or because your teammates are less skilled than your opponents? There is no easy way to tell without a lot of complicated math. “Make a system to figure it out, then!” you say. Well, Glicko does this for us already. Matchmaking assumes its data is correct by pairing opponents that appear to be evenly matched, then Glicko does its job to adjust ratings in case the outcome wasn’t what was expected. Second, random matches make your rating update verrry slowly. Every match won against a much less skilled opponent won’t make you go up because of course you were suppose to beat them, your rating must already be accurate. Every match lost against much higher rated opponents won’t make you go down because of course you were suppose to lose, your rating must already be accurate.
I honestly do not know why people think 50/50 matchmaking is a bad thing. My best guess is that people just want want something to blame when they lose.
EDIT: I agree 50/50 matchmaking sucked for the pip-based league. My comment is more general, regarding desired match quality and rating accuracy.
Really love all the work you guys have put in Evan, this is already looking like it’s leaps and bounds ahead of last season. If all goes according to plan you could end up with a very rewarding and competitive league system which would really spike the pvp population.
Regarding changes I do have a few questions and any answers you can provide (even if it’s just “I don’t know because I’m not on that team”) would be greatly appreciated:
1. Have the balance devs been doing this level of change to classes so we will see a large shakeup in game function as well or has it been small so we can expect similar play to now?
2. Are there any plans to look into class stacking within matchmaking or is that not a really big concern of the PvP team from a matchmaking POV?
3. Will the new armor skins be available only through the season or for all time? And if only for the season will the currency be common enough to allow us to get all the skins within the season?
4. How will the extra tickets and shards of glory we have from previous seasons be utilized after we have the backpack? Is there more stuff coming for us to save and spend on?
Thank you for all your time answering questions, listening to feedback, and doing these amazing changes for the game! Have a good one.
1) There will be the expected periodic balance updates, but I cannot speak to the extend of them.
2) Class stacking has not been a priority because of the leagues rework, and negative feedback regarding character locking when queuing. This might be something worth polling.
3) The vendor will always be around, but the new currency is only obtainable from the league pip system. You won’t be getting all the skins in a single season.
4) We’ve added the double currency requirement for ascended items to sink out shards of glory and will use the league tickets for more rare, longer term rewards as we make them.
Achievements id=2562 returns “bits”:[{"type":“Text”,“text”:“Ally: Canach”},{"type":“Text”,“text”:“Ally: Caithe”},{"type":“Text”,“text”:“Ally: Braham”}] , but now I have noticed that these bits are of another achievement, so for the id=2562 “bits” should be null.
Blergh, that’s super annoying. Those bits are associated with the achievement, they’re just not referenced from the achievement tracking part. Will have to add some stuff to filter those out. Tracking issue.
Achievement id =1257 is still missing as I have read from other posts on the forum.
That’s a bugged achievement — need to look into whether it is actually counted towards total AP. Empirically I’ve heard that it doesn’t. Debating on whether to whitelist it on /v2/achievements or to blacklist it from /v2/account/achievements.
I can for see one massive problem with this system so far with out playing it.
If a good player where to throw there 10 placement matches, they would be placed with the new players, then they could just farm wins and rewards for fun. What’s the incentive not to do this? You know there will be people out there who will do this if there is more to be gained from doing it.
There are three possible outcomes:
You throw your matches and then get yourself stuck in glicko hell.
You throw your matches and actually start playing poorly like continuing to make a funny face when your mom told you not to.
You throw your matches and then as you win, your rating shoots right back up again.
Rip. If the extra pips were only a thing for the winning team I don’t think anyone would go after them to the extent of hurting their teams chances of winning, but if you can get them on the losing team as well I could totally see some idiots out there trying it.
We think the top stats will be generic enough that they do not cause detrimental gameplay, but it does then we will have to adjust how top stats are given out or whether you get extra pips at all.
I totally agree with what you said about rewards. Same o’l thing. Why don’t you give us PVP players a chance to get the legendaries? Why would I need an ascended gear for? At least make some vendor with account bound skins or something.
