Out of curiosity, is the random roll going to be set up to roll in advance of a reset, so servers have a general idea on who they will be facing in the next match maybe a day or two ahead of time, or will this roll only happen after the current match ends?
Devon please can you fix the EU matchup – how did Gunnars end up against Piken and Riverside? Jon Peters made clear this sort of thing wasn’t going to happen, and the maths had it at less than 0.3%.
I kind of like the idea of not knowing the random rolls until the match has already started. It prevents ‘gaming’ of the rating system to try to get favorable matchups.
Devon please can you fix the EU matchup – how did Gunnars end up against Piken and Riverside? Jon Peters made clear this sort of thing wasn’t going to happen, and the maths had it at less than 0.3%.
I’m thinking the matchups that took place (albeit short) actually counted towards the servers ratings leading to some servers rating going up or down on drastic levels that should not have happened.
I’m thinking the matchups that took place (albeit short) actually counted towards the servers ratings leading to some servers rating going up or down on drastic levels that should not have happened.
That makes a lot of sense. Because the matchups did change from the original. This also likely answers the question will we know beforehand the answer is no but we will have a good guess.
I’m thinking the matchups that took place (albeit short) actually counted towards the servers ratings leading to some servers rating going up or down on drastic levels that should not have happened.
That makes a lot of sense. Because the matchups did change from the original. This also likely answers the question will we know beforehand the answer is no but we will have a good guess.
I guess we find out if this holds some weight when NA resets. If our matchups look a lot better than theirs, I think we can conclude this is likely what happened.
The NA matchups won’t be available until the reset.
In the future will we have a bit more of a heads up, especially as to what color the server will be? I know on TC the guild leaders like to get together and decide what guilds are going on what maps so that not all the guilds go on the same map, and the color of the map decides if they’re bay side, hills side, etc.
TC Golden Dolyak – [DOLY]
Mesmer – FURY
Rank 55 – Bunker Engi, Top 300
Sounds like the devs are waiting for a final match score to determine the ratings for next week. Might be nice to base the matchmaking ratings based on the scores at, for example, midnight Thursday PST while keeping Friday relevant by including it in a final rating. Might sound a little confusing, but trust me, there are enough clever people on every server to figure out even the most insanely complicated scoring scheme. And the tiny amount the rating might change on Friday should be drowned out by the RNG roll.
Just an option for those wanting to know the match-ups in advance.
Altoholic Luciana Delaluna, Ranger
Phoenix Ascendant [ASH] – Gate of Madness
“This space intentionally left blank.” ~ Zork
Sounds like the devs are waiting for a final match score to determine the ratings for next week. Might be nice to base the matchmaking ratings based on the scores at, for example, midnight Thursday PST while keeping Friday relevant by including it in a final rating. Might sound a little confusing, but trust me, there are enough clever people on every server to figure out even the most insanely complicated scoring scheme. And the tiny amount the rating might change on Friday should be drowned out by the RNG roll.
Just an option for those wanting to know the match-ups in advance.
If the random portion of the calculation for each server was released (the -1.0 to 1.0 part) on, say, Thursday night it would allow people to calculate their new match rating and plan accordingly.
At the same time that should be short enough notice to prevent anybody from trying to game the system.
“I’m not a PvE, WvW, or PvP player – I am a Guild Wars 2 player”
Tarnished Coast – Dissentient [DIS]
All classes
There’s no way to know the actual matchups in advance unless you know the random numbers that are going to be used. Personally I think this is a good thing.
But you can do some probability analysis to get an idea of what matchups are likely and what matchups aren’t. I had intended to do this last night, but it turned out that I needed some information I didn’t have, so I’ve been writing software to fill in the gaps. Hopefully I’ll be able to post some probability data next Thursday, assuming I finish writing and debugging code in time.
Sounds like the devs are waiting for a final match score to determine the ratings for next week. Might be nice to base the matchmaking ratings based on the scores at, for example, midnight Thursday PST while keeping Friday relevant by including it in a final rating. Might sound a little confusing, but trust me, there are enough clever people on every server to figure out even the most insanely complicated scoring scheme. And the tiny amount the rating might change on Friday should be drowned out by the RNG roll.
