The state of EU Gold

[ksanaks.2380]


 

This was an insta pop at ∼1300 around 23 gmt+1, no way these guys are beyond G2/G3.

Can you blame them? I honestly dont.

Theyre not doing this for titles, what theyre doing (I presume) is getting to quality matches in the fastest and easiest way possible, skipping the gold zoo.

How can you fix this, besides enforcing 2/3/5 man party?

 

[Alatar.7364]

> @"ksanaks.2380" said:

>

 

I am honestly a bit confused as to what is that you're trying to point out? Or what you're saying actually.

 

If it's about the score difference, than frankly 385 - 500 is on the good side of match balance. It's the >300 - 500 that are obvious.

[SoulSin.5682]

He believes that its unfair that two duos are fighting a team with a single duo.

 

Soon enough you will understand that having too much duos in a team is actually bad.

I can't even fathon to start mentioning the times I got, at Gold 2~3, tagged with duos of Silvers just to get obliterated against a team of Gold 3/Plat 1 pugs.

 

The game is really harsh for duoQs. Or perhaps I am just cursed.

[Megametzler.5729]

> @"SoulSin.5682" said:

> He believes that its unfair that two duos are fighting a team with a single duo.

>

> Soon enough you will understand that having too much duos in a team is actually bad.

> I can't even fathon to start mentioning the times I got, at Gold 2~3, tagged with duos of Silvers just to get obliterated against a team of Gold 3/Plat 1 pugs.

>

> The game is really harsh for duoQs. Or perhaps I am just cursed.

 

So far we do not have any information on whether duos are even considered in the matchmaker - apart from them getting in one team of course. So far I actually doubt duos get a "+20 MMR for the average team-MMR calculation". Nothing in the matchmaker points in that direction and I believe we've never been told anything. High ranked double duos against 5 randoms happens quite often.

 

So I fear you've just been unlucky. :tongue:

 

Maybe OP also means the renegade and scrapper on his team.

[ksanaks.2380]

> @"Megametzler.5729" said:

> Maybe OP also means the renegade and scrapper on his team.

Or maybe what OP meant is that 6 players synchro queued at the same time to fish for a match that they could fix

Maybe, you know

The more I read the pvp forums the more I realize that people who post here dont actually play the game

[Megametzler.5729]

> @"ksanaks.2380" said:

> > @"Megametzler.5729" said:

> > Maybe OP also means the renegade and scrapper on his team.

> Or maybe what OP meant is that 6 players synchro queued at the same time to fish for a match that they could fix

> Maybe, you know

> The more I read the pvp forums the more I realize that people who post here dont actually play the game

 

Yes. I don't know. We are all just guessing, it seems. Care to elaborate? How do you think a ~400 to 500 match is a problem?

[saerni.2584]

> @"ksanaks.2380" said:

> > @"Megametzler.5729" said:

> > Maybe OP also means the renegade and scrapper on his team.

> Or maybe what OP meant is that 6 players synchro queued at the same time to fish for a match that they could fix

> Maybe, you know

> The more I read the pvp forums the more I realize that people who post here dont actually play the game

 

That isn’t clear at all from your post. Rather than making assumptions about who does and doesn’t play the game you should focus on writing coherently.

 

No one is jumping on you unfairly.

[JusticeRetroHunter.7684]

> @"Megametzler.5729" said:

> > @"SoulSin.5682" said:

> > He believes that its unfair that two duos are fighting a team with a single duo.

> >

> > Soon enough you will understand that having too much duos in a team is actually bad.

> > I can't even fathon to start mentioning the times I got, at Gold 2~3, tagged with duos of Silvers just to get obliterated against a team of Gold 3/Plat 1 pugs.

> >

> > The game is really harsh for duoQs. Or perhaps I am just cursed.

>

> So far we do not have any information on whether duos are even considered in the matchmaker - apart from them getting in one team of course. So far I actually doubt duos get a "+20 MMR for the average team-MMR calculation". Nothing in the matchmaker points in that direction and I believe we've never been told anything. High ranked double duos against 5 randoms happens quite often.

>

 

They are in fact considered. According to the wiki, they heavily influence the matchmaking. If you look through the code you’ll notice the phrase “number of rosters.” “Rosters” is a team like a duo or a 5 man that is que’d.

