A Proposal for Unifying the Quick and Ranked Match Queues

Age of Empires IV IV - Discussion
#1

Objective

Reduce wait times and increase player matching variety by unifying the Quick and Ranked Match queues.

Background

AoE4 is one of the most played RTS games on Steam. However, this doesn’t mean there are a numerically large number of players (recent data suggests 8-9K players, see Age of Empires IV - Steam Charts). As a result, matching can take time (particularly for high ELO players) and frequent players may end up matching with the same player.

Part of the problem is that while the Ranked queue has been a benefit to the game, it has split the playerbase between Ranked and Quick matches. Based on data from AoE4 World, there are as many as 40k unique players for Ranked matches in the last month or so since patch 23349 (Ranked Match 1v1 – Ladder – AoE4 World). Quick matches have around half that number, with around 20k (Quick Match 1v1 – Ladder – AoE4 World). This level of split has likely had a significant impact on match finding.

Proposal: Keep Quick and Ranked Matches, but Unify their Queues

There are two fundamental changes required:

  1. Unify the Quick and Ranked queues to use the same matching pool and ELO system.
  2. Give each player two separate identities on the new unified Ranked ladder. One for Quick Match play and one for Ranked Match play.

While I’m not privy to the details of how the Ranked and Quick matching systems work under the hood, they both fundamentally work off of an ELO-like system. If they aren’t using the exact same algorithm, they could be shifted to do so. Then you need to point them to the same database of ladder player identities. Having two separate identities is necessary to allow for a different playstyle (e.g. trying out new strategies or playing casually) on Quick Match vs Ranked Match. Otherwise the different play style could negatively impact the player’s ELO.

Pros:

  • Unifies the two queues with limited negative impact to player experience.
  • Algorithmically simple, no complicated logic for matching between two pools and managing the zero-sum ELO between them.

Cons:

  • Depending on how the Ranked ladder database is structured, this could require a significant migration. For example, if the primary key is tied to the user’s game account rather than a separate internal ID (e.g. UUID V4), a migration will likely be required.
  • Players on the Ranked ladder may be less casual than Quick Match, leading to mismatched intensity levels (and a potentially negative player experience).

Note: after typing this up, I found out that someone had already mentioned (https://www.reddit.com/r/aoe4/comments/u7zvb1/please_combine_quickmatch_and_ranked_queues_for) the idea previously. Apparently this is how SC2 handles the queues.

The Algorithm

The only change I would suggest is a bias towards matching Quick Match players with each other (same for Ranked Match) if they’re available. This restriction can be removed if no match is found after a period of time.

Backfilling

The Quick Match player pool can simply be merged into the Ranked Match player pool. It already has a zero-sum ELO distribution, so it shouldn’t throw off the distribution in the Ranked pool. The Ranked account stays the same, the Quick Match account gets turned into a Ranked account with the same ELO it had in Quick Match.

Pros:

  • Simple.

Cons:

  • You may have some Quick Match players that end up at high levels in the Ranked ladder. Some may find this unfair due to the associated rewards.
  • There will likely be a readjustment period where the ELOs have to resettle. Assume for the moment that people who play Ranked are more serious about the game. This means that people on Quick Match might have a higher ELO than they would on the Ranked ladder. Initially when merging the pools this could result in some imbalanced matchups. However, it would stabilize fairly quickly and internally the AOE 4 team could make your ELO more flexible for the first few matches after merging the pools.

How to Handle Map Selection with Cross-Pool Matches

If a Ranked and Quick Match player play a game, you will need to select a map based on one or the other map pool (or both). Probably the fairest option here (while maintaining the Ranked map pool) is to only select a map from the Ranked pool but weight the preferences of the Quick Match player more if their preferred maps are not available.