Just an option for those wanting to know the match-ups in advance.
If the random portion of the calculation for each server was released (the -1.0 to 1.0 part) on, say, Thursday night it would allow people to calculate their new match rating and plan accordingly.
At the same time that should be short enough notice to prevent anybody from trying to game the system.
the t7 is gonna rage quit and WvW lose more ppls VERY GOOD IDEA!
(leave the game not just WvW)
Ah but this would be a mistake. A server at a vast rating disadvantage (to use your example, a 10 Seed versus a 20 Seed, it makes little sense to use tiers now that there is match-to-match mobility) can come out ahead in rating even if it scores less than 0.5 of the points against the higher rated server. That is, there is also pressure on the 10 Seeded server to dominate the match; if the 20 Seeded server is able to put up enough of a struggle (gaining valuable experience against a better team, as well) it might be able to move up a spot—19th Seed. This is the motor behind rank mobility with the newer system.
There’s no way to know the actual matchups in advance unless you know the random numbers that are going to be used. Personally I think this is a good thing.
But you can do some probability analysis to get an idea of what matchups are likely and what matchups aren’t. I had intended to do this last night, but it turned out that I needed some information I didn’t have, so I’ve been writing software to fill in the gaps. Hopefully I’ll be able to post some probability data next Thursday, assuming I finish writing and debugging code in time.
-ken
Yeah, I ran some simulations based on last week’s leader board and guessing some values to use in some of the sketchy details of their methods. I came up with only a 40% chance that we would face either of our current opponents next week. Can’t tell you much about Ehmry Bay except that I came up with only about a 5% chance that you would meet Gate of Madness next week. ATM I am trying to sort out if the EU match-up is a reasonable outcome based on my analysis.
Altoholic Luciana Delaluna, Ranger
Phoenix Ascendant [ASH] – Gate of Madness
“This space intentionally left blank.” ~ Zork
the t7 is gonna rage quit and WvW lose more ppls VERY GOOD IDEA!
(leave the game not just WvW)
Ah but this would be a mistake. A server at a vast rating disadvantage (to use your example, a 10 Seed versus a 20 Seed, it makes little sense to use tiers now that there is match-to-match mobility) can come out ahead in rating even if it scores less than 0.5 of the points against the higher rated server. That is, there is also pressure on the 10 Seeded server to dominate the match; if the 20 Seeded server is able to put up enough of a struggle (gaining valuable experience against a better team, as well) it might be able to move up a spot—19th Seed. This is the motor behind rank mobility with the newer system.
This guy gets it.
Lower seeded servers could still gain rating even while getting blown out. The math would then re-adjust the following week with higher probability of finding accurate matches in the newly-adjusted rating spread.
Essentially, this should fix glicko’s tendency to only absorb rating points from within your own tier and lead to much more volatility in the positions.
@DevonCarver thank you for your hard work and communication. You can’t please everyone, but a majority of the community is excited to fight some new opponents.
the t7 is gonna rage quit and WvW lose more ppls VERY GOOD IDEA!
(leave the game not just WvW)
Ah but this would be a mistake. A server at a vast rating disadvantage (to use your example, a 10 Seed versus a 20 Seed, it makes little sense to use tiers now that there is match-to-match mobility) can come out ahead in rating even if it scores less than 0.5 of the points against the higher rated server. That is, there is also pressure on the 10 Seeded server to dominate the match; if the 20 Seeded server is able to put up enough of a struggle (gaining valuable experience against a better team, as well) it might be able to move up a spot—19th Seed. This is the motor behind rank mobility with the newer system.
This guy gets it.
Lower seeded servers could still gain rating even while getting blown out. The math would then re-adjust the following week with higher probability of finding accurate matches in the newly-adjusted rating spread.
Essentially, this should fix glicko’s tendency to only absorb rating points from within your own tier and lead to much more volatility in the positions.
Oh you’re nice.
