|
On March 03 2015 14:55 ZAiNs wrote:Show nested quote +On March 03 2015 09:01 NonY wrote:During our analysis, we found that [MMR decay] was helping create even matches for players who were coming back from breaks and looking to ease back into StarCraft II. I wish. I didn't play on an account for over 5 months, came back and matched against top 50 GM. I guess the evidence shows that it takes more than 5 months for MMR decay to activate but skill in SC2 certainly drops much quicker than that. And I imagine that many of the people complaining about MMR decay were talking about periods of less than 5 months, and thus crying wolf the whole time. Unless the behind-the-scenes picture is much more complicated (inconsistent) and even worse at accomplishing its goals than we imagined. One change I hope to see for LotV is a different system for top ranks on the ladder. I don't know why the top 200 gets locked out from the rest of the system. The transition from divisions and leagues to a straightforward numerical ranking needs to be smoother. I say make it top 1000 or something, so any player who is decent (as in has any chance of going pro in the near future) can be ranked, and definitely do not let it get locked up. Make a system where people enter and leave based on performance, not activity, all season long. I had a similar experience with MMR decay on a smurf. I've also had a friend who 'decayed' from Master to Bronze, but that's because he got into Master after about 25 games the week HotS came out and then lost a placement game every season for a year. Most people crying about MMR decay probably haven't played enough games.
Well, MMR decay in the first place is by design a feature FOR people who don't play enough games. I don't want to win 10 games in a row to start to meet remotely challenging opponents...
|
On March 03 2015 12:02 NonY wrote:Show nested quote +On March 03 2015 09:23 KrazyTrumpet wrote:On March 03 2015 09:13 NonY wrote:On March 03 2015 09:05 Cyro wrote:I wish. I didn't play on an account for over 5 months, came back and matched against top 50 GM. I guess the evidence shows that it takes more than 5 months for MMR decay to activate but skill in SC2 certainly drops much quicker than that. And I imagine that many of the people complaining about MMR decay were talking about periods of less than 5 months, and thus crying wolf the whole time. Unless the behind-the-scenes picture is much more complicated (inconsistent) and even worse at accomplishing its goals than we imagined. Every time i play the game with a few friends (sometimes after having not played for ~2-9 months) our MMR has decayed to hitting bronze-golds in 2v2. We're pretty much all previous 1v1 master players, and that usually results in winning 5-10 games for free. Yeah.. I've just spent some time reading various accounts of MMR decay. I don't think there's any way so many people are inaccurately reporting. I'm leaning heavily toward the system being complex and faulty and obviously inconsistent. My hypothesis to explain at least part of it is that average MMR shifts and depending on when you quit and when you return, your MMR decay could either be nullified by the whole system shifting along with it or amplified by the whole system moving the other way. But if that's totally wrong, then there must be something else weird going on. This post from Blizz explains the MMR decay system: http://us.battle.net/sc2/en/blog/12055065/situation-report-the-starcraft-ii-multiplayer-ladder-12-18-2013The highlighted/important parts:-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. The problem seemed to stem from people who only played a few games every 4+ weeks (like me!), which is beyond the max decay time. This would continually kick in the MMR decay system, causing those sorts of players to steadily fall over time. In your case, a period of not playing *at all* for a full 5 months would cause only that initial 4 week decay period. Someone else feel free to correct me if I am wrong, but this is how I understand everything to work, based on what Blizzard has said. edit: for some clarity and because I accidentally some words I see, but that still leaves unexplained some other things that can affect your effective MMR over time. I wonder what other factors have been at play that have made it seem as though MMR decay was doing more or less than that. If MMR decay is removed and those factors are still at play, then people will continue to observe weird things when they take breaks. If the idea is for people to be able to take any length break and always be able to match against the same people, then presumably the system needs to be adjusted in other ways.
Those "other factors" you mentioned might be referred to in this section from this most recent post on ladder:
"As we’ve gathered more data over the years, our analysis has allowed us to more precisely understand how various factors affect league placement and matchmaking over time. As part of this, we’ve identified a few internal functions that we feel can be adjusted or removed to improve the overall experience and allow for more consistent matching with evenly skilled opponents."
So, I suppose we just have to wait for the changes, and see if they make enough of a difference in the long run for everything to "feel normal" again.
|
When they say "this season" when will that be the case? right now? In couple of weeks?
|
On March 03 2015 15:07 zerK wrote:
When they say "this season" when will that be the case? right now? In couple of weeks?
That's the understanding, yes.
|
I love that they keep us posted. I don´t think it is a bad thing that little actual changes were introduced. They are just not ready yet and we were told in advance that they are aware problem of the problem, what the problem is in their opinion that they are working on it and how they plan to approach it and I appreciate it.
I don´t think the mid-season demotion is needed, since leagues are just a cosmetic thing. What I strongly dislike is the hidden MMR, not knowing where I stand. What should be implemented is matching cross-server if no opponent is found on current server, where one players gets reconnected to the other server, of course proper scaling would have to be initially done for the different MMRs. Another thing, that may already be implemented is adaptive MMR range based on opponents available.
|
Yes, pls fix it! Having recently started playing again 1v1 (usually I am playing diamond/masters) and 2v2 with a friend (usually plat/dia) who also just started again, we experienced weird shit concerning our matches. I got placed placed in gold 1v1 where I just facerolled everyone for 20 games now, while in 2v2 we got straight to plat and now play against masters/high diamond and get rolled ourself.
