Teamgame queue might not survive this summer.

[Clemensor]

At this moment it seems clear that with the new calculation of ratings, there shouldve been a ladder reset as well.

 

[AttilaTheHun]

Team ratings are still unreiable and i have no idea how the matching system works...

In the game in which i queue solo i have seen up to 500 elo difference between my allies and the average elo of my team and the enemies is sometimes also 200-300 elo. These things make no sense to me. Sometimes it looks like all 4 best players (based on elo) are playing for the same team. The devs really have to take care of these issues.

 

[doacid]

and i have no idea how the matching system works...

Ask "thx DE", current top 1.
It seems he worked it out

 

[z3us]

Ask "thx DE", current top 1.
It seems he worked it out

I think he got banned. According to some posts in Arena discord.

 

[drunkenhalb]

Yeah I agree about TG not surviving. But prob for different reasons. People are going out a whole lot more. They are sick of being indoors. Video games, streaming, and wtv hobby was picked up during the pandemic are being replaced. And yeah, the ALT F4’s do not help. The DEVS should have just implemented a single maps Q with a random map Q since day 1. Instead of throwing everyone into a mix bag of maps. The devs essentially split the community into those who prefer single map Q and those who prefer random map Q...As a result, WTV solution they come up with will alienate one group of people, thereby causing some to leave the game. Making matters worse, inaction alienates everyone.

 

[drunkenhalb]

Great solution:D:D:D

Just go play somewhere else and keep the bad system while TG ladder is dying:D:D

Exactly. What he fails to see is that a healthy TG system keeps the game alive. People lose interest in tg, 60 percent of the online player base is gone and DE becomes just like HD.

 

[Helichaos]

Biggest problem at my elo (2k6 tg) is AltF4.

The majority of players are content to play most maps, a minority are ruining the experience for the rest of us.

Once that is addressed, I would like to see position picking removed and random civ being chosen if 50% or more players have it enabled.

But even without this I would be fine, just time out the AltF4ckers already.

 

[lecracheursagacite]

I think what needs to be considered is the fact that what works for 1v1 doesn't need to work for teamgames. The type of players, the mindset setting up/starting a match and overall how those games a played to me at least are very different when comparing teamgames vs 1v1 and therefore also the way a match making or lobby should be designed.
Perhaps there are clever ways to combine this?

Yup. As I have said numerous times before matchmade TG are a joke without civ draft (or all random).

 

[Memeluke]

I think he got banned. According to some posts in Arena discord.

I wonder if they'll ban vivi and lyx for exactly the same reason next LUL

 

[[KoBHV]venivero]

Give people the same map again, if matchmaking fails. So if you are queueing and matchmaking fails (for what ever reasons 11), after you were already assigned a game/map etc. give players only this map, until they have won or lost a game on this map.
So if you were happy to see you favourite map, and someone drops you dont have to feel sad! Because you will get it next time again. AND AGAIN AND AGAIN TILL ETERNITY HAHAHA

If 2 players with different maps "selected" get togehter, they get a 50/50 chance for Chaos pit or michi with 45 min treaty.

 

[linetyReborn]

i think forçing players to play maps they don't like is just stupid .
why not allow unlimited map bans.
for me like many other players we like arabia or arabia like maps . we played those maps for years and we still enjoy those maps. so just forçing to play other maps from time to time won't change my opinion , it only results in bad matchups and lost time

 

[leppard10]

TY GMAN

 

[Tarsiz]

Okay so we know the calculation for Elo is flawed, as it awards the same number of points to every member of the team regardless of the individual elos, which leads to the exploits we have seen at the top of the ladder with people teaming up with players much lower rated than them in order to farm Elo.

Obviously that should not possible so that lead me to think about how we would want Elo to be updated after a match, in order of importance:
- The total Elo won by the winning team should be equal to the total Elo lost by the losing team, to avoid Elo inflation over the ladder.
- Players should earn and lose points proportionally to their individual Elos and the average Elo of the match, to avoid players teaming up with much lower rated mates to farm Elo indefinitely.

My first idea was to compare each player of each team to the average Elo of the opponent team, and award points individually. Essentially, take the match as 8 (or 6, or 4) separate 1v1s between a player and the "average of the opponent team". This means that players will earn more or less points depending on how far they are from the average. Thus, someone with a much higher rating than the average will earn few and lose many points, while someone much lower than the average will earn big and lose little.