 

Other factors influence the matchmaking, including TOTAL number of games played...that means that if you have played 8000 games your “value” will be though of as being higher than someone with 100 games...

 

The code basically takes a bunch of variables and attempts to balance them to a singular value...so the above example would mean you get placed into a team that have played less games, have lower mmr etc...in order to balance out the amount of games you’ve played.

 

Edit: what should also be mentioned is that this singular value the matchmaker tries to balance matches on is not a set equal number. The singular value is an estimate of all the available variables involved and then based on this it will determine whether one team is slated to win, and the other slated to lose. Without knowing all the variables of the OP’s match we can’t determine if he was slated to win or lose. Simply put, just because one team has 2 rosters vs a team with one roster is not enough information to determine if it was fair or not, because there are other variables (games played, rank, hidden rank etc) that go into the calculation to balance the teams.

 

[Crystal Paladin.3871]

> @"ksanaks.2380" said:

> 

>

> This was an insta pop at ∼1300 around 23 gmt+1, no way these guys are beyond G2/G3.

> Can you blame them? I honestly dont.

> Theyre not doing this for titles, what theyre doing (I presume) is getting to quality matches in the fastest and easiest way possible, skipping the gold zoo.

> How can you fix this, besides enforcing 2/3/5 man party?

>

 

I guess he's trying to point out that he's secured 17 kills and yet he lost the match while playing in red team. The blue had double duo advantage and he hard carried the team to 385... If he didn't squeeze every bit of his performance, the match would've ended up in 100 - 500...

 

He could've mentioned that

[rank eleven monk.9502]

> @"JusticeRetroHunter.7684" said:

> > @"Megametzler.5729" said:

> > > @"SoulSin.5682" said:

> > > He believes that its unfair that two duos are fighting a team with a single duo.

> > >

> > > Soon enough you will understand that having too much duos in a team is actually bad.

> > > I can't even fathon to start mentioning the times I got, at Gold 2~3, tagged with duos of Silvers just to get obliterated against a team of Gold 3/Plat 1 pugs.

> > >

> > > The game is really harsh for duoQs. Or perhaps I am just cursed.

> >

> > So far we do not have any information on whether duos are even considered in the matchmaker - apart from them getting in one team of course. So far I actually doubt duos get a "+20 MMR for the average team-MMR calculation". Nothing in the matchmaker points in that direction and I believe we've never been told anything. High ranked double duos against 5 randoms happens quite often.

> >

>

> They are in fact considered. According to the wiki, they heavily influence the matchmaking. If you look through the code you’ll notice the phrase “number of rosters.” “Rosters” is a team like a duo or a 5 man that is que’d.

>

> Other factors influence the matchmaking, including TOTAL number of games played...that means that if you have played 8000 games your “value” will be though of as being higher than someone with 100 games...

>

> The code basically takes a bunch of variables and attempts to balance them to a singular value...so the above example would mean you get placed into a team that have played less games, have lower mmr etc...in order to balance out the amount of games you’ve played.

>

I cant check the algorythm right now but I am pretty sure you are talking about unranked matchmaking here (e.g. Number of played games)

[DanAlcedo.3281]

He has 17 kill and not even the most kills on his team.

 

Looks like Red farmed Blue but they didn’t cap points.

 

Or something.

[Dadnir.5038]

> @"Crystal Paladin.3871" said:

> I guess he's trying to point out that he's secured 17 kills and yet he lost the match while playing in red team. The blue had double duo advantage and he hard carried the team to 385... If he didn't squeeze every bit of his performance, the match would've ended up in 100 - 500...

>

> He could've mentioned that

 

Strange, from the same numbers I conclude that it's team sucked at keeping point but exceled at killing foes. Which mean that it's team didn't really play to win or lacked the will to play "conquest". Ultimately, capturing and keeping points is important, focusing on killing players won't give enough to catch up and win the game.

 

There were 101 kills in the game, it's team was responsible for 72 kills and he merely did 17 of them. That was a slugfest on mid I guess.

