Hi all! I want to show some interesting data I came across while visiting aoepulse.com. I would like to discuss some of the conclusions we can extract (as always, my threads are fully open for discussion, that’s my aim), so feel free to comment on them.
So I opened the aforementioned website and chose “All maps” (no filter), 1v1 Standard (not Empire Wars) and all patches/updates since Dynasties of India was released. Then, I filtered by ELO (<800, then by groups of 200, until >2000). I made the average win rates of each ELO range, and calculated the Standard Deviation (this value tells you “how far from the average the values are”). For those who aren’t familiar with Statistics, if civ A has 51% and 49% Win Rates (WR) and civ B has 52% and 48%, both civs’ average WR will be 50, but the Standard Deviation will be higher for civ B. Why is this value important? Because it highlights the shifts of WRs among all ELO ranges. I’ll show you the complete tables:
I divided in four groups the civs for discussion:
No, I’m not talking about civs that rely on Cavalry. These are civs that, across all the ELO range, show similar WRs and thus, a similar performance, no matter the level of the player. I will mention them: Bohemians, Britons, Burgundians, Burmese, Franks, Hindustanis, Incas, Japanese, Khmer, Koreans, Lithuanians, Malians, Mongols, Portuguese, Saracens, Tatars, Vikings. Here, we can find high WR civs like Franks (53.2%) and Hindustanis (54.4%) and low WR civs, like Koreans (46.1%) and Tatars (46.5%). What do they have in common? My guess is that they all have the same strengths and weaknesses and, as the level of players grow (and you add extra elements like new strats, harder to micro units, etc.), their overall performance doesn’t change. For these civs, it would be interesting to see if they also have similar civ matchups as ELO grows, but we will leave that for another thread.
Lower WR as ELO grows civs:
There is a group of civs that seem to be countered the best as ELO grows. In this group we find Bulgarians, Dravidians, Goths, Magyars and Persians. I feel each of the cases is singular. Bulgarians have no early eco bonus and you have to use the Blacksmith for early savings. They have strong melee options but can be played as Steppe civs. Krepost drops are really popular and effective at low ELO. I guess their flaws (no eco and no xbow) are exploited by higher level players. Persians have an insane boom into mounted units, but they need time to get there, and aggressive players can take the momentum of early game civs to kill them. Goths are a late-game oriented civ, and that favors less aggressive players, which are more common at low ELO. Magyars have access to the best late game composition, and they are very smooth to play, but experienced players will know how to prevent them from getting to Huss(z)ar + HCA, mostly by taking advantage from eco bonuses. Finally, Dravidians rely on micro units ( from Archery Range) and food hungry units that you need to feed with Farms.
There is also another group of civs that peak at the lowest ELOs, but after 1000-1200 seem to stabilize. These are Bengalis, Ethiopians, Malay and Poles (who have the highest WR at <800 ELO). I’m not sure about the reasons for these to perform like this. Bengalis would want to get to Elephant spam (because using Rathas is a hard task at 700 ELO), Malay would use Trashswordsmen with Arbs/Siege and Ethiopians would be Torsion Engine-ing opponents. As for Poles, they have a smooth eco and a straightforward Knight play, and that is juicy for us low ELO people.
Higher WR as ELO grows civs:
Celts, Chinese, Gurjaras, Mayans. We all expected Chinese to be on this section. It really seems that, in hands of a experienced player, Chinese perform the best thanks to their insane eco and versatility. Gurjaras have one of the highest WRs (if not the highest for nearly all ELOs), but at lower levels, they seem to be managed a little bit worse (perhaps when they find a good counter to their Camels?). Mayans lack of Cavalry (which at low ELO is the most popular play) is maybe the reason for their lower performance. Finally Celts suffer from lacking a strong, easily manageable Castle Age unit. Celt Siege is very strong, but micro intensive, and not every player knows how to Hoang.
Berbers, Huns, Spanish belong to a separate category, in which they show similar performance except for the lowest ELO. I can’t think the reason why 2 FC civs and the “I don’t need to remember to build Houses” civ seem to perform worse at >800 ELO, especially as they all have access to strong Cavalry and “Cavalry Archers”. Also notice the peak at 1.8-2.0k for Spanish. Is that the Nomad site? What do you think?
Aztecs and Byzantines are the opposite. While they show a pretty stable WR among all ELOs, they have a peak WR for >2000 ELO. No pro player wants to face Aztecs on Arabia unless playing with Bulgarians or Byzantines, but Aztecs rely on many support units for their Eagles and they need a strong eco to afford them. Byzantines depend a lot on the production of discounted units to force the enemy to spend “standard prices” on their units, and need a good timing for their Imperial Age. That is difficult to achieve with low experience.
If we remove highest and lowest ELOs, we can also find civs that seem to overperform or underperform at certain ELOs.
Valley civs: Sicilians and Vietnamese get their highest WRs for highest and lowest ELOs. Both get some powerhouses (Hauberk Cavaliers or Serjeants for Sicilians; Battle Elephants or Rattans for Viets) appreciated by low ELO players, and unique strats or features exploited by pro players (5TCs into First Crusade, Donjon “raiding” for Sicilians, Vietnamese Anti-Archer theme).
Mountain civs: Cumans, Slavs and Teutons seem to perform their best at different but certain ELOs. Slavs have their highest WR at 1k-1.2k. Cumans perform their worst at extreme ELOs. Teutons have an above average performance between 1k and 1.6k ELO. The three of them are Cavalry oriented civs with a boomy eco, and their raw power, difficulty to boom and diversity of units make them shine at different levels.
Erratic civs: Last but not least, Italians and Turks show a crazy curve, with peak performances at diverse ELOs. Turks seem to be well used by pro players (mainly closed maps) and at 800-1400 ELO, but not at higher or lower levels. Meanwhile, Italians show three peaks: for pro players, 1.4k-1.6k and 800-1000. Can you have a guess at why do these civs perform like this?
Well, I hope you enjoyed the read. If you feel it, you can discuss your experience on certain civs to help explain the data. Happy New Year and may your Castle drops be successful!