Guild Wars 2

Pixels.6532:

30% is not as good as 0%. Just Sayin’.

Actually, it’s 25%, I was corrected in another thread. Dropping trolling from 100% to 25% with one feature? I’d count that as a win, but you can’t please everyone.

Deimos Tel Arin.7391:

what are the odds of you releasing the map editor tools to us, so we can fix the map ourselves?

Basically nil. The tools we use are very customized to a particular workflow, and there’s a lot of tribal knowledge involved as I understand it. We’d need to allocate a fairly large team dedicated to just this.

That said, I think it would be a lot of fun if we ever held some sort of map concept contest. I believe something like that was done for weapons in GW1.

SaintSnow.6593:

Couldn’t it just allow someone already in que to join by giving someone qued up a message to join in a on ongoing match?

The system will do that if someone declines or fails to confirm for the match.

I don’t think we want to do that for mid-match DCs. It would means pulling a player into a match in progress, which can be disorienting. There is also a good chance they’d be put on the losing team as the other team has had an advantage while this is happening. This just wouldn’t be fair to that player. Sure we could ask, but what if no one says yes?

Also, what do we do with the player that DC’d? Do they get dishonor? Do we give them a grace period?

What about the other players? Do we keep them idle until a replacement is found? How long do we wait? What happens after that?

There are other options as well. We could just cancel the match and no one gets ladder credit.

It might be doable, but we’d need a good user experience for all involved.

Lopez.7369:

Would you consider more extreme measures, like removing the breakable panels or making them take even longer to break?

The last update included making them take longer to break, and made the canon dodge-able. Are you asking if we’d make it even longer? I’m not sure. This is more a question for the designers.

No, nothing Dec 2nd, other than all the ways we are trying to prevent the 4v5. What ways would you guys like to see this?

I believe some people suggested queuing respawns to compensate, others suggested adjusting the prediction rules. What ever we do, if anything, would need to be exploit free, which makes the problem a lot harder.

Harbinger.8637:

I have a question though, have the development team ever thrown out ideas of just removing the water and making slight tweaks to bring it back to pvp?

Yes, we’ve talked about it. We’ve had higher priorities, so it hasn’t been on the task list yet. I don’t know if it will ever come back though, I believe most players would rather see us spend designer and environmental artist time working on new game modes.

tichorum.2415:

sPvP wouldn’t exist if they listened to the majority of players. It’d be GvG and PvE.

Probably not even GvG. :\

Luckily, that isn’t the case for PvP and WvW.

Also, if GvG is your thing… https://forum-en.gw2archive.eu/forum/game/gw2/GvG-CDI/first#post4575075

Harbinger.8637:

I preferred Raid on the Capricorn and it was removed!

It was removed because we decided to exclude water combat from PvP, not because of anything wrong with the map’s mechanics (outside the water, obviously.)

I loved the map… very few people knew how to work the canon, so they’d leave me alone as I repeatedly blasted them.

Shanks.2907:

I suppose the fact that you ignored the suggestion of dueling in HotM to make waiting for queue less boring should be taken as a “no”. xD

Yes, ‘no’ in that it won’t be available Dec 2nd. No, ‘yes’ in that it’s something we might consider adding in the future.

This answer it for you?

Lopez.7369:

If you guys look at the metrics and see Skyhammer is clearly avoided once people can vote on it, will you consider doing something about it?

We’ve already done multiple things to the map based on community feedback, including making changes to the map, removing it from the tournament configuration rules, and from team arena.

If you’re asking if we’ll remove the map entirely? Probably not, not so long as some people prefer it.

cheese.4739:

It’s that we believe Skyhammer is not fit to be used for ranked PvP. Not while it allows players to get kills while dealing 0 (or very little) damage with a bunkery build.

There will be a limited set of maps available during the season. All maps will only be available during off-season play.

John mentions that in his post, but doesn’t specify the maps.
https://www.guildwars2.com/en/news/announcing-improvements-to-pvp-and-the-ladder-test-season/

To fill in the missing info, the limited maps = tournament configuration maps… meaning not Skyhammer or Spirit Watch.

