How good is each civ on the highest level? The top tournament(AoE3 Global Championship 2021) civ stats

country pick pick rate win win rate mirror
Russians 16 3% 11 69% 0
Hauds 22 5% 15 68% 2
Spanish 41 9% 26 63% 3
Portuguese 52 11% 29 56% 6
Swedes 47 10% 25 53% 9
Dutch 37 8% 19 51% 1
Ottomans 10 2% 5 50% 0
Indians 38 8% 19 50% 0
Aztecs 12 3% 6 50% 0
USA 6 1% 3 50% 0
Chinese 31 7% 15 48% 1
Japanese 39 8% 18 46% 1
German 21 5% 9 43% 1
British 34 7% 14 41% 2
Inca 19 4% 7 37% 0
Lakota 7 2% 2 29% 0
French 30 6% 8 27% 0

These stats were done by economist an ESOC user, this table also excludes mirrors stats but count them.

Some remarks:

Russia is the top civ on WR% but note that it was picked a lot by Haitch, the best player by a good margin nowadays, so take this with a grain of salt.

France sucks now, no matter how you see it.

Japan is very overrated by the community, it doesn’t need a nerf, just a redesign maybe.

British. I hear rumors that it will get nerfed, but it doesn’t seem OP or even strong at the highest level.

Spain, Ports and Sweden were by far the most picked civs and all of them have good WR%, so the top players think they are the strongest civs and the data supports them. I expect a nerf for those 3.

That’s it, what are your opinions on these data?


still OP in treaty, see i found a way.


Still? I heard it got significant nerfs there as well (no more instant Gendarms and Gerdarms late game stats got nerfed as well)

doesn’t really change the fact the faction has 120 military pop with full eco, a fairly strong eco, amazing skirmishers and some of the best art in the game.

only issue france has is that their anti cavalry is fairly mediocre but when you got the best combo (melee cav and anti infantry) then that doesn’t matter much.


The sample size is too small here. Russia is not OP yet has the highest win rate. Lakota bests Hauds yet this small data set wouldn’t tell you that.


The data does not allow an interpretation of which best civ is, as it doesnt show matchups some civs just lose to certain match ups…

From this table you could interpret that french are bad and aztec are good - and that is not the case.

What i think this table brings is that Sweden, Ports and Spanish - that were heavily picked won a lot.

But as stated in other threads, that may be in part influenced by the maps that were played on.


I don’t have any opinions about the Aztecs, but judging by the pick rates(which IMO are way more important than the WR% for analysis) they are not seen as a strong civ by the top players.

But France? They are in a bad state and hopefully will get a buff, f.e on the amateur championship (which is a bit weaker) they are also one of the most picked and also has the least WR% there. This already shows a pattern.

I don’t think there will be many arguments against the nerf of Sweden, Ports and Spanish. They are almost universally thought to be too strong.

It really is possible to gauge OP civs, but you need to combine win rate and pick rate.

Russia has high win rate and low pick rate → niche civ, very good against specific encounters but not really a general purpose civ.Not OP, not UP, but might need a redistribution of power between its units to achieve balanced state. There’s already lots of suggestions about that.

ports, spain and sweden: high pick rate high win rate => slightly OP but not by a large margin. Small adjustments can bring them inline.

all others: in need of buffs.
China, Japan, German, British, Inca, Lakota and French can be considered UP.


As mentioned in other posts here and on ESOC, these statistics are useless without context.

The format of the tournament skews these results so that they can’t be taken at face value. The first game in the series is predetermined by the players as a fair match-up. After that, the winner can no longer play that civ for the series and has to pick first. The other player gets to counter pick. This repeats after every game: winner can no longer play that civ, winner picks civ first in the next game, loser gets to counter pick.

Since we do not know what the win % is compared to whether the civ was picked first or if it was a counter pick, we can’t really determine anything from win rates. Civs in this game can hard counter other civs in 1vs1.

For specifically French: a LOT of pro players have this civ as a comfort civ. They probably locked in this civ first and then their opponent counter picked them with a custom strategy. Going just based on win percentage really isn’t useful here.

Based on the dataset, the pick rate is probably the most interesting. It (probably) indicates which civs the top players think will bring them the best success.

“There are three kinds of lies: lies, damned lies, and statistics.” - Mark Twain


the pick and pick rate % do not add up, probably on the account of mirrors. please add mirrors to the pick column for a more clear view, mirrors should only be left out of win rate statistics.

1 Like

Based on the dataset, the pick rate is probably the most interesting

Yeah, that’s what I said 2 replies above yours.

But, as I said before, I don’t agree with your France assessment> OFC if it was only those stats it wouldn’t mean much, but France was the most picked on the Amateur League too and also had the worst WR% there(around 30%).

@newaoeiiiai The mirror stats are on the table, it’s the last column.