To provide an example outside of our complicated system (and then apply it to that system), consider an FIDE rated chess game between an 1800 rated and 2200 rated player. If they draw (1/2 – 1/2), the 1800 rated player gains 6.4 points!
In GW2, results are between 0 and 1 depending on the ratio of your points and the total of your and your opponent’s points. The higher the rating differential, the less you have to score (perhaps even as low as 0.25—a 3-1 defeat) to get a positive rating change.
the t7 is gonna rage quit and WvW lose more ppls VERY GOOD IDEA!
(leave the game not just WvW)
Ah but this would be a mistake. A server at a vast rating disadvantage (to use your example, a 10 Seed versus a 20 Seed, it makes little sense to use tiers now that there is match-to-match mobility) can come out ahead in rating even if it scores less than 0.5 of the points against the higher rated server. That is, there is also pressure on the 10 Seeded server to dominate the match; if the 20 Seeded server is able to put up enough of a struggle (gaining valuable experience against a better team, as well) it might be able to move up a spot—19th Seed. This is the motor behind rank mobility with the newer system.
This guy gets it.
Lower seeded servers could still gain rating even while getting blown out. The math would then re-adjust the following week with higher probability of finding accurate matches in the newly-adjusted rating spread.
Essentially, this should fix glicko’s tendency to only absorb rating points from within your own tier and lead to much more volatility in the positions.
In practice, most players quit playing a particular match for the week when it becomes evident they have no hope of winning. Despite the good points of this system, rankings will remain inaccurate as losing servers fail to field players. Fighting your way up is all well and good in concept, but most players have no comprehension of how rankings work right now, and morale will be low when every fight is a roflstomp.
In the long term, this may be a good change. Right now, it’s going to suck for everyone who likes a fair fight.
In practice, the servers full of people who give up in the face of adversity kind of deserve to get lower ratings.
The higher-rated servers are the ones that continue to fight, no matter what. Under the old system, tier stagnation allowed some servers an easy ride once they fought their way into a tier, but under the new system each server will have to continually prove that they deserve their rating, or risk falling in the rankings.
Not knowing ahead of time if fine for 90% of the population. The other 10% will be upset but not really a big deal. A well coordinated server can still plan what they want to do.
the t7 is gonna rage quit and WvW lose more ppls VERY GOOD IDEA!
(leave the game not just WvW)
Ah but this would be a mistake. A server at a vast rating disadvantage (to use your example, a 10 Seed versus a 20 Seed, it makes little sense to use tiers now that there is match-to-match mobility) can come out ahead in rating even if it scores less than 0.5 of the points against the higher rated server. That is, there is also pressure on the 10 Seeded server to dominate the match; if the 20 Seeded server is able to put up enough of a struggle (gaining valuable experience against a better team, as well) it might be able to move up a spot—19th Seed. This is the motor behind rank mobility with the newer system.
This guy gets it.
Lower seeded servers could still gain rating even while getting blown out. The math would then re-adjust the following week with higher probability of finding accurate matches in the newly-adjusted rating spread.
Essentially, this should fix glicko’s tendency to only absorb rating points from within your own tier and lead to much more volatility in the positions.
In practice, most players quit playing a particular match for the week when it becomes evident they have no hope of winning. Despite the good points of this system, rankings will remain inaccurate as losing servers fail to field players. Fighting your way up is all well and good in concept, but most players have no comprehension of how rankings work right now, and morale will be low when every fight is a roflstomp.
In the long term, this may be a good change. Right now, it’s going to suck for everyone who likes a fair fight.
Unfortunately, math can’t fix morale. That’s up to each server to figure out.
But what the system WILL do is reward the servers that keep fighting in a blow out rather than quit. If the lower server fights their butts off all week, they might jump up a few places. If the higher server tries to coast through the week, they might drop in ranking even while winning by 200k.
I think the end result will be more accurate rankings for all of the servers. And once that stabilizes a bit, the deviation variable should keep things fresh since you have increasingly less chance of facing an opponent that’s either way too strong or weak for you. WvW has never been about fair fights, but at least the lines between tiers have been blurred a good bit.
the t7 is gonna rage quit and WvW lose more ppls VERY GOOD IDEA!
