|
juicyjames
United States3815 Posts
http://us.battle.net//sc2/en/blog/12055065
http://eu.battle.net//sc2/en/blog/12055065
![[image loading]](http://i.imgur.com/wlrq8wq.jpg)
Since we introduced a new ladder with the launch of Heart of the Swarm, we’ve been continuously reviewing your feedback about league placement and the matchmaker in StarCraft II. In particular, recently we’ve seen some discussion in the community about the current state of the ladder. We wanted to take this opportunity to address those discussions, clarify some misconceptions, and mention some changes we plan to make.
Right before we launched Heart of the Swarm, we announced that we were changing our targets for player distribution across leagues. For various reasons since that announcement, the distribution of players across leagues slowly shifted to no longer match those desired targets. Currently the lower leagues like bronze and silver have a larger percentage of players than desired. Meanwhile the upper leagues like platinum and diamond are under-represented. One of the reasons this occurred is due to the way that we maintain those target percentages.
Accordingly, around the time we release the 2.1 patch, we are changing how league boundaries are determined so that we can more closely match the desired percentage of players in each league. Thereafter, we expect to see a shift in the distribution of players across the leagues to their correct placement within the ladder. Because of the current bottom heavy distribution, many players may see a league promotion as a result of this change.
Recently there has also been some community discussion over what some players are calling “MMR decay”. This is the slight adjustment to a player’s hidden rating that occurs after that player has been inactive for a period of time. We implemented this automatic adjustment with Heart of the Swarm, after we found that players who returned to the ladder after a period of inactivity were frequently suffering a string of consecutive losses. That experience can be demotivating and discouraging.
Since there is some confusion in the community, we want to make sure that the way this adjustment works is fully understood. Specifically:- Each ladder queue is adjusted separately. For example, an absence from games in team queues will not adjust your 1v1 matchmaking rating.
- The adjustment only kicks in after a player has played no games (ranked or unranked) in a given queue for more than two weeks.
- Once it kicks in, the adjustment ramps from zero adjustment to our maximum adjustment over a period of two weeks. There is no further adjustment after four weeks of inactivity.
- At its maximum value, the adjustment is small; it’s the equivalent of losing a few games.
- If a player plays even a single game on the ladder every two weeks, their hidden rating will never be adjusted downward.
Prior to implementing this feature, data gathered from games played on Battle.net showed that even an absence of only 7 to 14 days results in a drop of a player’s win percentage. The drop becomes more pronounced as the length of time away increases. After we introduced this change, the data showed that we see a win percentage for returning players that’s right around 50%. This is especially important for the first few games played after a player returns. Additionally, to address the misconception that the majority of players are being adjusted in this way, the data shows that less than 6% of all StarCraft II games played on Battle.net are affected by this adjustment.
While we have evidence that the adjustment is helping players, we’ve also been reading your feedback and analyzing the way the system is working. As a result we are currently looking at the following areas:- We’re looking into possible changes to the way adjustments are made in team games.
- We’re considering altering the time period before the adjustment kicks in.
- We’re also considering altering the amount of the adjustment.
We continue to investigate ways to improve the experience of climbing the ladder. We want players to know when they’re moving up, and for those who want more granular details, better indications of how they’re doing competitively. We’re exploring ways to let you know when you are getting close to a promotion.
Finally, our balance team continues to monitor play results at all levels as they make their fine-tuning decisions. We've received a lot of good feedback from the community on that subject, and look forward to evaluating that aspect of the game separately after the release of patch 2.1.
We greatly appreciate your feedback and encourage you to share your thoughts in the comments below.
|
I think the leagues are pretty retarded to be honest, I keep facing low gold players whom i beat really easily or diamond/masters players who beat me really easily and im still gold. Sort this shit out Blizzard , either give me plat or players my own skill level
|
I'm interested in how the 2.1 changes will effect player motivation. I know I was heavily demotivated upon being demoted to diamond, and I've been playing for years. I imagine a lower league player would be even more so demotivated when being demoted from diamond to plat, or something likewise. Seems logical for people to be motivated if a lot of promotions start happening again.
|
Australia528 Posts
I think the 6% figure is if they only count people who are active. A lot of people play very few games per season and have kept their leagues that way for a long time . That doesn't mean those people have gotten so bad they need to be placed two or three leagues below their original league.
|
Good to see Blizzard is noticing that the MMR decay and league distribution is a problem right now.