Please also relate the team MMR to the 1v1 MMR, if the player is active in both.
|
On March 03 2015 14:58 ZenithM wrote:Show nested quote +On March 03 2015 14:55 ZAiNs wrote:On March 03 2015 09:01 NonY wrote:During our analysis, we found that [MMR decay] was helping create even matches for players who were coming back from breaks and looking to ease back into StarCraft II. I wish. I didn't play on an account for over 5 months, came back and matched against top 50 GM. I guess the evidence shows that it takes more than 5 months for MMR decay to activate but skill in SC2 certainly drops much quicker than that. And I imagine that many of the people complaining about MMR decay were talking about periods of less than 5 months, and thus crying wolf the whole time. Unless the behind-the-scenes picture is much more complicated (inconsistent) and even worse at accomplishing its goals than we imagined. One change I hope to see for LotV is a different system for top ranks on the ladder. I don't know why the top 200 gets locked out from the rest of the system. The transition from divisions and leagues to a straightforward numerical ranking needs to be smoother. I say make it top 1000 or something, so any player who is decent (as in has any chance of going pro in the near future) can be ranked, and definitely do not let it get locked up. Make a system where people enter and leave based on performance, not activity, all season long. I had a similar experience with MMR decay on a smurf. I've also had a friend who 'decayed' from Master to Bronze, but that's because he got into Master after about 25 games the week HotS came out and then lost a placement game every season for a year. Most people crying about MMR decay probably haven't played enough games. Well, MMR decay in the first place is by design a feature FOR people who don't play enough games. I don't want to win 10 games in a row to start to meet remotely challenging opponents...
If you haven't played many games, then you DO want to play challenging opponents. From what I can tell the system is (was?) very good at putting you where you deserved to be as long as you put in enough games. Not playing many games, being on a loss/win streak makes the system more 'uncertain' about your MMR being an accurate reflection of your skill, and when its uncertain it can pair you vs. stronger/weaker players in an attempt to get you where you should be quickly. Like if you get a new account you can be GM MMR if you go 20-0, but if you lose your 2nd match it'll take a lot more games to get to the GM MMR. From my understanding this is separate from MMR decay which just lowers your MMR, people just seem to mistake the system being uncertain about their skill as MMR decay.
|
Hmmm, yes, I don't get what that has to do with what I'm saying, I know how MMR works. While I get the reasoning behind decaying MMR to not "discourage" people who've been away from the game for some time (imo they should just suck it up but eh ;D), they're way overdoing it right now. With my friend we won like 20 games in 2v2 the other day, constantly being matched up against golds, until FINALLY Blizzard decided that maybe it was time to make us play against platinum. Ridiculous.
And one decay timer per ladder mode (1v1, 2v2, ...) and per team is obviously flawed anyway, regardless of how well you implement the actual decay.
|
On March 03 2015 19:38 ZenithM wrote:
Hmmm, yes, I don't get what that has to do with what I'm saying, I know how MMR works. While I get the reasoning behind decaying MMR to not "discourage" people who've been away from the game for some time (imo they should just suck it up but eh ;D), they're way overdoing it right now. With my friend we won like 20 games in 2v2 the other day, constantly being matched up against golds, until FINALLY Blizzard decided that maybe it was time to make us play against platinum. Ridiculous.
And one decay timer per ladder mode (1v1, 2v2, ...) and per team is obviously flawed anyway, regardless of how well you implement the actual decay.
What I was trying to demonstrate is that MMR decay isn't for people who haven't played enough games, but people who have stopped playing for a period of time. People love to blame MMR decay for everything when it's that they've never actually played a solid chunk of matches.
|
It depends what you call a solid chunk of matches. Imo if you played about a thousand 2v2s with the same teammate, that should do it :D
Edit: Now that I think about it, I seem to recall that there are some settings where Blizzard resets completely your MMR (this has nothing to do with MMR decay). Something like not having played any game in a season, right? That's maybe where my grudge lies then.
|
On March 03 2015 19:57 ZenithM wrote:
It depends what you call a solid chunk of matches. Imo if you played about a thousand 2v2s with the same teammate, that should do it :D
Edit: Now that I think about it, I seem to recall that there are some settings where Blizzard resets completely your MMR (this has nothing to do with MMR decay). Something like not having played any game in a season, right? That's maybe where my grudge lies then.
I think I read something along those lines before, I think it would have to be at least a year though. I do think 2v2 in general is a lot more inconsistent than 1v1 just because of how irregularly most people play it.
|
|
|
Haha so much for 2/18/20/20/20/20. I think that's what they were gunning for.
|
According Blizzard's post, they are "reevaluating the ladder system" for LotV.