(leave the game not just WvW)
Ah but this would be a mistake. A server at a vast rating disadvantage (to use your example, a 10 Seed versus a 20 Seed, it makes little sense to use tiers now that there is match-to-match mobility) can come out ahead in rating even if it scores less than 0.5 of the points against the higher rated server. That is, there is also pressure on the 10 Seeded server to dominate the match; if the 20 Seeded server is able to put up enough of a struggle (gaining valuable experience against a better team, as well) it might be able to move up a spot—19th Seed. This is the motor behind rank mobility with the newer system.
This guy gets it.
Lower seeded servers could still gain rating even while getting blown out. The math would then re-adjust the following week with higher probability of finding accurate matches in the newly-adjusted rating spread.
Essentially, this should fix glicko’s tendency to only absorb rating points from within your own tier and lead to much more volatility in the positions.
You can see this right now on Riverside/Piken Square/Gunnar’s Hold.
Gunnar’s Hold is lowest rated and expected to lose. At the moment, they are … but if the match were to end right now their rating would go up 68 points, mostly at the expense of Piken.
Basically they aren’t losing as badly as their rating suggests they would, and Piken is not winning by as much as they should.
Of course it’s early days yet.
“I’m not a PvE, WvW, or PvP player – I am a Guild Wars 2 player”
Tarnished Coast – Dissentient [DIS]
All classes
Thanks Devon. Variety and uncertainty are good things! Personally, I’d love to get crushed for a few weeks, but a few hours’ delicious anticipation is ample compensation for a week of terror or a week of cruise control. On Maguuma, we’ll enjoy either with our usual, cynical poise. hugs
The NA matchups won’t be available until the reset.
In the future will we have a bit more of a heads up, especially as to what color the server will be? I know on TC the guild leaders like to get together and decide what guilds are going on what maps so that not all the guilds go on the same map, and the color of the map decides if they’re bay side, hills side, etc.
Handin, there’s an easy workaround for you. Ooh, two easy workarounds, actually! Both will assume you’ve got your EB & BL groups planned, as it doesn’t matter what color you are – EB & your BL will always be EB & your BL. As for the other two BLs, there are two options to coordinate without knowing who is what color:
1. Have a team go “left” and a team go “right” from your BL. If you’re in the middle, it’s nice & obvious what’s left & what’s right; just imagine wrapping the edges of the map together, so that if you end up in the green position (at far left), red (middle) is right of you and blue (far right) is left of you.
2. Have a “top” & “bottom” team – your top team goes after the higher (green or blue) borderlands, and your bottom team goes after the lower (blue or red) borderlands. This one’s probably less confusing… but I thought of #1 first, so I decided to put it in anyway.
On another note entirely, I also want to thank the devs for introducing this element of randomness into the matchups. I am perfectly comfortable getting wtfpwned every now & again; losing hard is just a part of the game sometimes. The proper response to a roflstomp is not to get demoralized & quit, but to take stock of the situation and adapt. I’m not saying such adaptation will result in victory, I’m saying it will result in learning (so long as you’re paying attention). Adversity is often the greatest teacher.
I have studied philosophy, jurisprudence, and medicine too;
And worst of all, theology, with keen endeavor through and through;
Yet still I am, for all my lore, the wretched fool I was before…
the randomness didn’t do much for tier 5. EB had finally pushed their rating high enough to get into T4 after weeks and weeks of playing AR and BP.
so what matchup does the new system give us? AR and BP again. except this time we get to be blue instead of green. wheee.
I know it’s random, but it’s still very disappointing. This was basically the only random outcome we didn’t want (EB is as sick of fighting AR and BP as AR and BP are of fighting us), and it was exactly the one we ended up with.
it won’t matter next week… why so upset over one week? … what would have happened had you gotten knocked down a tier again only to have to play weeks and weeks to get back up – at least this way you have a chance at more mobility / variety
Both will assume you’ve got your EB & BL groups planned, as it doesn’t matter what color you are – EB & your BL will always be EB & your BL. As for the other two BLs, there are two options to coordinate without knowing who is what color: