Why we need seperate ELO for each map type - Open, Closed, Hybrid, Water,

There has been divide in the community around queue dodging, and while that behavior can be looked at as selfish, I can understand and empathize with some reasons for it.

For example - let’s say someone mostly plays Arabia, as a typical example. One day the matchmaking places him on Islands or Arena. He has never played those maps before, and thus placed at a great disadvantage at his current skill level.

The solution? We have a separate ELO for each map type, which combines to the player average.

Open - Arabia, Valley, etc
Closed - Arena, Hideout
Hybrid - Scandinavia, Four Lakes
Water - Islands

Wins give a weighted value. For example, if you win on Islands your ELO for the Hybrid map type also goes up slightly.

This makes it less intimidating for people to play on maps they are not familiar with, and makes the ranked ladder more friendly. With such a steep learning curve, the game also becomes more accessible to newer players.

Thanks for reading,


What’s next, separate Elo for each civ as well?
Wouldn’t it be better to propose an Elo for each map for each civilization?
We would have 39 different Elos for Arabia, 39 Elos for Arena, etc.


No, that’s silly.

Matchmaking finds a player of similar ELO first, then you get to choose the civilization

So that wouldn’t work

I made a similar proposal, but I used different categories, like agressive, versatile and restrictive.

I think the separation open/closed is irritating as there are “open” maps that are agressive and those who are very versatile like arabia/runestones.

This false categorisation always lead to the false narrative that arabia should be an “agressive” map, but the “versatile” catagorization fits it way better and is also the explanation why it is so popular and many “agressive” open maps like socotra arent because they are just not balanced in the way arabia is.

But I am in general for this method, I think it would make matchmaking much easier. WIthout the need to punish anybody because he just doesn’t likes most of the maps.

I know that it is
But your argument can be use the same way for a civ

  • let’s say someone plays with Franks, as a typical example. One day the he plays as Portuguese. He has never played that civ before, and thus placed at a great disadvantage at his current skill level.

Just saying

Who will say which map falls into the category of aggressive or versatile? open or closed?
The best approach to this is ranked lobbies.

The best way to use the Elo as intended is to play any map with any civilization (which is not happening at the moment but is ideal)

Yeah these seem really tough to categorize. Hideout for example is walled enough to prevent Dark Age laming, but it doesn’t keep you safe from Feudal units like Arena does. Is Highland open or closed (it has one choke point to wall like Black Forest or Amazon Tunnel, but you can easily transport across the river). Is Acropolis a hybrid map because it has a lake in the middle, or is 4 lakes the minimum water for hybrid classification?

I don’t see any subdividing attempt succeeding, it would just cause more confusion.

1 Like

I agree that one ELO per map would be ideal, if it is easy to implement in the backend. I don’t know how the queuing system work, but if it is possible to easily match players using one ELO per map, it would be a great feature.

Actually, we do not necessarily need to categorize. We could watch the ELO of every players over every map and find correlations between the maps and the ELO to figure out which map give similar ELOs to the players.

I woulnd mind having different ELOs for different maps, as for me it is natural that I can be good at playing certain maps and bad with other maps. I don’t know if it is confusing to other players to have several ELOs and thus not being able to compare themselves.

I think the tricky thing there is that maps are not considered or selected until after two players have been matched. If their Elo is close then they get paired up, followed by looking a map bans and favorites. At least that’s my understanding of the matchmaking system.

While I don’t see anything wrong with the suggestion, we have massively more glaring problems at the moment that affect Elo, to name only the most visible ones:

  • One exploit to create new accounts at 1000 Elo (smurfs) in ~5 minutes
  • A sizeable part of the playerbase who alt-f4’s when they get maps they don’t like, temporarily losing their own Elo and artificially inflating the Elo of non-popular map players.
  • Multiple exploits to achieve artificially higher/lower ELO, including some that are impossible by conventional means (5000 Elo currently)

As a result, to me it should be common knowledge that TG Elo is already meaningless and probably 1v1 midrange Elo as well. We are now 2 years after release, and these issues have always been there, so, they won’t magically get patched soon. Separate map-specific Elo is not a bad idea, but it’s really just a nice-to-have at this point, I don’t see why devs should allocate any time implementing it when we have such massive flaws in the game (and here, I’m only talking about Elo problems itself, not the other very contentious issues of the game).

Btw, unlimited map bans and map-specific ELO are completely unrelated features, both can be implemented independently, and actually the current system is already affected by the “artifically inflated Elo” issue you describe, because 1-2 maps take 70% of the total pickrate (Arabia/Arena), so a player can artificially increase their Elo above their true skill by banning these maps only, and practicing the unconventional build orders more than their opponents.

1 Like

I honestly do not see map specific elo to be solving the Alt+F4 problems - e.g. what if I don’t want to play valley and only arabia, tho both of them are grouped as open map?

Generally I think it is a very nice-to-have feature