Here's one way to fix up the ladder system: http://www.teamliquid.net/forum/starcraft-2/389449-improving-the-ladder-system-for-hots
This suggestion bends over backwards to accommodate Blizzard's ridiculous and distortionary bonus pool system and fixes up it within the constraints of Blizzard's supposed "feel-good", "positive-psychology" philosophy (although bonus pool does the opposite). Recently, the main gripe with the current system is league distribution. This suggestion would bring transparency with very clear, very simple, very understandable promotion criteria: If diamond is the top 20%, then if you're in the top 20%, you get put in Diamond (or above).
Of course, we could just abandon these overrated constraints by Blizzard, and do a more radical, but more effective, overhaul:
-Just rank by MMR accounting for uncertainty of MMR.
-Display this metric.
-Get rid of leagues and replace it with "ladder levels", where you are in ladder level X, when you're in X percentile (i.e. instead of 7 leagues, have 100 leagues, and instantly promote or demote based on percentile). End the charade, abolish leagues.
A clear, transparent, and most importantly, truthful and non-obfuscated comparable rank for everyone.
Of course, Blizzard mostly realizes the benefits of this system as it is what they're already using in Hearthstone and Heroes of the Storm (except with uncertainty of MMR replaced by activity and 50 levels instead of 100).
2 unrelated points: Firstly, Blizzard's justification for removing MMR decay makes no sense. They praise it, but then they say that they decided to remove it. If people's skill decays due to inactivity, then MMR should decay too, because MMR measures skill. If it doesn't decay, then it is not accurate. Secondly, it beggars belief that Blizzard's system is so bad that they seemingly don't even know that their league distribution is off-target without people complaining on the forums, and then seemingly needing to manually check that it is true. In fact, it is surprising that they can't set up the logic of the system in such a way that it automatically enforces the league distribution approximately, without any manual intervention required. Moreover, Blizzard's fix appears to simply be calculating the MMR threshold based on the correct distribution for last season and applying it to next season. So why can't this simple computation (or whatever it is) be automated every season, instead of being triggered manually only when forum complaints flow in large volumes?
|
For me as an on-off player who has been playing in master / diamond for most of his sc2 life, the possible changes sound good. I've picked up the game again a few weeks ago and found myself in gold league, which wasn't really where I belonged. Got back to diamond with a 90% win rate, but I ran into lots of players who were complaining about opponents of higher skill level. Once you've reached a certain level, you just don't fall below that anymore. Especially if your mechanics are kind of solid. I really hope they'll get rid of mmr decay and make the whole ladder more transparent.
|
On March 04 2015 00:02 virpi wrote:
For me as an on-off player who has been playing in master / diamond for most of his sc2 life, the possible changes sound good. I've picked up the game again a few weeks ago and found myself in gold league, which wasn't really where I belonged. Got back to diamond with a 90% win rate, but I ran into lots of players who were complaining about opponents of higher skill level. Once you've reached a certain level, you just don't fall below that anymore. Especially if your mechanics are kind of solid. I really hope they'll get rid of mmr decay and make the whole ladder more transparent.
Yeah I agree. I got my overall level on the back of mechanics alone (mostly, I don't really play smart, nor am I interested in doing so), and that shit is like riding a bike, after 1-2 games I'm back to full speed and I don't need no damn decay to make me win 10 games in a row.
|
i just feel even talking about this is just creating drama and another 2 min waste of life. If you play the game a lot good or bad you will be placed against people fo similar skill, if you play now and again yes this will be kinda odd. I play a hell of a lot fo games and i can clearly tell you that i get about 2-4 mismatches per day, most days none. Its when you start flicking between ur ranked and your unranked or even other accounts where people start thinking there is a problem. My main account plays plats and dias with the occasional golds regularly, my other accnt which i dont play very much is dias and masters players mostly./ When i play i watch streams and they seem to be getting their leagues all time as well. noone on is complaining
so i have to ask., Who is actually complaining about the ladder? Lower people? If you are low then dont worry cos if you were better you would have the shiniest golden looking outline on your name or if not a glowing blue one . . . .just play the game youre only as good as you are. Im pretty sure all this is coming from those playwers who think they ought to be higher than they are with playing a few games. Play more games, after about 200 i would say ur mmr is exactly where it needs to be
|
United Kingdom20322 Posts
On March 03 2015 23:33 ZenithM wrote:
Haha so much for 2/18/20/20/20/20. I think that's what they were gunning for.
They changed it a while ago, making gold something like 35% target IIRC
however actual distribution is miles away from their targets.
Since people can't be demoted, if getting top 2% MMR meant masters promotion, the result would be an inflation in masters because there were people there that didn't deserve to be there, and new people got promoted in as they hit top 2% MMR. Masters would go to ~3-4% (the extra 1.5x for example from people who should be demoted, but have not been) but instead it's at like ~0.8%. It's way off
|
This might be the most vague statement blizzard has ever released. What are you going to do I wonder, you are just pointing out issues.
|
hey, that's cool. It actually makes me feel better to see that while we are "just plat" in 2v2 EU (the mode I play the most), we are among top 12.5 percent of teams (which sould have be diamond under the target percntages).
Still doesn't mean anything, but the Blizzard feel-good-system needs to be fixed so that more people feel good 
|
|
|
|
|
|
EPT
