|
On April 03 2011 10:24 Incanus wrote:I've been following this whole ladder analysis chain of threads since beta, but something very interesting just happened which I cannot explain. Last season I was ~35th diamond. This season, my placement match was against a diamond player who went mass void rays (yes, cannons straight into void rays), and whined about the matchup being broken when I walked into his base. As a result I got into platinum. Nothing too special so far... However, 4 games later, after accumulating 98 points total, I got into diamond, with 122 points. What's the explanation behind this?  Edit: Ah, I guess... When you get promoted, your new rating is C + bonus pool, regardless of what your previous points were?
Edit: As Lysenko said bellow, your points are reset to 73 + spent bonus pool, but your matchmaking rating won't change because of a promoting/demoting.
|
Thanks a lot for posting this !
|
Edit: Ah, I guess... When you get promoted, your new rating is C + bonus pool, regardless of what your previous points were?
Your rating (MMR) does not change, but your point total does.
|
@Excalibur_Z
I was wondering about something. I saw on the EU forums that someone noticed that he is facing a lot of Mirror Matches. I've kept the stats of every ladder match I played in a spreadsheet and I conclude this is actually the case for me too: (btw I play zerg, EU, Platinum)
Season 1: (EU server): 88x ZvZ, 66x ZvT & 77x ZvP (note the funny coincidence it's all a multitude of 11)
Now in Season 2: (EU server), (I realize it's less matches): 13x ZvZ, 6x ZvT & 9x ZvP
Eventhough Zerg isn't the most played race (in platinum, EU) Over the course of 259 games I've faced 39% mirrors (ZvZ), 28% vs Terran, and 33% vs Protoss.
Now is this coincidence? Or is there an explanation in the system? I have at least a small hint that this could be on purpose. I win more than 50% of my ZvZ games, while I win less than 50% in my ZvT & ZvP.. since the system tries to keep my W/L-ratio at 50%, could it be it searches for more ZvZ matches, helping me out in keeping my W/L-ratio to 50%? What do you think?
|
United States12252 Posts
On April 05 2011 05:21 Gulzt wrote:
@Excalibur_Z
I was wondering about something. I saw on the EU forums that someone noticed that he is facing a lot of Mirror Matches. I've kept the stats of every ladder match I played in a spreadsheet and I conclude this is actually the case for me too: (btw I play zerg, EU, Platinum)
Season 1: (EU server): 88x ZvZ, 66x ZvT & 77x ZvP (note the funny coincidence it's all a multitude of 11)
Now in Season 2: (EU server), (I realize it's less matches): 13x ZvZ, 6x ZvT & 9x ZvP
Eventhough Zerg isn't the most played race (in platinum, EU) Over the course of 259 games I've faced 39% mirrors (ZvZ), 28% vs Terran, and 33% vs Protoss.
Now is this coincidence? Or is there an explanation in the system? I have at least a small hint that this could be on purpose. I win more than 50% of my ZvZ games, while I win less than 50% in my ZvT & ZvP.. since the system tries to keep my W/L-ratio at 50%, could it be it searches for more ZvZ matches, helping me out in keeping my W/L-ratio to 50%? What do you think?
I really really doubt it. There's not enough evidence for or against to say for sure, but a truly unbiased system would want to ignore such and instead match players solely by MMR. If anything you're getting more mirrors because the maps you vetoed aren't good for your race in general, and the other people who play your race feel the same way.
|
On April 05 2011 06:29 Excalibur_Z wrote:Show nested quote +On April 05 2011 05:21 Gulzt wrote:
@Excalibur_Z
I was wondering about something. I saw on the EU forums that someone noticed that he is facing a lot of Mirror Matches. I've kept the stats of every ladder match I played in a spreadsheet and I conclude this is actually the case for me too: (btw I play zerg, EU, Platinum)
Season 1: (EU server): 88x ZvZ, 66x ZvT & 77x ZvP (note the funny coincidence it's all a multitude of 11)
Now in Season 2: (EU server), (I realize it's less matches): 13x ZvZ, 6x ZvT & 9x ZvP
Eventhough Zerg isn't the most played race (in platinum, EU) Over the course of 259 games I've faced 39% mirrors (ZvZ), 28% vs Terran, and 33% vs Protoss.
