StarCraft® II

Evidence of Broken Team Matchmaking

Posts: 606
I know a lot of people complain about team match making and how unfair it is, but they never provide any replays or proof. Well, here is some proof. I played several matches today, had a losing streak (9 games lost IN A ROW), and despite all of this kept getting matched against pre-made teams all with higher rankings. This isn't really an uncommon occurrence either, it happens fairly regularly.

Here are some pics I took before my matches:

http://img.photobucket.com/albums/v326/Deathleech/photo5.jpg
Played this team 3 or 4 times in a row, full pre-made of 4 players, all from same clan. Highest player was a Master on their team, our highest? Me and some other guy at gold.

http://img.photobucket.com/albums/v326/Deathleech/photo4-1.jpg
Another pre-made I faced 2 or 3 times in a row. Only two of them in the same clan but the other two were with them every time I faced them so they were definetly a pre-made team. 2 Masters players and a bronze vs us who had me (gold), a bronze, and 2 unranked.

http://img.photobucket.com/albums/v326/Deathleech/photo3.jpg
Not a pre-made, but they had 2 diamond and a gold vs us who had 2 gold and 2 unranked.

http://img.photobucket.com/albums/v326/Deathleech/photo2-1.jpg
Half pre-made from same clan, 1 plat, 1 diamond, and 2 masters vs 3 unranked and a gold

http://img.photobucket.com/albums/v326/Deathleech/photo1.jpg
Another half pre-made with two from same clan. Our team had me (gold), a silver, and 2 unranked vs a master, platinum, and 2 unranked

And here is my match history for today, 9 losses and all:
http://img.photobucket.com/albums/v326/Deathleech/photo-3.jpg

As you can see, not only did I play people with combined higher rankings most game, a lot of them were also on pre-made teams. Can someone please explain to me how this happens? Especially when I am losing game after game? I mean I can see getting placed against slightly higher ranked players if I am on a winning streak, but after I have lost several games in a row? It just takes all the fun out of this game.
Edited by BuffaloMeat on 6/2/2013 3:11 PM PDT
Reply Quote
Posts: 582
Yeah, I made anothere thread today with this same issue. I'm gold in 3v3 (bronze in 1v1) and together with 2 unranked friends we fought a premade team during our classification matches 2 times. One of them was a GM, the other a diamond, and the third was unranked in 1v1.

There's simply no way that could be balanced, and no reason to fight them twice in 5 classification matches. What can the game learn if I lose to the same team twice? Will it help placing me and my team mates in the correct league? I shouldn't even see the GM-colored border around a portrait while playing ladder, the difference in skill is simply absurd. I can provide replays if necessary, I have them here. But considering its Blizzard who has to do something about it I dont feel its necessary. They know what happens, if they want they can simply fix it.
Reply Quote
Posts: 39
the print screen button saves screenshots in your starcraft 2 folder

just sayin
Reply Quote
Posts: 1
The reason behind this is simple...even these teams have to play placement matches and it is just pure bad luck that you happened to encounter these teams so often. I just don't see how Blizzard could fix this apart from only allowing players to make RT matches if they are not in a group. But i think you will have to wait very long to find a game then....
Reply Quote
Posts: 606
The reason behind this is simple...even these teams have to play placement matches and it is just pure bad luck that you happened to encounter these teams so often. I just don't see how Blizzard could fix this apart from only allowing players to make RT matches if they are not in a group. But i think you will have to wait very long to find a game then....

So the team match making is more lenient to quicker place you in games? I can understand that and figured it had a little something to do with things . The only problem is, most of these matches aren't even close in skill level. Two masters players vs highest gold on the other team? Masters is almost twice as high ranking as gold. Then if that weren't bad enough, add to it they are a team and the matches are completely skewed.

I honestly would rather wait double or even triple the time to be placed in a match if it were actually even sides. Playing games where your chances or winning are next to zilch isn't fun at all, especially when they all out rank you and are a team and acting cocky because they know they will win. Losing ten games in a row because you are severely out ranked every match is lame.
Reply Quote
Posts: 742
there is honestly no way to perfectly balance team games
Reply Quote
Posts: 606
there is honestly no way to perfectly balance team games

Perfectly balanced is one thing. Currently it's no where even remotely near balanced though.