[Crann.1342]

Nothing to see here, just your typical delusional everyone is a wintrader post. Yes wintrading happens but not nearly as often as these conspiracy theorists on the forums make it out to. Realise you are being paranoid before it starts to effect your mental health irl

[JusticeRetroHunter.7684]

> @"rank eleven monk.9502" said:

> > @"JusticeRetroHunter.7684" said:

> > > @"Megametzler.5729" said:

> > > > @"SoulSin.5682" said:

> > > > He believes that its unfair that two duos are fighting a team with a single duo.

> > > >

> > > > Soon enough you will understand that having too much duos in a team is actually bad.

> > > > I can't even fathon to start mentioning the times I got, at Gold 2~3, tagged with duos of Silvers just to get obliterated against a team of Gold 3/Plat 1 pugs.

> > > >

> > > > The game is really harsh for duoQs. Or perhaps I am just cursed.

> > >

> > > So far we do not have any information on whether duos are even considered in the matchmaker - apart from them getting in one team of course. So far I actually doubt duos get a "+20 MMR for the average team-MMR calculation". Nothing in the matchmaker points in that direction and I believe we've never been told anything. High ranked double duos against 5 randoms happens quite often.

> > >

> >

> > They are in fact considered. According to the wiki, they heavily influence the matchmaking. If you look through the code you’ll notice the phrase “number of rosters.” “Rosters” is a team like a duo or a 5 man that is que’d.

> >

> > Other factors influence the matchmaking, including TOTAL number of games played...that means that if you have played 8000 games your “value” will be though of as being higher than someone with 100 games...

> >

> > The code basically takes a bunch of variables and attempts to balance them to a singular value...so the above example would mean you get placed into a team that have played less games, have lower mmr etc...in order to balance out the amount of games you’ve played.

> >

> I cant check the algorythm right now but I am pretty sure you are talking about unranked matchmaking here (e.g. Number of played games)

 

Na, it's for ranked as well.

 

Here is a copy paste from the wiki of the newest code we have available...keep in mind it is considered a "psuedo-code" which i don't exactly know how these codes are created, lets at least assume that the code is accurate to what we have in game with the current matchmaker;

______

Pseudo-Code (New February 7th 2017)

A new matchmaker has been written to solve some of the failings of the previous while maintaining a similar flow. This new matcher will score rosters against both teams and the entire match instead of only considering alternating target teams. This is most notable when scoring ratings as a roster's fit is based on how it will balance team ratings instead of just how close it is to the target team's rating. One additional scoring parameter includes a bonus for balancing profession counts.

______

def createMatches(queue, config):

rosters = queue

failed = []

 

# try to make a match for each roster in the queue

while len(rosters) > 0:

roster = rosters.pop()

queue.remove(roster)

if not tryMakeMatch(roster, queue, config):

failed.append(roster)

 

# move rosters we couldn't find match for to the end of the queue

queue.append(failed)

 

 

def tryMakeMatch(target, queue, config):

# gather rosters that are good potential matches

potentials = gatherPotentials(target, queue, config)

 

# enforcing a minimum allows some flexibility in choices

if len(potentials) < config.potentials.min:

return False

 

team1 = []

team2 = []

match = []

 

team1.append(target)

match.append(target)

 

while len(match) < config.teamSize * 2:

 

bestRoster = None

bestTeam1Score = -infinity

bestTeam2Score = -infinity

 

for roster in potentials:

# score the roster against team1 and the match

team1Score = -infinity

if canJoinTeam(roster, team1, team2, config):

team1Score = scoreRoster(roster, team1, team2, match, config)

 

# score the roster against team2 and the match

team2Score = -infinity

if canJoinTeam(roster, team2, team1, config):

team2Score = scoreRoster(roster, team2, team1, match, config)

 

# found a better roster!

if bestRoster is None or max(team1Score, team2Score) > max(bestTeam1Score, bestTeam2Score):

bestRoster = roster

bestTeam1Score = team1Score

bestTeam2Score = team2Score

 

# could not find any roster for the match, abort

if bestRoster is None:

return False

 

# add this player to whichever team is a better fit

match.append(bestRoster)

if bestTeam1Score > bestTeam2Score:

team1.append(bestRoster)

else

team2.append(bestRoster)

 

queue.remove(team1)

queue.remove(team2)

 

createMatch(team1, team2)