However the issue with that is that, since you're comparing Team A individually to the average of Team B, and Team B individually to the average of Team A, you're not ensuring that you have a zero-sum game, meaning Elo can be created or removed from nothing. We absolutely want to avoid that, because this would lead to similar issues as we had in the past, were people would end up earning more by simply playing a lot and have ratings that do not reflect their skill level.

In other words, it's okay to be carried by higher rated mates in a zero-sum game, as you'll lose more games as soon as you don't team up with them and end up to your real rating. If it's not a zero-sum game you'll maintain by simply playing, although it's not your true elo.

Example:

However we can refine that.

What we can do is compare the averages, i.e. viewing the match as a 1v1 between "Team A average" and "Team B average". We calculate how many points should be won or lost, and multiply that by the number of players in the team. That gives us a total amount of points (won or lost) that must be distributed among the teams. In our example above, Team A's average is 3227.5 while Team B's average is 3236.5. By winning the "1v1", Team A earns 16.5 points that amount to 16.5 x 4 = 66 points to split between the four members of the team.

This ensures we will have a zero sum game.

The current system in place splits the hoard equally between players. That's very communist, but it also doesn't work as it leads to higher rated players teaming up with noobs to artificially inflate their rating.

So, in order to split fairly, with respect to the Elos of all players, what I did was to use the previous sytem to calculate the ratio of the total number of points earned each individual player should get. I.e., you get the ratio of the points a player would theoretically earn if matching against the average Elo of the opponent team by the total amount of points theoretically earned by that team. That gives us a ratio we can then use to split the points that should be eventually won or lost.

I suck at explaining it so I made this other super intuitive table:

Essentially with the previous example we see that Team B should lose 63 points meaning 3 points are created out of the blue. We take the ratios of that 63 and apply that to the 66 points they should lose and voilà. Small difference - 3 players lose one more point - but it makes sure we have both a zero-sum game and players earning and losing points proportionally to their rate.

 

[nimanoe]

Okay so we know the calculation for Elo is flawed, as it awards the same number of points to every member of the team regardless of the individual elos, which leads to the exploits we have seen at the top of the ladder with people teaming up with players much lower rated than them in order to farm Elo.

This exploit is indeed terrible, but I'm not sure your solution is more fair. It pretty much means that the highest rated player will always get punished in matchmaking. If their team wins, they'll get the least amount of points and if their team wins, they'll lose the most amount of points.

I think the most obvious example of this is when two friends team together occasionally. They both grinded on their own and got somewhat accurate ratings. One is around 2k and the other at 3k. Now everytime they play together their ratings will converge to each other. Then when they don't play together, the lower rated player will have really tough match-ups, because they're too high rated and the higher rated player will have easy match-ups as they're rated too low.

I think an ideal scenario would be somewhat inbetween your solution and the one we currently have, so the top rated player doesn't gain as many points, but doesn't get punished as hard as in your solution. Perhaps make it so that if you have over 1000 points more than the average ELO of the team you don't get any points.

 

[Zbyszek]

Force players to play their first 10 ranked games without premade teams. Assign an according elo. If someone throws game on purpose ban that player.

 

[nimanoe]

Force players to play their first 10 ranked games without premade teams. Assign an according elo. If someone throws game on purpose ban that player.

I think this will help a lot for the accuracy of the ELOs, but it will probably hurt too much for new players trying the game. There's probably a lot of people who won't start playing teamgames if they don't have a friend to queue up with. They'll probably also have less fun without a friend making the chances they'll play more ranked games lower.

 

[IYIyTh]

There will always be problems. I only play TG's with an 1100 and a barely 1200, and am floating near 1650 1v1. I think of my TG's I might have 2-3 without those players playing with me.

My skill (w/o tryharding because it's not very fun to just insta-win with a cheese feudal rush,) only brings us to around 2300 since there is only so much one player can do without intending to end the game early. We routinely barely beat teams of 13-1500's, occasionally matchmaking will throw a 1600 and two rando's against us which is about where we should be.