Potential Risks

  1. One risk is people using their Quick Match identity to manipulate other ELOs. Random matching ensures there’s no way to do targeted ELO transfers. The one viable option is that if enough players intentionally tank their Quick Match ELO, it could potentially boost the overall ELO of Ranked Match players. I think this level of coordination and dedication is unlikely, but the system could punish significant deviations in win rate between Ranked and Quick Match play.
  2. What impact would ELO transfer from Ranked players to Quick Match players or vice versa have on the overall distribution / player experience?
    2.1 Imagine for the moment if every pairing for a week between a Ranked and a Quick Match player resulted in the Quick Match player winning. This would skew the ELO points distribution for both Quick Match and Ranked players. However, it should not negatively impact the player experience because the skill levels between the two ladders should stabilize, resulting in more equal matchups between players overall.

Alternatives Considered

Allow Players to Play Without Affecting their ELO

This would be an option to play games that wouldn’t affect your ELO. This isn’t viable because it’s too easy to abuse. Players could use this to intentionally target below their skill level.

Removing Quick Match Entirely

This is a fairly simple option with the downside that it removes casual play options outside of open games. Open games have no matching system, which leads to inferior player experiences.

3 Likes
#2

Yeah this would be nice. It would probably help reduce wait times at higher ELOs as well.

2 Likes
#3

Interesting idea.

Side note, people do need to remember that Steam is not where all AoE’s players come from. I would imagine a large number of players play on Xbox PC Game Pass, I do and my 4 friends that play do as well. I actually don’t know anyone who plays on Steam because everyone I know has Xbox Game Pass Ultimate. :wink:

#4

I appreciate the goal of this idea.

How would it work with high ranked players using quickmatch to target similar players playing ranked?

It seems that at a certain point, players do not want to hurt their rank, so they will use quickmatch to strengthen their position on the ladder (try to beat players around their rank using quickmatch instead). This would be comparable to using a smurf account.

If possible, it seems that making the ranked ladder the most attractive option would be best. Not easy to solve.

1 Like
#5

Would this apply to TG’s too? If so then no it would be stupid idea to have quickmatch and upcoming ranked queues mixed up. In 1v1 there is no similar problems as TG’s.

Many ppl who play TG QM are ppl who are not considerate of their rank or elo and don’t care about it. Such players are nuisance to players who try their hardest to rank up those it will result problems.

I like alternative of not having separate QM or Ranked at all much more simpler, but this also brings similar problems in TG’s. For 1v1 it wouldn’t change things. Just have season rewards like now and if you dont care about your rank then play the way you desire and those who want to rank up can rely on their skill

#6

A lot of difficulties faced with the proposal for both 1v1 and team games. Also does not seem practical to have one rule for one and the other.

Maybe you could check a box that you are willing to play a quickmatch game after x minutes if you have not found a game in ranked but on the separate ladders.

Maybe the new expansion will just breathe life into the competitive modes, here is hoping.

Again merging the two ladders seems to be preferable. However I also like to queue for some quickmatch 1v1 after a few high stake ranked games. If you merge the ladders then you remove that lower stake option to play. This might result in more players stopping once they reach a certain rank and/or using a smurf with no other matchmaking outlet.

I can’t see combining the queues working but I will try and learn about how SC2 dealt with this so I am more informed.

In respect of team games, my view is you are not really being competitive if you are not in a premade team. It is like chopping one of your legs off before a race and hoping you get given a good replacement (team mate) in time for the start of the game. Such random factors are gambling rather than competiting on a more level playing field. Such lends itself well to players making more excuses.

Splitting up team game player base into premade and non premade also seems unlikely. Changing how teams are made in line with ratings seems the most effective change in the short term. Again you could maybe check a box that you are willing to wait longer to match at more even skills.

#7

This here is not remotely true. Its completely plausible to be competitive in TG’s without premade in terms of mindset and willingness to coordinate. Look at other competitive games like mobas, fps etc. Plenty of ppl play with mindset of being competitive even without premade and plenty of ppl are at high rankings without playing with premade.

The issue here is in AOE4 how much it favors premade over solo. If you play with premade and you are decently skilled player its instantly free 200-500 elo on your account vs alternative. Then the fact that premades can abuse matchmaking to lower the elo of team to levels where they gain more than they should and lose less than they should by having one of the players playing with lower elo account.