 

return True

 

 

def gatherPotentials(queue, target, config):

potentials = []

 

for roster in queue:

# check conditions where rosters are never allowed to match

if roster.gameModes != target.gameModes:

continue

if roster.rating > target.ratingHigh:

continue

if roster.rating < target.ratingLow:

continue

 

potentials.append(roster)

 

# limit choices for performance

if len(potentials) >= config.filter.potentials.max:

break

 

return potentials

 

 

def canJoinTeam(roster, team, otherTeam, match, config):

# roster is too big for this team

if len(team) + len(roster) > config.teamSize:

return False

 

# don't pick rosters that are much different size than what exists

if abs(len(roster) - otherTeam.maxRosterSize) > config.rosterSize.maxDiff:

return False

 

return True

 

 

def scoreRoster(roster, team, otherTeam, match, config):

score = 0

 

# adjust score by time queued

score += roster.age * config.age.seconds

 

# adjust score by games played difference

distance = abs(match.averageGames - roster.games)

score += distance * config.rating.distance

 

# adjust score by rank difference

distance = abs(match.averageRank - roster.rank)

score += distance * config.rank.distance

 

# adjust score by roster size difference

distance = abs(len(roster) - otherTeam.maxRosterSize)

score += distance * config.rosterSize.distance

 

# adjust score by POTENTIAL rating difference

distance = abs(team.ratingWithRoster(roster) - otherTeam.rating)

score += distance * config.rating.distance

 

# adjust score by profession counts

for profession in allProfessions:

# roster has none of these professions

if roster.count(profession) == 0:

continue

 

# too many of the same profession

totalCount = roster.count(profession) + team.count(profession)

if totalCount > config.professions.max:

score += (totalCount - config.professions.max) * config.professions.common

 

# otherwise favor the variety

elif team.count(profession) == 0:

score += config.professions.unique

 

# favor matching professions between teams

if team.count(profession) < otherTeam.count(profession):

score += config.professions.matching

 

return score

_______

 

If you look carefully, "Return Score" is the singular value that the matchmaker will try to attain. Earlier i was wrong about the "match prediction" as it's now out of use...so for now it's just a single score that the matchmaker will, rather than assuming a team will win or lose, it will just do it's best to attain this singular value.

 

Now, notice the "Roster" terminology here. A "Roster" consists of a minimum of 1 player. So one person que'ing into a game is considered a roster. 2 people in a duo is considers a single roster with 2 sets of variables, that it will then combine, value and match it against the variables of the other rosters it's considering to create a match.

 

Now notice the "Score Roster" section. It will add or take away value by judging your games played, your time in the que, your roster size, proffesion type, number of professions in the current pool of players it's considering for a match, Dishonor, and "potential" rating difference. (this has to do with a deviation calculation...it's important but a bit complicated to explain, best to just look at the table on the wiki.") and rating. You will notice that there are different formula's based on each of these scoring sections. For example;

**_ # adjust score by roster size difference

distance = abs(len(roster) - otherTeam.maxRosterSize)

score += distance * config.rosterSize.distance

 

verses

# adjust score by games played difference

distance = abs(match.averageGames - roster.games)

score += distance * config.rating.distance_**

 

To keep it simple, these formulas determine a part of the score, by weighting it against other variables in the matchmaker. With each "step" it will adjust the score, one after the other until it returns the final score. It will then use these scores to determine at the end if the match gets the "Okay" to begin and that's when you get the match maker to pop for you (where you click accept and all that jazz), here is the line of code that this is referring to;

**_ # score the roster against team1 and the match

team1Score = -infinity

if canJoinTeam(roster, team1, team2, config):

team1Score = scoreRoster(roster, team1, team2, match, config)

 

# score the roster against team2 and the match

team2Score = -infinity

if canJoinTeam(roster, team2, team1, config):

team2Score = scoreRoster(roster, team2, team1, match, config)

 

# found a better roster!

if bestRoster is None or max(team1Score, team2Score) > max(bestTeam1Score, bestTeam2Score):

bestRoster = roster

bestTeam1Score = team1Score

bestTeam2Score = team2Score_**

 

**_ # could not find any roster for the match, abort

if bestRoster is None:

return False

 

# add this player to whichever team is a better fit

match.append(bestRoster)

if bestTeam1Score > bestTeam2Score:

team1.append(bestRoster)

else

team2.append(bestRoster)

 

queue.remove(team1)

queue.remove(team2)

 

createMatch(team1, team2)_**

 

_____

Edit: Really sorry about the formatting but it's really bad...not really sure why it looks like the way it does but ehh...