I hope that the distribution change won't make it like in WoL, where being Masters meant nothing.
I'd also love to see a less aggressive hidden rating decay.
|
Serbia681 Posts
Hmm it took hem a while...
Hopefully they will really fix it.
|
I got dropped in league hard after a long absence and started playing again only recently, just a couple of games every few days. Absolutely crushing most people whilst barely even trying, not even using proper builds. I still think they made the decay too hardcore, you don't drop THAT much even after a couple months away. :-\
I mean hell I think I've got about a 85-90%+ overall winrate at the moment, and the ones I lost were because I was playing drunk. I have literally double or triple the APM of anyone I'm playing; its almost sad to just destroy them.
|
At its maximum value, the adjustment is small; it’s the equivalent of losing a few games.
Tell that to all the players who dropped down 2 or 3 leagues below in a couple of month. Going from master to gold in 2 seasons because you did not play isnt the "equivalent of losing a few games" for fuck's sake. Really not convinced by those explanations, the ladder system needs a more drastic revamping, imho.
|
On December 19 2013 01:18 fastr wrote:Show nested quote +At its maximum value, the adjustment is small; it’s the equivalent of losing a few games. Tell that to all the players who dropped down 2 or 3 leagues below in a couple of month. Going from master to gold in 2 seasons because you did not play isnt the "equivalent of losing a few games" for fuck's sake. Really not convinced by those explanations, the ladder system needs a more drastic revamping, imho.
What people aren't dropping 2 to 3 ranks. I didn't play for 3 months and I placed masters still on my first game back.
|
On December 19 2013 01:26 blade55555 wrote:Show nested quote +On December 19 2013 01:18 fastr wrote:At its maximum value, the adjustment is small; it’s the equivalent of losing a few games. Tell that to all the players who dropped down 2 or 3 leagues below in a couple of month. Going from master to gold in 2 seasons because you did not play isnt the "equivalent of losing a few games" for fuck's sake. Really not convinced by those explanations, the ladder system needs a more drastic revamping, imho. What people aren't dropping 2 to 3 ranks. I didn't play for 3 months and I placed masters still on my first game back.
Some have reported drops of Diamond or Masters to Silver level so...yeah.
Broken AND inconsistent it seems. >_>
|
United States248 Posts
On December 19 2013 01:26 blade55555 wrote:Show nested quote +On December 19 2013 01:18 fastr wrote:At its maximum value, the adjustment is small; it’s the equivalent of losing a few games. Tell that to all the players who dropped down 2 or 3 leagues below in a couple of month. Going from master to gold in 2 seasons because you did not play isnt the "equivalent of losing a few games" for fuck's sake. Really not convinced by those explanations, the ladder system needs a more drastic revamping, imho. What people aren't dropping 2 to 3 ranks. I didn't play for 3 months and I placed masters still on my first game back.
This change isn't affecting people in high masters +, its just low masters/diamond and below.
|
On December 19 2013 01:26 blade55555 wrote:Show nested quote +On December 19 2013 01:18 fastr wrote:At its maximum value, the adjustment is small; it’s the equivalent of losing a few games. Tell that to all the players who dropped down 2 or 3 leagues below in a couple of month. Going from master to gold in 2 seasons because you did not play isnt the "equivalent of losing a few games" for fuck's sake. Really not convinced by those explanations, the ladder system needs a more drastic revamping, imho. What people aren't dropping 2 to 3 ranks. I didn't play for 3 months and I placed masters still on my first game back.
I dropped from platinum to bronze, and stomped my way back into platinum.
|
when is 2.1 supposed to hit?
|
On December 19 2013 01:48 MrMatt wrote:
when is 2.1 supposed to hit?
Question for the ages 
the PTR has been out for a while, so it should be too much longer.
P.S. I am sooooo happy about this!!