There is many fundamental issues with TG’s which are not addressed and won’t be, because devs are not thinking anything before they add things to game those ranked mode will be complete joke for TG’s.

Few days I ago I went through top 100 in 2v2. I found 2 that were very likely to play as solo at above 1800 elo and 2 more that were also likely, but less likely than previous 2. So 4 players in total in top100.

#8

Yes of course the system needs changing as it can be abused. I too would like to improve team games.

When I queue up without a premade team, I know that either team could be at a disadvantage. Therefore it is not as competitive as two of the best premade teams going up against each other for the top spots on the ladder.

I never said it was not competitive at all to queue up with randoms, you have inferred that. I said it was not really in my view as I’m of the opinion that you compete the best you can or not at all (meaning I accept I don’t have much time to play anymore so I do my best and good luck to those playing at the highest levels). Just as you have a self imposed way of playing team games to supposedly challenge yourself, I can respect that.

The nature of being competitive is to want to win or be more successful than other people. You can be competitive against yourself but this game requires you to play with others in team games. Giving yourself a handicap (by maybe or maybe not matching with a strong random team) to me means that you don’t want to win as much as others doing everything they can to win (like being in a pre made team). Of course you will strongly oppose this because of your philosophy to solo queuing.

Hopefully some good solutions are forthcoming but some of the new maps seem horrible which will certainly not help team games succeed (including trying to force players to play more water).

#9

There is many issues with TG’s rn but on this topic main problem isn’t exactly if you play or not with premade. Premade makes things considerable amount of easier but its not just because you play with premade its because how unfairly matchmaking works. It considers average team elo = balanced matchmaking which is never the case.

Your team can be as strong as its weakest link. So its more of which side gets weaker player than stronger. Strong can obviously compensate some, but if your ally is dead set being useless or leaves immediately when something minor goes wrong it just creates extremely unfavorable situation.

Lets take example from game I had today in 2v2. I throw the elo numbers from my head cuz too lazy to double check but it was me 1.7k with 1.1k vs 1k with 1.4k. Now I would obviously be very favorable to win this game, but getting matched up with someone with 1.1k is pure coinflip what ever it works or not. In this case I should’ve been able to carry the elements were there, but I made one critical mistake which was going infantry over cavalry another issue was that I wasn’t 100% focused on the game due I was doing something else IRL.

Now the major issue was my ally. He reached castle age at 30min mark while everyone else was imperial. Everyone was castle way before him and he had taken 0 dmg also he played as french so our team had complete map control. Also because he went 1v1 unit comp knight + archers which is super bad option to go in this situation. Also he lacked control over units and multitasking was none existing. Even tho I was able to defend against 2v1 and fight back, moment english hit imperial before me it was over. Especially when french was still in feudal.

Needless to say its complete coinflip if you happen to get ally that has somewhat decent understanding and so on. If you lose that coinflip you’re in massive disadvantage and anyone below 1.5k elo is huge coinflip which is vast majority of matches that I play.

I might get 900 elo ally that does several times more army or right things at right time because of luck and win a game then in other spectrum I get ally who is 900 elo and does nothing. I had ppl at 900 elo who are placing mining camp 5-10 tiles away from gold / stone just to gather both of them inefficiently.

Same was my last game where my english ally went for LB pressure against another english who went for dock into FC and my English ally was in feudal. At this point he tried to pressure and miserably lost and did nothing. Lost the game due I made single mistake which was didn’t palisade 3 tiles which I already had queued but during multitasking forgot I put gate to another part of wall which cancelled the build order command to those 3 tiles. English got 6 knights into my food eco after this because not having that single wall build in which slowed me down too much so I wasn’t able to carry.

So you’re right solos shooting themselves in leg while playing with random allies, but the issue is matchmaking and how it works. Matchmaking should never consider average elo of team good way to put matches together. Its not accurate and the fact that it increases search range of lobby after x minutes. So if I want decent matches which are not complete stomp on either side I have to cancel search every 1 minute and continue doing this like 15-20 times to find match even then the chances are that I get into lobby with low elo ally against high elo elo abusing premade is extremely high.