You want the ascended armor because its a new set of skins!
1.) I know that and did write it in other threads. I would be pleased on locking class in Q :-).
If it ever happen we should get a possibility to change our builds in resp zones. If I will be locked on one class with one build that sucks because in opposite team have some hard counter to me, then there is no fun in play.
I don’t imagine we’d ever lock builds, just your profession. You could even swap to another character of the same profession since they’re share the same MMR.
How would you do this though? The allowing of a switch from character to character as long as its the same class.
If you have the same MMR on two different classes (theoretically possible, not sure how hard it would be to do exactly) wouldn’t it then open the the floor to the argument that you should be able to switch to that one too? I’m sure your initial gut response is “no” for mmr tracking accuracy (among many other variables, like team composition), but if you can accurately determine that the player is swapping from say one thief character to another, could you not also as easily determine that they swapped to an ele with an equatable mmr value and apply any changes from the match played accordingly?
Technically allowing someone to switch to a profession with the same MMR would be fine. The only reason we want to lock is so you can’t switch to something that invalidates what matchmaking just finished working with. Even if we allowed a particular MMR range of available other professions, the user experience of that feature would be bad and not worth it.
1.) I know that and did write it in other threads. I would be pleased on locking class in Q :-).
If it ever happen we should get a possibility to change our builds in resp zones. If I will be locked on one class with one build that sucks because in opposite team have some hard counter to me, then there is no fun in play.
I don’t imagine we’d ever lock builds, just your profession. You could even swap to another character of the same profession since they’re share the same MMR.
There have been some threads on the past about it, but the community always seemed split. Some liked being able to counter pick after they got into a match, or to better fill out their team comp. Others wanted the accuracy.
I do agree with the players that said profession-based matchmaking needs to be better before locking classes.
There is a leaderboard for the top 250 players (plus a friends/guild leaderboard for everyone else). The leaderboard awards unique titles at the end of the season.
Is this like the old leaderboards that showed what percentage you were at so you could aim to progress even outside the top ranks? That was a good system.
I think it was more than top 250 but I may be misrembering.
The old leaderboard was top 1000 I believe and did not show MMR, only win rate and rank.
3.) This doesn´t realy help. Your MMR is dependent on your teammates which might be hard to overcome but is not realy good.
We’ve tried using your teammates’ rating to affect your own for a while, but it ended up being less accurate and effectively created Glicko Hell. I’m sure there is a way to factor them in, but MMR and matchmaking is delicate. In my 4ish years of adjusting our use of Glicko, I’ve come to learn that most of the time less is more.
I’m not really sure what all this fuss about “skill rating” is. First off how does that mechanic even work? How do you measure someone’s skill completely without individually watching them in action? Is a player that caps more and engages in 2v1’s truly more skilled than the player holding off 3 by himself while preventing a cap? Lmao my problem with going down this path is that people will use the leaderboard to claim who is and who isn’t skilled. But if you ask me, being able to kill two players out of 3 that are fighting you by yourself is a much better skill rating than capping and defending.
I suggest possibly changing that mechanic to TEAM skill rating. There is only one way to measure skill in this game and it’s in 1v2 wager duels.
This is an enormous topic and there are no right answers. If you’re curious about approximating player skill, you can do some research on Elo, Glicko, Glicko 2, or True Skill. TL;DR an approximation is the best we can do, but end up being quite accurate.
Curious about the decay. Will it happen at every division or will the bottom divisions not be affected. For example bronze to gold would not have decay while divisions above gold would
Decay happens in every division. There is a 3 day grace period, then 7 days of decay. Each game played will remove 1 day of decay.
Last season MM was not so bad and with full premades gone i expect a lot close matches.
There will be still the usual problems.
1.) Without per class MMR skill might vary largely (especiall if daily class stays ….)
2.) With even matches progress moves to grinding. That 50:50 rate won´t please people playing for a rank.
3.) System might calculate personal MMR on a team effort… Which is one of the culprits for frustration.)