I would be fine facing a silver, 2 gold, and platinum player with myself, 2 other gold and a bronze player for example. That is not perfectly balance but is still fairly even. Playing a bunch of master players who are in a pre-made when your highest is gold and your all random is just ridiculous though. Its like playing 3v4, or playing with 50% hp on all your units. The chances of winning are slim to none.
Reply Quote
Posts: 606
This particular match I played today was really fun:
http://img.photobucket.com/albums/v326/Deathleech/Screenshot2013-06-0417_06_00.jpg

Notice they are a pre-made and their highest ranked player is a GRAND MASTER, while our highest are two gold? They also have diamond and platinum players on their team. So not only is their lowest ranked player higher than our teams highest, they are also a pre-made.

What kind of explanation can there possibly be for this, besides the match making in team games doesn't even try to match you against equal skilled players so it can place you faster?
Edited by BuffaloMeat on 6/4/2013 3:49 PM PDT
Reply Quote
Posts: 174
What kind of explanation can there possibly be for this, besides the match making in team games doesn't even try to match you against equal skilled players so it can place you faster?

When you are playing as a team you are matched according to your teams performance not 1 vs 1 performance. You are mixing up team rank and 1 vs 1 rank. They are not the same. If you look at those teams you have played against you will find that they are not ranked very high.

There can be several possibilities for this. They may actually suck as a team, or they may be smurfing, offracing or playing placement matches. If these players are members of many other low level teams the matchmaker may initially guess that their new team is not very skilled. Exactly how this is done is not known.
Reply Quote
Posts: 606
The thing is though, team games relate to 1v1. They aren't completely independent from each other. If I am bronze in 1v1 I am not going to magically be diamond or higher in team games or vis versa. It doesn't work that way. The fundamentals in 1v1 still apply to team matches as do your APM and knowledge of unit comps and statistics. Sure there might be some differences like all in 6 minute hits or a few Zerg 10 pooling work better than in 1v1, but all in all things arn't THAT much different. The biggest difference I ever see between team games and 1v1 when checking a players profile is 1 or 2 ranks MAX (unless they are intentionally losing in one or the other). Even I only have a 2 rank difference, and really it's because I have played so few 1v1. I have played around 10 matches total and won 8/10 of them so really I am probably platinum at the worst and just need to play more to get promoted.

A grand master player MIGHT not be a grand master in team games, but you can't possibly tell me he is gold level (which was our highest rank in that particular game). I would realistically guess he is masters at the worst. Also, most of the games I linked are me playing totally random vs pre-mades, these pre-mades undoubtedly have better coordination and communication than 4 random people placed together. The fact I have lost so many in a row should be proof the games are not evenly matched in the slightest.

I have come to the conclusion team games aren't remotely even matched up like they are in 1v1. Very rarely do both sides seem "equal", or the games close. Usually it just feels like I got placed against a much higher ranking pre-made because the match making wanted to place me quicker rather than find suitably same level opponents.
Reply Quote
Posts: 174
The only thing the matchmaking system cares about is how good the team actually performs. You seem to think that you could make a better matchmaking system by including other variables, like how they play in other matchups. My guess is that you couldn't. It would just lead to other problems, like a team that plays badly but according to "BuffaloMeats matchmaking system V2" shouldnt, and therefore keeps getting matched up against high level teams.

And this includes randoms vs arranged teams. The matchmaking system automatically compensates for the arranged teams communication advantage. Trying to second guess instead of just looking at their performance would be more complicated and error prone, and unlike the current system it would not self correct. You will in the long run have the same win ratio against random and arranged teams.
Reply Quote
Posts: 606
06/05/2013 09:43 AMPosted by Mendelfist
The only thing the matchmaking system cares about is how good the team actually performs. You seem to think that you could make a better matchmaking system by including other variables, like how they play in other matchups. My guess is that you couldn't. It would just lead to other problems, like a team that plays badly but according to "BuffaloMeats matchmaking system V2" shouldnt, and therefore keeps getting matched up against high level teams.

I don't think you are understanding. Yes, the match making cares about how good the other team is obviously and not just their 1v1 ranking. However, it's not like a pre-made with with diamond, master, and grand master players in 1v1 is going to be so bad in team matches that they are at the gold level. It doesn't work like that. The max difference MIGHT be a rank or two from team to single player matches, but to think it's anymore is just naive. I check profiles after a game ends and most are the same rank in 1v1 as team games. Usually they are higher ranked in team games than 1v1 if anything.

And this includes randoms vs arranged teams. The matchmaking system automatically compensates for the arranged teams communication advantage. Trying to second guess instead of just looking at their performance would be more complicated and error prone, and unlike the current system it would not self correct. You will in the long run have the same win ratio against random and arranged teams.