Now is this coincidence? Or is there an explanation in the system? I have at least a small hint that this could be on purpose. I win more than 50% of my ZvZ games, while I win less than 50% in my ZvT & ZvP.. since the system tries to keep my W/L-ratio at 50%, could it be it searches for more ZvZ matches, helping me out in keeping my W/L-ratio to 50%? What do you think? I really really doubt it. There's not enough evidence for or against to say for sure, but a truly unbiased system would want to ignore such and instead match players solely by MMR. If anything you're getting more mirrors because the maps you vetoed aren't good for your race in general, and the other people who play your race feel the same way.
I thought all the speculation would seem to imply that map vetos don't matter, since the number is limited enough that whoever they pair you with, they can find a map that works for both of you. Is there some reason to believe map vetos effect matchmaking?
|
On April 05 2011 06:32 aristarchus wrote:Show nested quote +On April 05 2011 06:29 Excalibur_Z wrote:On April 05 2011 05:21 Gulzt wrote:
@Excalibur_Z
I was wondering about something. I saw on the EU forums that someone noticed that he is facing a lot of Mirror Matches. I've kept the stats of every ladder match I played in a spreadsheet and I conclude this is actually the case for me too: (btw I play zerg, EU, Platinum)
Season 1: (EU server): 88x ZvZ, 66x ZvT & 77x ZvP (note the funny coincidence it's all a multitude of 11)
Now in Season 2: (EU server), (I realize it's less matches): 13x ZvZ, 6x ZvT & 9x ZvP
Eventhough Zerg isn't the most played race (in platinum, EU) Over the course of 259 games I've faced 39% mirrors (ZvZ), 28% vs Terran, and 33% vs Protoss.
Now is this coincidence? Or is there an explanation in the system? I have at least a small hint that this could be on purpose. I win more than 50% of my ZvZ games, while I win less than 50% in my ZvT & ZvP.. since the system tries to keep my W/L-ratio at 50%, could it be it searches for more ZvZ matches, helping me out in keeping my W/L-ratio to 50%? What do you think? I really really doubt it. There's not enough evidence for or against to say for sure, but a truly unbiased system would want to ignore such and instead match players solely by MMR. If anything you're getting more mirrors because the maps you vetoed aren't good for your race in general, and the other people who play your race feel the same way. I thought all the speculation would seem to imply that map vetos don't matter, since the number is limited enough that whoever they pair you with, they can find a map that works for both of you. Is there some reason to believe map vetos effect matchmaking?
I think Excalibur_Z was trying to explain there are too many factors that could be attributed to Gulzt's pattern of opponents to start making any serious claim at an intentional mechanic in the MMR system...
As an example in addition to map veto:
Perhaps Gulzt also tends to play during times where those race ratios are common,...
(ie: More Zerg playing, less Terran playing)
There were definite trends in the Myers Briggs data collection that showed a clear personality bias for the race population distribution... so why not similar trends for daily/hourly active race density?
|
mhh im faceing many mirror matches too.
21 x ZvZ (63% winratio)
14 x ZvP (28% " )
14 x ZvT (57% " )
i didnt used sc2gears for season1 and dont have many replays from s1 ;|
i know thats not enough games to say anything about it, but it looks similar according to gulzt post.
|
@Excalibur_Z
I agree with that my profile alone is no proof, but it's interesting if it were to be true. How about your own stats?
I only voted down "delta quadrant", and only since season 2. I didn't vote down a single map in season 1. The time of play could play a role but that's not very likely either, since I play at varying times: morning, afternoon, evening, weekend, every day of the week, though I tend to play slightly more in the evenings during the week..
@ Luxer, in what league are you?
|
@Gulzt - In case you're interested, my sc2gears data is as follows:
ZvP - 96
ZvT - 96
ZvZ - 101
I veto'd maps I didn't like in Season 1. Trying to not do it any more.
|
People have speculated in other threads that ZvZ is the most common matchup for many Zerg players because ZvZ has the shortest average duration out of any matchup, meaning that at any time, there are more Z players searching for games than players of other races. This is purely speculation though.
|
On April 05 2011 23:48 TestSubject893 wrote:
People have speculated in other threads that ZvZ is the most common matchup for many Zerg players because ZvZ has the shortest average duration out of any matchup, meaning that at any time, there are more Z players searching for games than players of other races. This is purely speculation though.
Sounds like a reasonable speculation, but aren't PvP games the shortest currently? Wouldn't I face more toss players? Especially since Protoss is the most played race in every europe league...
|
On April 05 2011 23:55 Gulzt wrote:Show nested quote +On April 05 2011 23:48 TestSubject893 wrote:
People have speculated in other threads that ZvZ is the most common matchup for many Zerg players because ZvZ has the shortest average duration out of any matchup, meaning that at any time, there are more Z players searching for games than players of other races. This is purely speculation though. Sounds like a reasonable speculation, but aren't PvP games the shortest currently? Wouldn't I face more toss players? Especially since Protoss is the most played race in every europe league...