Presuming I try-harded it's not unreasonable to think if I teamed a couple of fellow 1600's my actual TG rating would be more like 28-2900+ -- but since I'll never do that and only play tg's with my two buds we just aggregate out lower.

 

[A B C 1]

Elo calculation isn't the sole problem, you're just moving the problem (granted, to a place that matters less if you mostly care about a correct Top 100) if you want to punish the higher rated player. The system can't tell which member of a team over/underperformed and should gain more or less points than their teammates.

Instead teams shouldn't just be matched by the sum of their elo but the elo of higher rated players should weigh more, depending on how much higher it is, so that in the end an uneven team still has around 50% winchance.

Idk how to find the perfect factor for that but the devs should have access to enough stats that they can find it out. There is no easy solution for TG elo calculation in any game since the impact the better player makes is different in every game.

 

[Dkastro]

I think it's fine to punish the higher ELO, this way you have incentives to play with people closer to your skill/elo. That is what we want, every other game matchmakes with similar rating players... (silver league, gold, platinium, etc.)

If you still want to play with a high ELO difference team mate you can do it but you get "punished" becouse you are forcing something hard in to the system (finding a fair matchup).

 

[A B C 1]

The primary point of matchmaking is giving fair matches, if you're worried that players that only play against Top 10 000 players get into the Top 100 then just do some minor adaptions in the Elo gain calculation but that shouldn't be the main fix to the current problem.

 

[fluppets]

Premade teams should simply never get matched with random player teams.

If you team up with friends, you should be able to fill a roster of 2/3/4 players to match up with other premade teams.

Bigger games with far more players and competition have come to the same conclusion; no need to invent hot water again.

After solo teams and premade teams are separated ELO calculation can get discussed, not before.

I don't often play team-games, but I myself have alt-f4'd once: a fully random team against a full 4-man premade team in hideout, their civs were clearly well-thought out specifically for hideout. I wouldnt be surprised if one of them had created a new/low-elo account.

There's no reason I should want to play that matchup in a competitive/ranked setting.

 

[IYIyTh]

I think you overestimate the amount of people playing tg's that you believe there is capacity to differentiate full pre-made vs. partial.

 

[bobdebuilder]

One of the reasons that is leading me toward stop playing AOC (the only video game I play now and played since ever) is the TG system and am sure many others stopped playing tg on DE and moved to voobly is due to this complete crap TG system where queue goes to 8 minutes then someone from other team doesn't like his team so he alt f4 (which is understandable) because otherwise the gap level between the two teams will be huge and the tryhard pick-civ and etc... it means if you don't have a team to play with you simply can't play TG !! and many players don't like 1v1s so they would rather to stop playing the game completely than start queuing 1v1s...
this pure imbalance is making aoe2 disappear ! the game is surviving thanks to the big tournaments that are constantly keep ongoing but am sure if the situation keep being this way the game will extinct.

 

[Bladesman]

i dont want to say i agree with you bob. But i agree, there is many people exploiting the system.

 

[SouFire]

One of the reasons that is leading me toward stop playing AOC (the only video game I play now and played since ever) is the TG system and am sure many others stopped playing tg on DE and moved to voobly is due to this complete crap TG system where queue goes to 8 minutes then someone from other team doesn't like his team so he alt f4 (which is understandable) because otherwise the gap level between the two teams will be huge and the tryhard pick-civ and etc... it means if you don't have a team to play with you simply can't play TG !! and many players don't like 1v1s so they would rather to stop playing the game completely than start queuing 1v1s...
this pure imbalance is making aoe2 disappear ! the game is surviving thanks to the big tournaments that are constantly keep ongoing but am sure if the situation keep being this way the game will extinct.

The game will survive as long as the single player is playable despite how bad the multiplayer is, we can take HD as the best example of this, but the community is just leaving the game.

The TG match making didn't work, pretending it did is just ignoring all the current issues, thinking that map dodgers are the real problem, but the TG MM was never attractive to the pro players for a reason, the more they take to fix this issues from the rots the more players will leave.

Tournaments just keep like top 50 players very active and the viewers entertained and the community active with all the news and drama but they don't represent all the active community by active one should understand those who actually play the game(multiplayer) not those who watch twitch or youtube, so even with redbull here and other big tournaments around that ain't enough to keep the numbers healthy.