#10

Your response to me focused on the premade aspect of my post.

Subject to the amount of players actually playing in a game mode, ideally players higher up the ladder should never be playing with extremely low rated players. I guess this is a reflection of the number of players and they must be aware that there are many uneven matchups going on.

Agreed average ELO matchmaking does not appear to be working very well and I guess it depends if people want to wait longer for a quality game or not. As I expect if you narrow the criteria for matchmaking, it will take much longer to get a game and it could make players play that game mode even less. On the flip side, the current matchmaking probably does not inspire weaker players to continue searching where they find themselves in a mismatched game.

One of the main reasons I stick more with 1v1 now. 2v2 can be bad but 4v4 is on another planet by comparison. Gamble city. However that is one thing I think is good about the new ranked is that you can rise in say 2v2 to conqueror then be matched with equivalent players in say 4v4 was my understanding?! At the end of the day you can always risk being matched up with someone who could just be having a bad day or is drinking heavily haha.

I guess the question of keeping players playing in a single game mode and keeping them happy from more casual to competitive is a very difficult one.

In any event very excited to see how it plays out on 25 October.

#11

If this is the case its one minor step into right direction. I would love if they referenced all modes when searching the match. While 1v1 and TG’s are different type of games and skill doesn’t translate 1:1 to either of them its still better have 1v1 conqueror in team than TG 1500 elo player with platinum rank in 1v1 especioally if they’re reached that elo with premade.

The thing about matchmaking is that players can freely manipulate it because average team elo system and no limitations who you can play with. Many 2k elo 2v2 players play with someone who has sub 1.5k or close to 1k. This ensures slightly faster search time but also less points lost and more points gained than playing with someone who has 2k elo.

This could be easily fixed by having premades to form a team/clan what ever you want and the teams elo is considered when matching. So they cant take low elo account just to boost their own elo. So if team has 2k elo then they got 2k elo even if one player is actually 0 elo player and if you play solo then you don’t have to form team.

Ofc there should be limitations how often team can be created or how often you can leave the team to stop players abusing the system.

Also what comes to queue times. My average wait time in queue is 5mins and im +1700 rn. I personally can wait even 10mins to get decent match because I got other things to do while Im in queue, but what I cant be arsed to do is stop search every 1-2mins and put myself queue again. Just too annoying. If there was box that would be “stricter matchmaking” I would use it all the time, even if my queue times are longer.

The TG ranked will be joke. Its going to split already small player base more into smaller groups and those making problem more transparent. Already did some math. This is from couple of days ago so number may be different rn.

Only numbers available for us is steamcharts and avg is 7k and peak is 15k. If we split those numbers equally during peak hours it means there is 3750 players playing each mode at same time which is 1v1 = 1875 matches, 2v2 = 937.5 matches, 3v3 = 625 matches, 4v4 468,75 matches. So once we move away from peak hours we might have 50% less games played which makes things even worse for split queue.

And once we split this into ranked TG’s, ranked 1v1 and QM 1v1 and TG’s we get playerbase thats so split that the matchmaking becomes worse.

#12

“Your Solo rank is completely distinct from your Team rank. That means you’ll need to complete your 5 placement matches for each queue. Your Team rank is a combination of 2v2, 3v3, and 4v4 – queueing into any of these ranked modes will affect your Team rank Points! And just like unranked, you can queue in with a party of friends, or allow matchmaking to find teammates.”

Yes this could at least match players up more evenly assuming it sticks to a narrower set of ranks or even down the line allows you to focus the search more.

To be fair season 2 1v1 greatly improved on the era of inflation in season 1 so there is hope.