 

[Megametzler.5729]

> @"JusticeRetroHunter.7684" said:

> (...)

 

Very much appreciate your effort! :smile:

 

Still I am not really convinced. So while I am not a programmer, I also tried to have a look. I think the most important lines are:

 

> # adjust score by roster size difference

> distance = abs(len(roster) - otherTeam.maxRosterSize)

> score += distance * config.rosterSize.distance

 

config.rosterSize.distance is defined above as "-100":

 

> RosterSize max-diff="3" distance="-100" perfect-fit="0"/

 

So the team's score might actually get changed. However, "abs" should always give a positive value, so the value "score" always gets lower. So, what exactly is "len(roster)"? Is it the queueing player's roster (1 or 2 for duo)?

 

So: It might actually be included, I admit that. But: Is there a mechanic preventing 2 duoQs on the other team? "otherTeam.maxRosterSize" would be 2 even if there were 2 duos. So no matter if you have 5 randoms and play against one duo or 5 randoms and playing against 2 duos, the score will be changed in the same way. (Correct me here if I am wrong please!) Same goes for allied duoQs - where are they accounted for?

 

Another thing that could prevent this the "max-diff="3"" mentioned above with:

 

> Scoring/RosterSize/@max-diff The maximum allowed roster size difference between teams.

 

Is this "roster size difference" just redundant with this:

 

> RosterSize min="1" max="2"/

 

since you can only queue as a duo, therefor the maximum difference to any other roster is... 1 at max?

 

I find interpreting code extremely difficult, I admit. There might be something adjusting score if you soloQ and play against a duoQ indeed. But are duoQ on your team accounted for? Is the number of enemy duoQs accounted for? I still don't know.

 

I am also still unsure what OP was trying to point out, but this topic is interesting anyways. :smile:

[Megametzler.5729]

> @"DanAlcedo.3281" said:

> He has 17 kill and not even the most kills on his team.

>

> Looks like Red farmed Blue but they didn’t cap points.

>

> Or something.

 

I make another post for this, because it is interesting, but a completely different topic.

 

Yeah, since you pointed it out: He has 23.9% of his team's kills **and 17.1% of the game's kills**. Does this mean, the enemy team won while killing much much less enemies (the deaths statistic shows something similarish)? Were they bunkered to death, hardcore decapped and outrotated or... I don't know. Why doesn't he explain? :weary:

[DanAlcedo.3281]

> @"Megametzler.5729" said:

> > @"DanAlcedo.3281" said:

> > He has 17 kill and not even the most kills on his team.

> >

> > Looks like Red farmed Blue but they didn’t cap points.

> >

> > Or something.

>

> I make another post for this, because it is interesting, but a completely different topic.

>

> Yeah, since you pointed it out: He has 23.9% of his team's kills **and 17.1% of the game's kills**. Does this mean, the enemy team won while killing much much less enemies (the deaths statistic shows something similarish)? Were they bunkered to death, hardcore decapped and outrotated or... I don't know. Why doesn't he explain? :weary:

 

This post is like one of these "you cant unsee it, IF you can see it" pictures.

 

I look at it for minutes but cant figure out wtf is so special about it.

[Megametzler.5729]

> @"DanAlcedo.3281" said:

> > @"Megametzler.5729" said:

> > > @"DanAlcedo.3281" said:

> > > He has 17 kill and not even the most kills on his team.

> > >

> > > Looks like Red farmed Blue but they didn’t cap points.

> > >

> > > Or something.

> >

> > I make another post for this, because it is interesting, but a completely different topic.

> >

> > Yeah, since you pointed it out: He has 23.9% of his team's kills **and 17.1% of the game's kills**. Does this mean, the enemy team won while killing much much less enemies (the deaths statistic shows something similarish)? Were they bunkered to death, hardcore decapped and outrotated or... I don't know. Why doesn't he explain? :weary:

>

> This post is like one of these "you cant unsee it, IF you can see it" pictures.