TwanBr.8340:

Are you planning on making HotM a more exciting place to be at to compensate?

I can’t speak about priorities, future plans, or promise changes, but it’s on our radar.

TwanBr.8340:

Aren’t you also afraid that you are limiting people of a certain skill level overly much? Or people who play primarily at less populated times of the day?

Yes, I am worried about making people wait too long, especially outliers. We have several settings we can change to reduce wait times at the cost of match quality. It may take some time to find a sweet spot.

Checkout http://wiki.guildwars2.com/wiki/PvP_Matchmaking_Algorithm for more details. Though the current configuration values on that page are from our dev environment, we haven’t decided on final numbers for Dec 2nd.

TwanBr.8340:

I don’t see how a map enforcement will help battle 4vs5.

This was one source of trouble with the old servers, we’re determined to not let it be one with the new servers. We might be a bit over-zealous right now. If so, it wouldn’t be wrong to relax restrictions in the future.

style.6173:

Why does ANET continue to push skyhammer when the majority of players do not like it?

Because if we only cared about what the majority of players like there wouldn’t even be a PvP mode?

We know some players like Skyhammer and some don’t. After Dec 2nd will know the exact numbers, but don’t expect to be surprised. We try to be inclusive as possible.

zELEVENz.6937:

1/6 + (5/6 * 1/5) + (5/6 * 4/5 *1/4) = 1/2

Ugg, you’re right… what was I thinking.

facepalm

But there you go… 50% chance of map showing, with 50% chance of being selected… assuming no one on other team votes for the map, and the troll team all votes for it.

Dondagora.9645:

If that were so, I wouldn’t be troubled, however from the wording and from Justin’s response, it sounds like ANet is putting “duplicate professions” in a very all-encompassing box and not differentiating by profession or build.

As much as I would love if my servers had constant and perfect understanding of the meta, it just isn’t possible. I’d have to put in a lot work and of hacky code to make these sort of exceptions, just to have them possibly change with each game build.

Also, a lot of people are just as annoyed by full mesmer teams because of all the clones. So making exceptions like this will always displease someone.

Dondagora.9645:

Why/How does this new match-making punish a single-profession team?

We reduce matchmaking score for duplicate professions, whether intentional or not. Do you think there are many others like your team?

Aaron.2413:

Great! This is awesome!!! My team (we win every team q) will play and some mid tier team is gonna run 5 engis … They know they can’t win in any other circumstance…. They build only for Skyhammer … Gj anet. This is what we meant when we said we wanted more competition in high lvl play

As Frenk.5917 mentioned, they won’t even have the ability to vote on Skyhammer every time there is a vote. If there are 6 maps, and 3 slots, the odds of any map showing is about 62% (1/6 + 1/5 + 1/4), so their cheeze comp is already out of luck 38% of the time. Then add to that they’re only going to win the vote 50% of the time. That means they’ve got a 31% of getting skyhammer during any vote.

Oh, and running 5 of any professions is also going to hurt them on matchmaking, so it’ll take them longer to find a match, and that match likely isn’t going to be a good one. We’ll probably toss them in with other cheeze comps.

Do you think they’ll still want to risk it?

Edit: Cleaned up language because I didn’t mean to call out Aaron.2413 on his theoretical post.

manveruppd.7601:

Justin, can’t you just make the time people get to respond to the match ready prompt longer so they’ll be able to click it after their cutscene or load screen finishes? 30" should do it for loading screens. As for cutscenes and cinematics, tbh you should be able to make a pop-up dialog that overrides them…

There are certainly many ways we can work around the issue, but right now our first priority is do everything possible to combat the 4v5 problem.

Marcos.3690:

They can make a premade group that is insanely op at skyhammer and win half of their matches..

Less than 50%, probably closer to 30% if they don’t adapt. You’re given 3 maps to vote on, and which 3 is chosen randomly, so there’s chance any map won’t even be a choice.

MarkPhilips.5169:

Probably stupid question but can i do Hotjoin or private custom arena or i have only to stay in the hotm running around?

Not stupid, but no, you’ll need to stay in the HotM.