4.) Human factor will still decide a lot of matches. (AFK, drunk, trolling, toilet, cat on keyboard, ….) You will never get rid of that (exeption is 5:5 team at high level) so don´t complain to much on that ;-)
1) We track per-class MMR, but it’s not very accurate right now. I really want to use it, but without class locking on queue, we can’t use it for matchmaking.
2) This is why rewards and skill rating are split. Your skill rating will only go up if you actually improve your gameplay.
3) Your personal MMR is only updated based on win/loss and enemy MMR.
4)
What is skill rating and what affects it?
Skill Rating is your Glicko Rating plus decay. Each match has a zero or one outcome that updates your Glicko rating against the average of your opponents’ rating.
So it’s still possible to lose a close game you weren’t expected to win and still gain some glicko, kind of like how you could gain pips on a lose in Season 1?
In practice, rating does not ever go up on a loss.
I think a team of coordinated folks would be particularly over powered against equally skilled but not pre-made folks in this gank meta. I feel there should be an option for teams to go team each other if they want, but it should be sperated from those who want to solo/duo queue.
Not enough full parties play to keep an entirely separate 5-man queue healthy 24/7. However, that doesn’t mean teams shouldn’t have a structured place to participate in PvP. This will be part of the the discussion around the followup poll regarding solo/duo queue.
Last season MM was not so bad and with full premades gone i expect a lot close matches.
There will be still the usual problems.
1.) Without per class MMR skill might vary largely (especiall if daily class stays ….)
2.) With even matches progress moves to grinding. That 50:50 rate won´t please people playing for a rank.
3.) System might calculate personal MMR on a team effort… Which is one of the culprits for frustration.)
4.) Human factor will still decide a lot of matches. (AFK, drunk, trolling, toilet, cat on keyboard, ….) You will never get rid of that (exeption is 5:5 team at high level) so don´t complain to much on that ;-)
1) We track per-class MMR, but it’s not very accurate right now. I really want to use it, but without class locking on queue, we can’t use it for matchmaking.
2) This is why rewards and skill rating are split. Your skill rating will only go up if you actually improve your gameplay.
3) Your personal MMR is only updated based on win/loss and enemy MMR.
4)
What is skill rating and what affects it?
Skill Rating is your Glicko Rating plus decay. Each match has a zero or one outcome that updates your Glicko rating against the average of your opponents’ rating.
And by extent, the character model render so we can have WoW-like armory inspired websites.
I briefly looked at the time it would take to implement and shed a thousand tears.
Maybe next year ;____;
The character proportions (and face/hair selections) are something I can easily knock together though. I don’t know why I didn’t link it originally, but here’s the tracking issue for that bit.
I think the Mumble Link data has a unique id in the “shard” field which could theoretically be used — that’d be more accurate than /ip (which is not necessarily unique).
But anyway, the event state bits probably aren’t ever coming back :<
I also thought what shakeup? No per class MMR which for me would be the base to have a good MMR matching. No Leaderboard? People seem to cry for that.
Lets see how it works. It´s definitly intresting to look at. It might be much better but doesn´t seem like a huge change. Well. If it creates good matches and exitement it will be a large step forward even if it doesn´t sound so.
There is a leaderboard for the top 250 players (plus a friends/guild leaderboard for everyone else). The leaderboard awards unique titles at the end of the season.
Last season MM was not so bad and with full premades gone i expect a lot close matches.
There will be still the usual problems.
1.) Without per class MMR skill might vary largely (especiall if daily class stays ….)
2.) With even matches progress moves to grinding. That 50:50 rate won´t please people playing for a rank.
3.) System might calculate personal MMR on a team effort… Which is one of the culprits for frustration.)
4.) Human factor will still decide a lot of matches. (AFK, drunk, trolling, toilet, cat on keyboard, ….) You will never get rid of that (exeption is 5:5 team at high level) so don´t complain to much on that ;-)
1) We track per-class MMR, but it’s not very accurate right now. I really want to use it, but without class locking on queue, we can’t use it for matchmaking.
2) This is why rewards and skill rating are split. Your skill rating will only go up if you actually improve your gameplay.