I know. but they should also be added to the pick column too. as it is it can be misleading.

currently the pick columns shows that swedes were picked by players just 47 times compared to the Portuguese 52, when in reality Sweden was chosen by players 65 times compared to 64 times for the Portuguese.

1 Like

You made a lot of claims in the first post that seemed to use the statistics as evidence. I am saying that the tournament statistics are not sufficient evidence. France may indeed need a boost, but you’re using statistics that conveniently leave out information. The amateur league has the same rules and should also come with massive disclaimers. If you’d like to go through and spend the hours needed to determine what was a first pick and what was a counter pick, then I think you can use win rate as a reasonable statistic.

You did mention pick rates being more important… in a buried reply that is in response to Aztecs. In my opinion, that isn’t sufficient. Only people you responded to/people keeping an eye out for Aztecs will see that comment.

I think most people will come in here, look at the win rates, and then use that to further whatever preconceptions they already had. I’m just pointing out that is dangerous to interpret this data the way it is presented.


No matter how you twist it, France is UP.
Even if it were a comfort pick (it isn’t, given the low pick rate), the fact that it won only 27% of the picks means it really needs help against it counter picks, cause a 27% win rate is essentially something you get when you are extremely hard countered.

A decently balanced civ has between 45-55% win rate, with 55% being borderline OP and 45% borderline UP.

Also, spain ports and swedes had the most mirrors - these are comfort picks. Pros don’t comfort pick weak civs.

Anyone remember the reason for dev nerfed Lakota was they has THE HIGHEST winrate.

Who is accurate.

This could be true. I’m not arguing that France is not UP.

Then this is an invitation to go look at the French games themselves. High level statistics like this are not evidence. It is an invitation to dive into the WHY something is happening. I can easily see a lot of matches going Sweden vs Sweden, Port vs Sweden, Port vs Port, France vs Port/Spain. The draft system also matters. Not everyone could draft all 3 of the (commonly referred to as) OP civs.

Another thing with these statistics are how competitive were the games? I’m not sure on the seeding, but the group stage could have had Haitch and the lowest seeded pro playing eachother. Should these games count in this statistic? Those wins are most likely attributed to skill and not civ balance.

Or looking at games over time. Are a lot of France losses early in the tournament before some of the OP strats came out? You may think all the players came in prepared against all the crazy stuff. If that is the case, I encourage you to watch the Kaiserklein vs Don Artie match. Kaiser clearly wasn’t prepped for Spain Logistician water play.

Very much disagree.

  1. Tournament setting: as explained above these summary statistics aren’t representative of reality. They might be, but they aren’t guarnteed to be
  2. Ladder setting: With the ELO system, everyone will have a 50% win rate. You keep rising until you stop winning and balance out there. This is balanced on skill not civ balance.

French is the most recommended civ to start with. Most of the pro players also played on TAD and EP where French was very very common. Spain was seen as weak (and even hyped people up when seeing it since it was rare in tournaments. SPAIN TRAIN), and Port was seen as just a water civ. I can easily see the Kaisers of the world defaulting to French as a first pick. Those group of people even occasionally have a sort of gentleman’s agreement to not play ‘Lame’ civs.

These summary statistics are interesting. They point in which way we should start further digging into the data. Taking them at face value at a high level is reckless. Making statements based on just the table isn’t accurate. If you want to make a case for French being UP, go to ESOC and get the data on a significant number of French games. Figure out what was the first pick. Figure out how good French is on the map. Then you can come back and say something like “French were chosen as a counter pick in 15 games. Despite being seen as a counter and on a favorable map, they only won 45% of the games. When first picked they won 30%. The probably need a boost to be competitive”


For the record. It looks like in the actual Pro Division tournament (Round of 16) French were only played once: Mitoe in a first round game (no counter-picking) on Saguenay playing against Perez using the Chinese. He lost.

The sample size is too small to mandate a buff based on win rate. But it does call into consideration that French were basically not used in the Pro Division. This alone isn’t enough to mandate a buff, but rather to put a red flag to look at French closer.

I’d invite other people to look into French and gather data.

source:Pro Division - AoE3 Global Championship 2021 - ESOCommunity

Nice to see the stats, means absolutely nothing in general though due to the counter-pick format of the tournament and only being able to use a civ once.
Not surprised france is low, just a good all round civ not sure they directly counter any civ.

I agree with the pick rates that show spain/port/swe as the strongest civs, still I think they were picked a lot more just because spain/ports are the go to on water maps for basically every player. Sweden definitely needs a nerf though, hopefully it’s coming soon.


They got nerfed, so it is natural to not be highest pick rate anymore. Same with aztecs.

Yes, but from the highest winrate stated by dev to second low pickrate and winrate.
French also got nerfed many times but still have average pickrate but lowest winrate.

I would consider these 2 civs are balance victim after DE.