>

> I look at it for minutes but cant figure out kitten is so special about it.

 

Ah, hold on, there might be another reason for this.^^

 

23.9% of his team's kills with 17 kills means: His team killed about 71 people. As it is 17.1% of the game's kills, there were probably 99 kills in total, so the enemy team had only 28 kills in total!

 

Now this _might_ mean they got outrotated, outbunkered on some points or stuff. It could also mean that his team _always ran together_. If they ran together with, say, 4 players and killed a total of 18, he wouldn't have got max kills (scrapper did), but the total kill count would actually be 18+17+(15+15 or so) = 65 already, leaving only 6 kills for duels of the last player.

 

The enemy team might have won the duels on the side nodes consistently, they might have killed 28 players, if all their kills were in duels, which is of course an extreme assumption. But then they might have actually more kills (28) than the losing team (18+6).

 

Okay, actually, those statistics are extremely hard to interpret. You would have to include the death ratio for more data. But I am too lazy (and tired) to do that now.^^

 

€: Wait a minute! There is still "Team vs. enemy team" which is 71.7%! Now we can calculate the actual kills: 17/71.7*100 = 24 - which is actually exactly what I calculated above, 24. xD Yeah, they were probably zerging around without proper map-movement.^^

[JusticeRetroHunter.7684]

> @"Megametzler.5729" said:

> > @"JusticeRetroHunter.7684" said:

> > (...)

>

> Very much appreciate your effort! :smile:

>

> Still I am not really convinced. So while I am not a programmer, I also tried to have a look. I think the most important lines are:

>

> > # adjust score by roster size difference

> > distance = abs(len(roster) - otherTeam.maxRosterSize)

> > score += distance * config.rosterSize.distance

>

> config.rosterSize.distance is defined above as "-100":

>

> > RosterSize max-diff="3" distance="-100" perfect-fit="0"/

>

> So the team's score might actually get changed. However, "abs" should always give a positive value, so the value "score" always gets lower. So, what exactly is "len(roster)"? Is it the queueing player's roster (1 or 2 for duo)?

>

> So: It might actually be included, I admit that. But: Is there a mechanic preventing 2 duoQs on the other team? "otherTeam.maxRosterSize" would be 2 even if there were 2 duos. So no matter if you have 5 randoms and play against one duo or 5 randoms and playing against 2 duos, the score will be changed in the same way. (Correct me here if I am wrong please!) Same goes for allied duoQs - where are they accounted for?

>

> Another thing that could prevent this the "max-diff="3"" mentioned above with:

>

> > Scoring/RosterSize/@max-diff The maximum allowed roster size difference between teams.

>

> Is this "roster size difference" just redundant with this:

>

> > RosterSize min="1" max="2"/

>

> since you can only queue as a duo, therefor the maximum difference to any other roster is... 1 at max?

>

> I find interpreting code extremely difficult, I admit. There might be something adjusting score if you soloQ and play against a duoQ indeed. But are duoQ on your team accounted for? Is the number of enemy duoQs accounted for? I still don't know.

>

> I am also still unsure what OP was trying to point out, but this topic is interesting anyways. :smile:

 

Well i think that's the big question...how much do these scoring differences effect how the matchmaker finds the pairings okay? Perhaps a math wizard who ran a simulation of a match using the code would be able to find out. What we do know is that the matchmaker will try to find the "best" possible matchup by throwing away matchups if it's able to find one that's better (which probably means that another grouping of teams would cause a lower score differential than the previous). In my eyes, the Matchmaker hinges on it's ability to have a pool of players to select from in the first place...this is the good'ol population size issue. If there are only 20 players in a que, the matchmaker will do it's best to find a grouping of those players to set up a game. With only so many variables at it's disposal to try and balance said games to the singular value...well it gets harder and harder for it to create good matches, because it will have less options to throw away bad roster groupings (Scores are also effected by how long you wait in the que as well.)

 

It's a real predicament.

[Tycura.1982]

> @"SoulSin.5682" said:

> The game is really harsh for duoQs. Or perhaps I am just cursed.

 

Well I don't think you're crazy.