3) Your personal MMR is only updated based on win/loss and enemy MMR.
I absolutely don’t get the radio silence we get on upcoming events, specifically holiday events. The harm in giving us start and end dates in advance is … ?
…. if they are delayed, due to technical issues or other causes, players can feel concern, even get upset. We prefer to give you dates in a reasonable timeframe but with minimal “risk,” as the saying goes.
You’ll be happy to know that you will get both the opening and end date for the Wintersday Festival. I agree with everyone who expressed opinions in the past that it’s very helpful to know the extent of a festival, and it’s my belief we’ll be more forthcoming about end dates in the future.
Hi mads. It’s funny you should ask, I was just writing a post about our new partnership with WeLoveFine, and with the opportunity that it gives our players to create merchandise. It seems to me if a fan creates a cool plushie, our new partner WeLoveFine may consider making it.
Looking at the Fan Forge page, I’d suggest that anyone offering items that don’t fall into the normal categories, such as hats or t-shirts, might enter their submissions in the “Accessories” section. This is a new partnership, but we believe WeLoveFine will be happy to look for entries there. If they get back to us with another suggestion about where to submit entries that don’t easily fall into the established categories, I’ll update to let everyone know.
Thanks for the suggestion. I’d dearly love to see an Aurene plushie myself!
You should read today’s blog post if you’d like to join in the new Fan Forge project we have kicked off with our partner, WeLoveFine. The article outlines an opportunity for you to be considered as a possible creative source for Guild Wars 2 merchandise. And if your work is selected, you would earn a commission on each item sold, too!
Not an artist? No worries! Visit the Fan Forge to vote on items you’d like to see.
We’re super happy about this new program, and hope to see a lot of cool new items on offer in the future!
To clarify you won’t be able to activate them while in combat.
So if you are transformed then enter combat you will not transform back to your regular model. So you can still fight as a kodan if you activate it before the fight.
We are aware of the Kodan tonic ‘spam’ and the use of tonics in WvW to trigger weapon-swap effects. I just wanted to give everyone a heads-up that in the next update the Endless Kodan Tonic will have a 5 second cool down and will not be useable while in combat. We also plan to make it so you can’t use any tonics in combat going forward, but that will be coming in a later update since doing this work to all tonics in the game is going to take more time to implement and test.
The wiki needs to be updated. binding/bound_to should be straightforward, but the stats.id references /v2/itemstats and indicates which stats were selected for a choosy-stat item.
It’s a long-requested feature; the main issue is that most of the chat (and the combat logs in their entirety) aren’t stored in a network-accessible manner — the implementation would have to use some sort of local API which is a pretty large undertaking. As such I don’t really have a timeline
We appreciate your Wintersday spirit! Information on the festival will be coming in the near future. In the meantime, don’t forget our hopes that you will share your arts and crafts on our Creative Corner forums, or on our Tumblr page! We love seeing your creations, your GW2 decorations, and anything you’d like to share. (On Tumblr, please use #GW2FanSubmission as your tag.)
I’m going to go ahead and whitelist 1257, but it’s a broken achievement (it doesn’t have a name — so the client won’t render it, IIRC). It’s … kind of strange that there are accounts with it? But it’s worth 5 AP, so there’s that.
https://api.guildwars2.com/v2/achievements/561 probably should be blacklisted because it gets converted to https://api.guildwars2.com/v2/achievements/863 upon logging in. It seems only people who did not log in between the first release of the super adventure box in April 2013 can have this achievement. (this is most likely because the achievement got an additional reward at the 2nd SAB release in 2013.
Seeing some of the IIS errors this forum gets guessing it’s built in .NET? Professionally develop .NET web applications so would gladly rip it apart and put a working search in!!!
Ruby on Rails, actually. IIS is load balancing, I think.
Do the scoreboard or other maps function at all while this isn’t working? I’m trying to establish if your client is losing a connection to all of the WvW data, or only pieces of it. If only pieces you may still be able to see changes to other maps or your scoreboard.