#13

I’m slow? But it sounds like you want ppl to que up for rank/QS simultaneously?? And potentially end up an either game??? BUT… there is a reason the QS pool/population is abysmal. (at least in 1 v 1)

  1. PPL do NOTTTT want to play all maps that are in QS and you can’t possibly veto all the non competitive non-fun (popular opinion) maps
  2. There is no reward associated with QS
  3. As you mentioned above some ppl play more earnestly in rank as oppose to QS and likewise anticipates reciprocal play.

MAYBE I’M JUST SLOW AND MISUNDERSTOOD???

#14

It’s definitely true that Game Pass likely makes up a large portion of the player base. Unfortunately, we don’t have daily active user counts for Game Pass. The stats on unique players over time will have to be a supplement.

It would essentially be a free smurf account for every player (except tied to your identity). I’m not fully sure I understand what your concern is here. They could try to beat players at the level of their Ranked identity with their Quick Match identity, but I don’t really see the benefit. They would have to grind all the way up to the top and then play just as intensely as if they were ranked. For the other player, it would make no difference if they were playing against someone on their Quick Match or Ranked identity (depending on the implementation, they may not even be able to functionally tell the difference). The only think I could see is that maybe players would feel less pressure if they’re not playing on their Ranked account, but I don’t see how that would help a highly rated player so much. I would think they’re used to it.

Let me know if I misunderstood you.

I don’t have a strong opinion on whether this should apply to team games. I would think we could try it out with 1v1 first and see how it helps queuing times.

Unfortunately forcing people to tank their ELO if they want to play casually is a really strong downside here. I believe totally deprecating QM would be a less appealing choice to the player base.

This is definitely an option.

This would functionally give everyone a smurf account on the Ranked ladder. Your QM plays would be a separate identity in the same player pool.

Y’all both make some good points, but I’m not familiar enough with TGs to say how this proposal might affect them. My preference would still be to try this change with 1v1s and then potentially figure out how it can work with TGs once we have data.

This is addressed in the proposal under “How to Handle Map Selection with Cross-Pool Matches”. Basically, if a QM player matches with a RM player, they would use the Ranked map pool.

You wouldn’t be queuing up for both, you would be queuing up in a single pool that holds both QM and RM. This would help unify the player base to reduce queuing times.

Correct, and this would not change. While you would functionally have two identities on the Ranked ladder, only your RM identity would earn awards.

The idea here is that people who play QM casually would eventually find themselves at a lower ELO in the ladder that contains both Quick and Ranked players. They would naturally find themselves with less intense players.

No worries, I think the key understanding is that we’re unifying the queues, not queuing up for both. Then we give people two separate identities in the unified queues, one for QM, one for RM.

#15

After thinking this bit more I think best option would simply remove quickmatch and have only ranked mode or otherway around. It doesn’t matter which one is removed.

And just simply place ppl to right leagues according elo. If it matters that much to some ppl add toggle to turn off display of your rank.

For casuals nothing will change other than they will get seasonal rewards even if they were originally afraid of playing ranked. If you don’t care about the “ranked” feature then you just keep playing and play as you did with QM and those who try harder will get ahead of them.

At first it might bring bit difference between mindsets of QM / Ranked players but this will eventually even out and places players to right elo spectrum

There is also one very strong positive point to this. There is many ppl who are afraid of playing ranked. Which is quite stupid. Player who is afraid of playing ranked won’t lose anything valuable. There is nothing on the line and having 1 mode will eventually remove this fear from players.

This would also work with TG’s as long as they tweak how matchmaking works and do adjustments outside of matchmaking.

For example if they force premades to make teams/clans that they cant abandon without penalty will remove elo abusing and this would be only way to play TG’s with premade. Like I mentioned earlier (maybe?) lobbies are created based on average elo of teams. This current system is so ###### ###### and doesn’t work at all and lets players manipulate system to gain more points and lose less.

For example many 2k elo players teamup with someone who is 800-1,2k elo range to tank their teams average elo to lower this ensures that matchmaking matches them frequently with player(s) with 800-1.2k elo while one of the opponents might have considerably higher elo. If premades were forced to make a team which has 2k elo then all players on that team would’ve 2k elo and which automatically searches much stronger opponents and in case they get matched with lower elo players they got huge risk of losing more points and won’t gain anything.