The alternative is to let people be anywhere, and risk them having a queue pop they can’t do anything about because they’re loading into another map. Which means not only would they no longer be queued (with possibly no notification), but they would also get a timeout and dishonor for queue dodging.

The new UI is also very obtrusive on purpose, so can be very annoying if your fighting something and it pops up.

If wait times are a significant issue, making waiting in HotM very boring, we can try to find alternative solutions. So be sure to share that opinion if you have it after Dec 2nd.

lvis.3824:

So you say, ( and Anet ) , the chance that someone will go AFK while in the queue is higher if he is actually playing the game in another part than standing in the mists while doing nothing ????
Interessting…

No, staying in the HotM is to prevent players from getting dishonor for failing to respond to a match confirmation because they’re loading another map. It also takes care of people who forgot they queued and went on to do something else in-game because it will remove you from the queue as soon as you leave.

KkIenJuju.9406:

If i understood this right: you que up with a spesific class( eg. Warrior) but then you see that your team needs a bunker guardian to win. So you relog to your guardian, but you get minus points for leaving the match, even if this desition may have won you the game?

Correct.

Lopez.7369:

Justin ODell.9517:

That said, I think character swapping mid game should be discouraged. Does anyone have an opinion on that?

Every single high-rated, competitive player knows this should be the case. Not thinks. Knows.

That’s what I figure, but I was opening it up to any options to the contrary.

Teutos.8620:

reset=“2014-02-01”

1.) I thought the rating is not going to get reset?

Sorry, I’ll put a notice on the wiki. The current configuration is straight from our development environment, and does not contain the values that will be set on Dec 2nd. Also, if we kept this value, your MMR would not reset if you played after Feb 1st 2014.

Teutos.8620:

Profession max=“2”

2.) Maximum of two players of the same profession per team?

Profession/@unique is scored for every profession under 2 (meaning only 1) the potential roster would add to the team. Profession/@common, which is negative, scores for every profession at or above 2 (meaning > 1) the potential roster would add to the team. This includes professions already selected in other rosters for the team.

I should clean up the code, config, and the wiki to make this more clear. When I wrote the XML parsing code, I wasn’t yet sure how exactly I wanted it to operate. This won’t happen until after Dec 2nd though.

Teutos.8620:

<Dishonor distance=“-100” stack=“-50”/>

3.) How should we read this pseudo-code? Is the dishonor distance the distance you may still play together? Can a player without any dishonor stacks play with a player with 104 stacks?

These numbers are exaggerated, because they are test values. Yes, someone with a lot of dishonor can play with someone who has no dishonor. The scoring will just make that less likely. If you are are in a party together, you will definitely be able to play together, assuming there is no active timeout.

manveruppd.7601:

I’m with the people who posted above on this. There is no reason to disallow profession swapping. It’s a perfectly valid tactical choice to punish an enemy team which has invested too heavily in an easily-countered profession. It seems that you’ve only chosen to disable it so you could enable your MMR-by-profession feature. Well, MMR-by-profession a neat idea, but I don’t think it’s necessary. The other improvements you’ve made to matchmaking should be enough to make matchmaking better. I confess, as someone who gets slaughtered every time I try a new build (because I get matched against people who are as skilled as I’ve gotten on my main), it would be nice to be able to slum it at a lower rank when playing something new, but since you’re introducing an unranked arena, that should be enough for people like me who have only started playing a warrior last week. I don’t want this enough to want to disable profession-swapping, and I doubt many other people do either – I don’t remember anyone raising this issue on the forums, so it’s not like this is a feature that many people think is necessary.

You may be right, but we’ll wait until we have actual data to look at. If the cost of having the feature out-weight the benefits we’d definitely be willing to change it. The new systems make this sort of thing much easier to do.

That said, I think character swapping mid game should be discouraged. Does anyone have an opinion on that?

Deimos Tel Arin.7391:

what if the new system only slap offenders with dishonor, only if they came back with a different profession?

While I can’t any we’ll make any changes, I can say the new systems give us way more flexibility to try more things. So feel free to bring up discussions about these issues throughout the season.

MarkPhilips.5169:

I hope it’s not in this way because it’s very very heavy.