The number of ex masters in gold is incredible.
|
On December 19 2013 00:20 Hadronsbecrazy wrote:
I think the leagues are pretty retarded to be honest, I keep facing low gold players whom i beat really easily or diamond/masters players who beat me really easily and im still gold. Sort this shit out Blizzard , either give me plat or players my own skill level
It doesn't matter which league you are in, you will face that either way so it doesn't really. Platinum has that just as much inconsistency when matching as gold does.
It is quite frustrating. The decay is too steep. Once you hit that two weeks you are guaranteed quite a few games against much worse people before you can play your level again.
|
United States12240 Posts
My notes:
- Decay must have really messed up their activity metric because it's separately tracking a player's real time gaps between games as well as their unspent bonus pool. If they're using apples to redistribute the leagues from season to season, and the apples are still pretty firmly where they should be each time, then you can't have oranges artificially pushing people down and thinking the skill definitions will remain intact. They can't.
- I'll put it another way. Let's say that in Season 6, 28% of the ladder population with bonus pools below 208 have less than 800 MMR. Perfect. Then for Season 7, we'll make 800 MMR the Silver/Gold league boundary. Uh oh, it turns out 80% of players under 1100 MMR have more than a one-month gap between games. Now in terms of Activity A we're fine, 28%, cool, but in terms of Activity B, now 800 MMR contains 60% of the players. They measure completely different things, and that's a problem.
- They mentioned how they're going to change the distribution. Inevitably that's going to have to factor in decayed players. How exactly they're going to do that I don't know. Maybe they could do a breakdown of what percentage of players per league experience decay and introduce that as a coefficient to their existing apportionment metric. It wouldn't be perfect because again, apples and oranges, but the boundaries and internal league percentages are hidden anyway, and there's always some slush involved besides.
- As I guessed, each bracket tracks MMR decay separately. That's actually a big problem because it's much harder to get the same arranged 2v2, 3v3, and 4v4 players together than it is to play 1v1 or Random Team. I know a bunch of players who have like 50 4v4 teams, many of them only went through 5 placement matches and that's it. When the same 4 players play again next season, they've decayed to a lower league. They haven't forgotten how to play because they're playing with 49 other team combinations, just not this specific group. They're not rusty, but they're penalized as though they're rusty. Team games probably shouldn't have decay at all: there are fewer team listings than solo players and it's too easy to mess up the distribution.
- Our data shows that the maximum impact of decay is over 310 rating. Though it does cap around this value, that's still the equivalent of losing 20 straight same-skill games. That's more than "a few" in my glossary.
- They said that decay starts at 14 days where it starts at 0 adjustment. That's something I've said as well and it's important for players to understand that. If you take a 15 day break, you're only a game and a half behind (in sports terms). Same with "if you play one game every two weeks, you won't decay" -- very important for players to remember that because people were getting paranoid.
- Their "to address the misconception that the majority of players are being adjusted in this way, the data shows that less than 6% of all StarCraft II games played on Battle.net are affected by this adjustment" is really misleading I think. It's another apples-to-oranges comparison. What we're asking about is how many players are affected, but what they reported is how many games are affected. In terms of activity, high level players play much more often than low level ones. I don't have the data here, but it wouldn't surprise me if only 10% of players played 94% of all the games on the ladder. That would still mean 6% of games were affected, but it would also mean that 90% of players decayed to some degree. Now, I know they have to have the "# of players affected" data, but maybe that number really does constitute a majority of players after all.
|
Is there a reason they can't just use an open ranking ELO system? Or would that be too imitating seeing your ELO jump up and down
|
On December 19 2013 02:54 BakedButters wrote:
Is there a reason they can't just use an open ranking ELO system? Or would that be too imitating seeing your ELO jump up and down
Yes and it is also a shitty way to track your on progress, since it will likely just level out. Match making numbers do not help people see improvement or provide them with useful information, it is mostly just to pair players.
|
At its maximum value, the adjustment is small; it’s the equivalent of losing a few games.
High diamond to low gold = losing a few games. Okay blizzard.
|
this post has so many words in it yet it says absolutely nothing. It's full of "we will look intos" and "we will consider"
|
|
|
|
|
|
EPT