I don't know...
I've seen a lot of Zerg rage quit on me the moment they saw me block my ramp with two Queens so their Baneling all in was inevitably going to fail...
I'm not sure how fast PvP is usually over, but ZvZ seems to only go more than 10 minutes when both players push for a macro game like I do,... and many of my ZvZs last 8 minutes or less. :-(
Does a PvP often end right after that 6 minute timing push?
Otherwise, it sounds like a valid thought to me,...
|
@Phinix in the GSL they tend to be shorter than the ZvZ matches if I'm not mistaken. Still, given that PvP is short too (even if it's longer), and more people play toss than zerg; it seems not to be the most convincing explanation (though I can't disproof it either)
|
thanks for summarizing this. it helped me a lot to understand some twists.
here are two cases that I found hard to understand. now I am a bit closer.
1) - instantly promoted to masters after inactivity
4 Month ago, I played around 50 games 33random (41-10) and got promoted to diamond. Until Season 2 I never played 33random again. Now I won 1 game and got instantly promoted to diamond. This is what I wasn't expecting at all.
2) - I am winning and winning in 44random, but don't get promoted.
I was on 100 games in 44random (diamond) last season and won ~75%. Not much, but not the worst thing.
Patch 1.3 was released and three days before league promotions were re-enabled, I started pumping mass 44random games. All in all, since patch1.3 I won around 40 games and only lost 3. I still don't get promoted and don't know what else I have to do. Now I am leading my division with 20% more points than next places. Winning game after game.. but I dont get any promotion.
Each game results in +4 up to +14 points.
Your post made become more patient for that whole promotion demotion thing.
Especially the activity informations enlightend me. Thanks
|
Thanks for the great work. Your guide has been invaluable in describing the ladder system.
On March 18 2011 05:52 Excalibur_Z wrote:
- Players really are "locked" into GM league based on their MMR's moving average for the entire season until they accumulate 270 bonus pool, at which point they're dropped and a slot opens up. If they stay below 270 they won't be removed even if better players emerge.
To clarify, this means that the people promoted to grandmasters after 2 weeks will not be the same 200 at the top of sc2ranks, but the 200 with the highest MMR (assuming they meet the bonus pool requirements).
|
I'll add my data about MUs too, in case anyone is interested:
Season 1(Platinum):
Total Games: 314
ZvZ: 110 - 44% Win ratio
ZvP: 109 - 56% Win ratio
ZvT: 95 - 50% Win ratio
Season 2(Platinum -> Diamond -> Platinum, 6 total games as plat):
Total Games: 52
ZvZ: 21 - 28% Win ratio
ZvP: 14 - 71% Win ratio
ZvT: 17 - 52% Win ratio
On a note, I'm below 50% on season 2(25-28, 1 disc)
|
Of my last 68 games:
28/68=41% PvP
26/68=38% PvT
14/68=21% PvZ
The highest % is mirror for me as well. If the theory about there being more zergs queuing due to short games were true, everyone should match up with more zergs. That isn't true in my case.
Comparing to race statistics for NA plat from sc2ranks: P-33% T-25% Z-31% R-11%. The random players should draw each race equally, so the opponent race distribution should look like P-37% T-28% Z-35%. It appears that in my games, Terrans have been overrepresented, and Zergs underrepresented. 68 games is not that many, but I would have thought it was enough for a 14% difference just by chance to be unlikely. I think I have scrap station and delta quadrant vetoed.
|
Excalibur, I have a quick question about placement into the GM league.
Its based purely on moving average MMR and not simply top 200 in points correct? So even though a player may not be currently in the top 200 in points, as long as he is only matched up against top 200 players, and appears even or favored from the top 200 player's persepctive, he should have no trouble getting in.
Yes, no?
|
United States12252 Posts
On April 06 2011 23:36 ch33psh33p wrote:
Excalibur, I have a quick question about placement into the GM league.
Its based purely on moving average MMR and not simply top 200 in points correct? So even though a player may not be currently in the top 200 in points, as long as he is only matched up against top 200 players, and appears even or favored from the top 200 player's persepctive, he should have no trouble getting in.
Yes, no?
That's right.
|
|
|
|
|
|
EPT