You queued as a warrior, we found a match for a warrior, using your warrior MMR. It wouldn’t be fair to others if we let bypass matchmaking rules, and it opens the door for exploits.

If the full loss for small disconnects becomes an issue, feel free to bring it up again and we can talk about less severe deterrents. As John mentioned, this is an experiment and we’ll be looking at the data and community reaction.

XGhoul.7426:

That’s an insane requirement given that some teams (teamq) run less than optimal comps for the sake of fun, but when they see a real team they swap classes. What you’re essentially putting is a barrier to always play serious which results in teams steamrolling other players and rather than building a stronger pvp community you get more people discouraged from playing.

I get you, the matchmaking system will do much more to prevent awkward compositions.

If you want full control over comp you must form a party. Evan’s post tomorrow (today for you, maybe?) will explain how we’ve simplified and improved the process. I hope it helps.

glock.6590:

Yeaa same here…. I play on a very crappy internet and PC, hopefully arenanet can tell if its DC or rage quit. Like giving 5 min to come back would be reasonable, that’s what it takes to reset the PC and stuff.

Sorry, but if you leave for any reason you’ll get dishonor, no matter the length of time. Even the shortest interrupts can have significant impact on the quality of the game for the other players.

Unfortunately there is no real way that I’m aware of to detect real disconnects from someone ripping out their Ethernet cable and setting it on fire in a fit or rage induced passion.

As usual though, if there is good reason too, we’ll take the time to look at this again. I can’t promise any we’ll take any action though.

Deimos Tel Arin.7391:

oooo. eh, is the beta ladder based on ratings?
as in, players with higher rating will be near the top etc.

Not based on ratings at all, other than compensating for uneven matches by increasing risk and reward in ladder points.

Deimos Tel Arin.7391:

so if someone stops actively gaining rating, they will eventually be pushed downwards by other players climbing up the ladder with their increasing rating due to more active winnings.

It’s possible to keep climbing higher and higher the more you play. How fast depends on your win/loss ratio and who you play against.

If you have a high win/loss ratio but constantly play against lesser skilled opponents you may find those with lower win/loss ratios passing you because they’ve been successful in matches that were more challenging.

The biggest thing driving the upward trend is the ability to not lose points, or even gain some, when you lose a close match.

I know how it works in simulations, and I’m really excited to see how it works in the real world.

No, spectator mode will not be available. I would like to see this though.

Deimos Tel Arin.7391:

MarkPhilips.5169:

Justin, can you answer pls ?

And what about the leaderboard decay?

Any change on that?

we temporarily lose ladder points when we don’t play and we’ll re-get the same amount when we play 1 game? Or do you have changed something?

i’m guessing decay will cause position in the leader board / ladder to drop, but the actual ratings will remain unchanged.

does that sound reasonable?

that’s my guess though.

The beta ladder will not include any decay, and may not need any. We want to wait and see before making a decision for future ladders.

Deimos Tel Arin.7391:

do the rewards get better or worse with the -3 / +3 rating change?

for example, in terms of gold rewards,

rating / gold
-3 = 10 silvers (equivalent to hotjoin lose)
-2 = 12.5 silvers (equivalent to solo arena lose)
-1 = 15 silvers (equivalent to team arena lose)
0 = 25 silvers (equivalent to solo arena win)
+1 = 30 silvers (equivalent to team arena win)
+2 = 60 silvers
+3 = 1 gold

how would it be oooo justin?

thanks!

I honestly don’t know. Those rewards are part of content and not handled by my servers. Not a bad idea though, I’ll pass it along.

glock.6590:

So there will only be a " Ranked Q" instead of solo Q and team Q. I was wondering if the wins/losses accumulated on both leaderboards will be either merged , reset or completely deleted for the new system ?

Your record of solo and team arena wins/losses will be kept and migrated over to ranked arena.

Your rank on the leaderboard will be reset.

Teutos.8620:

Is it also an individual Glicko2-Rating Algorithm per profession?

It’s calculated exactly the same.

Teutos.8620:

Do I have the same MMR in ‘unranked’ as I have in ‘ranked’? F.e. if I have a high MMR on my mesmer in ranked, will I automatically have a high MMR in unranked? Are both values exact the same?