Incorrect. I have played 150+ team matches this season and my win rate is MUCH worse than 50/50. This is largely due to unfair team match ups where I am with random gold and silver players facing diamond or higher players on a pre-made who are obviously significantly better. Also, if I lost a game twice in a row against a pre-made with higher rankings, why on earth would the match making system place me against the same team two more times with similar random players on my team? I am just going to lose again.

Maybe if you were actually in these matches experiencing them you would better understand. That game with the GM player? Ya, he single handedly killed my silver and unranked partner. I am not sure how you can try to defend this system when I linked proof of me facing much higher ranked players and as a result losing over 9 games in a row versus them. It would be one thing to face higher ranked players and lose as a "test" to see if I should be promoted, but I face them over and over again even if I am on a losing streak. Also the 1v1 obviously does count for something because these players are handing me and my random team our asses, the matches aren't even close. I even tell my team exactly what the enemy is going to do if I face them previously and we STILL lose when they do the exact same builds.
Edited by BuffaloMeat on 6/5/2013 10:43 AM PDT
Reply Quote
Posts: 174
06/05/2013 10:43 AMPosted by BuffaloMeat
I don't think you are understanding.

I don't think you are reading. I didn't say that all matches are even. I even gave you examples of why they sometimes aren't. I gave you reasons why the system is designed the way it is, and that I don't think that you wouldn't be able to design a better one.

This is largely due to unfair team match ups where I am with random gold and silver players facing diamond or higher players on a pre-made who are obviously significantly better.

I have looked into this more than you have. A lot more. I have methodically gone through literally hundreds of team games, calculating and comparing differences between win ratios against arranged and random teams. They don't exist. This is also what Blizzard says, and they have access to more raw data than me. Unless something has changed lately, this is still the case. This doesn't prevent people from complaining about it however. They have done that since WoL beta.
Reply Quote
Posts: 606
Then please explain it better and enlighten me because all I have gotten from your responses are "team rank is all that matters" despite other variables that SHOULD come into play. If they are a pre-made, that should count for something, and also their 1v1 AND team ranks. I don't see why they shouldn't?

Also your argument just seems fundamentally wrong because you are saying that the matches are actually even (usually) based on team ranks. Well, if that were so wouldn't all these games where I am facing masters and GM players be actually close then? Instead my random team with gold 1v1 players gets SLAUGHTERED by the pre-made with master and GM players. You make it sound like a GM 1v1 is going to be gold in team and that's all that matters when placing them in games.

I mean I could understand if these were isolated incidences, or my team had really good team rankings and the matches were actually fairly close. They aren't though. I get slaughtered in team games because I am facing MUCH better players who are also on a pre-made. I linked 9 games where this happened and I lost every game. Every one of those games was against significantly better players and most were pre-mades.

I am not expecting the team match making to be perfect, but being relatively close would at least be nice. 3 gold and a bronze player vs 3 silver and a platinum pre-made is pretty close but not exactly even. 2 gold and 2 unranked randoms vs 2 diamond and a master pre-made is in no way even in no universe.
Reply Quote
Posts: 174
Then please explain it better and enlighten me because all I have gotten from your responses are "team rank is all that matters" despite other variables that SHOULD come into play. If they are a pre-made, that should count for something, and also their 1v1 AND team ranks. I don't see why they shouldn't?


The system does take other things into account. As I said, when you start a new team the system uses your other teams and possibly even 1 vs 1 rank as a starting point. If you are only member of master teams your first placement will be against a master opponent. From then on however, the new teams MMR is updated only according to wins and losses.
It's possible that the system could be made better, but it's not easy without breaking other things.

I linked 9 games where this happened and I lost every game. Every one of those games was against significantly better players and most were pre-mades.

Are you sure you are looking at every game? Memories tend to be selective. I didn't have to go far back in your match history to find this game:
Your team: Diamond, Gold (you), Unranked, Unranked (previously gold).
Opponents: Unranked (previously gold), Master, Master, Master.

You won this game. It looks grossly unbalanced, but as I said. using 1 vs 1 rank other than as a starting point would be dangerous. I don't think it's easy to improve the current system, even though the matches often are impossible to win.
Edited by Mendelfist on 6/6/2013 4:57 AM PDT
Reply Quote
Posts: 606
The system does take other things into account. As I said, when you start a new team the system uses your other teams and possibly even 1 vs 1 rank as a starting point. If you are only member of master teams your first placement will be against a master opponent. From then on however, the new teams MMR is updated only according to wins and losses.