Ofc leaving or abandoning the team should’ve cooldown and penalty so that players cant just re create teams every day or switch between teams constantly and obviously cap of how many players can be on team. For example 4 players in 2v2 team.

This might also bring some other problems to ppl, but it would at least create more fair environment between solos and premades and stop elo abusing which is crucial for ranked TG’s, because otherwise the ladder is complete shitfest.

Stricter matchmaking would also need to be put in place. As of currently matchmaking can throw anyone in the lobby. I have played with 200 elo player while myself being +1600. Everyday few of my games I got anywhere from 0-1k elo rated player in my match which is not fair situation at all to anyone and this should never happen especially when player(s) are just so far ahead in terms of skill

#17

Ty for the clarification but after considering your truest implications it would make QS only perk being you could play at your respective level with rank impunity; so ppl trying new stuff out and casual play will be incentivized.

1 Like
#18

Suggested by a number of people, also on Reddit. It’s a good idea. Works the same in starcraft. Priority on matching against players in same queue, only shifts queue if time exceeded. Keep TG ranked and QM separate until further study due to potential clash in team playstyle. Possibly lower everyone’s QM elo by 100 so when they face ranked players it isn’t as hard or use an elo gate between queues.

#19

I think you’re not considering the folks that play both QM and RM. I know of folks that will play some Ranked, but switch over to QM so they don’t have to stress as much. It’s also useful for people who want to try out odd strategies that they wouldn’t if they’re trying to achieve some particular ELO. High ranked players do this all the time when they try out new strategies, but setting up smurf accounts is unreasonable for most people.

I guess I’m not clear on the downside of this approach vs getting rid of QM entirely. The difference is only the amount of work required from the devs, which, while a reasonable concern, doesn’t outweigh a better player experience (IMO).

That said, I would rather get rid of QM than keep the pools separate. So, if this proposal weren’t viable, I would be in agreement with you.

Yup, that’s the idea.

Yeah, I noticed an old Reddit post (linked in my OP) about this after typing up my proposal…a little disappointing to realize the devs are already probably aware of this style of matching. I’m hoping this brings more attention to the idea.

1 Like
#20

Do devs read all this stuff? Very long and confusing thread

#21

While you’re right about this that, but the in AOE4 case probably biggest problem for player to try something new in ranked is being afraid of losing the rank they currently hold if its the highest they have gone, because the rewards are tied to your existing rank and not the highest achieved in season. If this was changed then barrier to do something new or weird would be much lower in ranked.

Also best place to test something new / weird etc is actually ranked especially if you’re higher elo player. It gives you much greater indicator whatever it can work or not. Lets say you go into QM and try things out and get good results then you go to ranked and get horrible results.

Also almost always simplest system is the best, because there is very little that can go wrong. If you keep piling system after another then there is much greater chance of someone finding way to exploit it.

For example if we had unified queue for ranked and QM someone might dodge all the ranked players trying to find QM players to ensure higher chance of winning also with this system you would be switching elo points between QM / Ranked players. For me I can’t find reason how this could be exploitable, but something could rise up after experiencing it those getting rid of one mode is simplest solution.

Also how would this work with TG’s is another issue. If you got a friend that wants to play QM only and one that wants to play ranked. This might also open possibility of abusing the system by having one player ticking QM while others chose ranked. This could result facing players only from QM side which could be much easier than facing team full of ranked players. So essentially this should result another system to be placed for TG’s which disables ability to play with mixed QM/ranked players on same team. And I can easily prove how TG matchmaking is one of the ######### I have ever seen as of rn.

image

My last game. This should never happen in any situation no matter how long I had to be in queue but for this I was in queue for like 5 mins which is completely standard for me and btw im the 1746 elo player. Game was dull, boring, annoying and not worth of my time, but if I left then I would take -25-30 point hit and had to play 5-10 games to recover from it.

1 Like