No, profession ratings, like the core rating, are per rating type. Unranked and Ranked each have their own rating type exclusive to them.

Iason Evan.3806:

I still have a concern about multiple same professions on an account having different MMR. If all I have to do to game the system on a Warrior is just roll a different Warrior, then that is an issue for sure.

No, it’s profession specific, but not character specific. Your warrior rating will be the same no matter what warrior you’re on.

Saiyan.1704:

Just to clarify, that counts for each individual person instead of that team as a whole? For example, GuardianBob’s odds of victory could be 0-19% while his partner MesmerJoe could be 20-39%. Or are both player’s odds of victory the same?

The odds are for the entire team. They also compliment the enemy’s odds, i.e. your team’s odd + the enemy team’s odds = 100%.

Nebilim.5127:

Isn’t possible to give players a choice? Some box where they can tick to either choose a fast match, or a quality match that will take however possible.

There are several ways we might be able to do something like that, but there is no way they will be available with this build. I also don’t know if it be would be wise to split the queue like that. I’ll look into it, but can’t promise anything.

Iason Evan.3806:

What I mean is, I enter the match on my garbage Thief. We are waiting for the match to begin. I then swap to my really good Ranger. How is that handled?

You’ll receive dishonor if you leave a match and take a full loss on the leaderboard.

Meglobob.8620:

One question I have will the wait time for a game be increased by a lot? You can wait 10-20 mins+ now for a team/solo q game on occasions….

It should actually be faster, but we’ll need to wait and see how it performs in the real world (and not just our internal testing and simulations.) We have several ways we can reduce queue time by sacrificing match quality, or vice versa.

skullmount.1758:

This has only to do with solo and team arenas, correct? Hotjoin, at least manually choosing a server/match, isn’t going to be affected by these changes?

Correct. Custom arenas and play now are unaffected.

Iason Evan.3806:

What’s to keep people from queue-ing as a profession they don’t regularly play so that they enter the match at a lower rank and then swapping to a profession they do play a lot?

Leaving the PvP lobby will remove you from the queue.

Iason Evan.3806:

Does the system take into account your MMR as a bigger percentage and then your rank?

Yes. Once I’m done with the wiki page you’ll see exact numbers once they’re ready. We’ll be changing them often before the ladder starts, but they should be stable after that point.

Iason Evan.3806:

I am also wondering if your different professions will have different ranks. Is that the case?

No. Rank is still account wide.

Iason Evan.3806:

I would love to see that be how it is handled and I would also love to see a leaderboard with “Best overall” that showed players who were really good on numerous professions. It would be a great way to differentiate players and have them be identified as really good on a profession or two or really well-rounded because they are solid on a lot of different professions.

The new PvP backend servers allow for this sort of flexability, but it isn’t currently planned.

Iason Evan.3806:

Thanks for any info you can provide in regards to my questions! Super excited that this could end up being where I spend a ton of time again!

Glad to hear! I’ve been dying waiting to talk about these changes.

MarkPhilips.5169:

Basically, your ladder positions would be the sum of every points (-3 – +3) you can gain in a match?

Correct.

MarkPhilips.5169:

Team MMR would be the simply average about every single MMR player or other formula?

Yes, simple average. We compensate for the advantage of party size through the matchmaking and match prediction algorithms.

MarkPhilips.5169:

And another thing…in the new leaderboard, can you show us our positions AND our total ladder points?

Yes, that’s the plan.

Elegant Avenger.8042:

Does this mean that if you end up facing players from the top of the leaderboards, you have a lot more to gain than to lose? Or vice versa; if you’re at the top, you can’t gain much but you can lose a lot?

Yes. You can potentially gain more points by losing than they can by winning.

Elegant Avenger.8042:

Additionally, are you going to show ladder points on the leaderboard, or before/after games on a pop up?

On the leaderboard, yes. In game, no, not with this build, We did not want to devote resources into new UI work until we get the community’s reactions to this leaderboard.

Aerveor.9617:

My concerns are with ladder rank and the ranked area. Based on what I’ve read, I see the potential to abuse this system by initially climbing the ladder on your main profession. Then, when competition gets stiff, swap to a profession and build that you don’t have many games on that traditionally does well at lower ratings.