It's possible that the system could be made better, but it's not easy without breaking other things.

So you are saying I could make a team, we all could purposely play bad to bomb our rating and smurf, then make another team where we actually try just to move that team up when we face bad players? In essence one team counter balances the other since the MMR would be in the middle because on one team we purposely play bad and the other we play our best, thereby playing worse players than we should and winning on the "real team". If this is how it works that seems like it needs to seriously be looked at. I mean sure you can do it in 1v1, but at least in 1v1 you only have one team - yourself, so it's not exactly the same. You are stuck in bronze if you smurf until you start winning a lot, thereby not really gaining anything except easy pointless wins (for achievements I suppose). You don't cheat the system and make your rank go up more than it should by playing worse players.

Are you sure you are looking at every game? Memories tend to be selective. I didn't have to go far back in your match history to find this game:
Your team: Diamond, Gold (you), Unranked, Unranked (previously gold).
Opponents: Unranked (previously gold), Master, Master, Master.

You won this game. It looks grossly unbalanced, but as I said. using 1 vs 1 rank other than as a starting point would be dangerous. I don't think it's easy to improve the current system, even though the matches often are impossible to win.

To answer your question, I have tried to look at the whole picture. Sure there are some games where my team is much higher rank than the opponent, or my team looks like it's weaker and we somehow manages to win, but these cases are fairly rare. That game you are referring to is one such isolated case. I believe that game was played yesterday if you are referring to the game I think you are. Also, it was a long hard battle. We got lucky and did a time push and knocked out two of their players while the other two expanded. We pretty much were 2v4 most the game and even still it was a long, close game that lasted much longer than it should of if the teams had been "even".

I really wish the system tracked unranked games. I started playing unranked because I got frustrated constantly getting matched against significantly better teams and losing. At least in unranked I am not losing anything, it's just not as fun.
Reply Quote
Posts: 174
06/06/2013 09:37 AMPosted by BuffaloMeat
So you are saying I could make a team, we all could purposely play bad to bomb our rating and smurf, then make another team where we actually try just to move that team up when we face bad players?

I think so, but it's only the first placement match as far as I know that uses other teams as a starting point. The starting MMR has to be guessed anyway. If there isn't anything else to use it uses middle gold or something. In any case, I don't think it's exactly known how this works.
Reply Quote
Posts: 21
An approach I found interesting in other sports (I am thinking league Pool off the top of my head) was that of a 'points' system, where players at certain levels were attributed x amount of points and teams could be made up of any combination but the bands they could play in would be limited according to their points totals.. in this instance to use a very basic example

M - 10
D - 8
P - 6
G - 4
S - 2
B - 1

the lowest team score in a 4v4 being 4 compared to that of 40 for a team of Masters, the combination banding could be something along the lines of 4-10, 11-20, 21-30, 31-40, even opened so that you could play one band above/below you as well ifa game isn't found in say... 2/3 minutes with your own band. I appreciate MMR makes the low totals I have used as an example redundant and that this scoring level exists in-game, and whether there could be something like a 5-10 point 'penalty' for arranged teams, not enough to discourage them from playing, but enough that perhaps this doesn't occur as frequently.

Obviously this is a very loose set of examples but other than separating 'arranged' vs 'casual' into another separate league, perhaps as the numbers are already there for the system to interpret?

On second thoughts.... I may have just described Matchmaking.... never mind, it has been a long day :s
Reply Quote
Posts: 30
I partially agree with that matchmaking is broken because, a few days ago, I got placed into bronze league, and i've been facing silver/gold players. I can beat them, but why go from bronze players immediately to gold or silver?
Reply Quote
Posts: 64
I partially agree with that matchmaking is broken because, a few days ago, I got placed into bronze league, and i've been facing silver/gold players. I can beat them, but why go from bronze players immediately to gold or silver?


Your MMR is still volative, and the system tries to find your rightful place. Be a patient and it will soon pair you with opponents of your skill. In your case, it's working as intended.
Reply Quote

Please report any Code of Conduct violations, including:

Threats of violence. We take these seriously and will alert the proper authorities.

Posts containing personal information about other players. This includes physical addresses, e-mail addresses, phone numbers, and inappropriate photos and/or videos.

Harassing or discriminatory language. This will not be tolerated.

Forums Code of Conduct

Report Post # written by

Reason
Explain (256 characters max)
Submit Cancel

Reported!

[Close]