We’ll need to wait and see. If it becomes an issue we can definitely change the rules.

Aerveor.9617:

For example, what’s the incentive, besides integrity, for a someone who never plays engineer to climb to the top 300 of the new leaders boards with their main profession…and then play turret engi, have their MMR adjusted, and boost their ladder points with the wins at this MMR?

If they don’t play a profession often then they should have a high RD, causing their rating to adapt quickly. We’ll keep an eye on it to see if its fast enough to prevent this sort of thing.

Dead Muppet.9718:

I call shenanigans, this is setup for a 50/50 win loss ratio. Meaning some games the odds will be stacked against you mmr rating wise, and other games they will be stacked in your favor.

I’m not sure what you mean, but yes… the algorithm is not perfect and even in the best-cast scenario there is almost always a slight advantage on one side. Glicko2 factors in the enemy rating though when making adjustments though.

Echiel.8196:

Would it be possible for for example for ranks 96%-0% to implement another rule to count a team lvl.

I’d like to see something added that, at the very least, increased the odds new players are grouped together. People have suggested a number of ways this could be done. Does anyone remember the big thread for that?

SYPHA.9283:

Anytime a comment was made or evidence was shown of the opposite the Dev stated, he had a generic recant ready.

I disagree with this. We’re dealing with anecdotal and disjointed evidence here, so I tried to explain what could be going on but it involves a bit of guesswork. As a general rule, we should take any comparisons involving screenshots and leaderboard data with a grain of salt because they simply don’t tell the whole story.

SYPHA.9283:

By what the Dev said of the MMr and the same ppl of the same skill level are matched up….do u mean i will only ever play the same people over and over and over, because god knows the randomness of pvp, W and L, and Anet only seems to ALLOW a select few to be in top 25 consistently. Or make it above 50% ish.

I’m confused, you seem to be contradicting yourself… are you experiencing random matches, or do you feel matchmaking is preventing you from playing with players outside your skill level? (Is this even something you’d want?)

You are more likely to play with people around your skill level, but it also depends on who is in the queue at the same time you are and how long you’ve been waiting.

SYPHA.9283:

MMR to be accurate needs to take into account the individuals contributions to the team as well. Not just as a whole of Win, Loss.

This would be awesome, but would be very hard to get right and would deserve a very large discussion of its own.

Zapion.2967:

As far as I can see your MMR can sink if ppl roflstomp you but if you have a 4on4 group fight your own actions have a weight of 25% of the result. And if your mate is a brainless daily farmer who gives a kitten about teamplay this should effect you? Are you serious?

I think this goes along with what SYPHA.9283 was saying. Does anyone see any low hanging fruit here? It would probably need to be something that would be generic enough to be true for any map and/or gamemode.

Edit: We should totally start a new thread for this.

Please include the IP address of the server if possible. We believe this may be a hardware issue.

Is this a crash or just an error popup?

Should be fixed now.

skullfaerie.7203:

Out of curiosity, what about those random matches that pop seconds after hitting the queue button

If it isn’t you that had to wait then it was someone else.

Seiishizo.7162:

how would you explain this?
http://imgur.com/jjHRu8W

Worst case scenerio, assuming this is all solo players queued for team arena, this likely because team arena does not shuffle teams to make them even and you were put into best fitting team available at the time the team was formed.

Best case, this spread is caused more by the self selection of players forming their own rosters to play together.

The truth is probably somewhere in between.

However, there is a known (to us) flaw with team arena matchmaking. It will never try to reform a team (to try to find a better fit) once its been created. So you are stuck with the best team it could do at the time, and have to wait until it finds another team within range comes along. This is also why team arena queues can sometimes take longer, or why even a bad match can pop quickly.

Seiishizo.7162:

But if it’s based on MMR i would assume that the MMR of a #3 player would be way way higher than a 97% player.

That is an incorrect assumption. The MMR spread between a rank 3 and 97% can actually be very small. Your leaderboard rank (for the top players) is your position in a sorted list, and does not represent the distance between players other than to say one is lower than the other.