The impact of civ choice at 1000 vs 1600+ elo 

For me Britons and Celts were the "main characters" of the game, due to only owning the demo for 4-5 years and learning to play on the William Wallace campaign.

For me around 1000-1100 ELO, the thing that seems to matter most about a civ is my own familiarity with it. I feel like I play MUCH better with a civ if I've played it many times and know how to use their bonuses, and I would guess that at high levels one advantage players have is that they're familiar with most or all of the civs in that way. So I typically pick between a handful of civs I feel most confident with.

I really would like to see stats for below 900 ELO games. Admittedly anecdotally, most people I have seen hovering around 1000 ELO understand basic civ matchups and have build orders. To see an environment where most people do not possess or utilize this knowledge, I feel like we would need to go below 900 ELO.

I can't imagine the amount of information you have given that enhances the experience of plebs like me, pros, or even developers. Chaps like you keep a solid community engaged with these objective standpoints you've had consistently throughout all of your videos. AVE!

I remember choosing my first civilization all those years ago. Boy was it fun to be a Trade Federation main

I think this video is very flawed because 900-1100 rated players aren't new. They generally know some build orders, keep relatively consistent villager production, and know the strengths and weaknesses of civs somewhat well. You'll see very few 1000 elo players using spears against archers or something. If you want players that are somewhat like your analogy of clicking random buttons until someone resigns, you're going to have to look at something like 600 elo.

Why does Age of Vampires sound so interesting as an alternative to Age of empires? Like Age of Mythology, but for the Victorian age, and still as supernatural?

It could be interesting to compare how well low rated players do with civilizations with passive bonuses vs civilizations with bonuses that require player attention. For example if civilizations could be grouped by how many of their bonuses are passive, there are probably other features like unique units that should also be looked at but this can be starting point. This could perhaps be developed into a ranking of civilizations based on their ease of play.

This phenomenon is happening with much more recognition in the FPS I play regularly (MechWarrior Online). The top-tier players are taking a direct role in balancing the game, literally telling the developers what stats to change, with their rationale, and the developers are taking them at their word and making the recommended changes. Meanwhile, the lowest-tier players, which includes all of the brand new players, who don't have any of the same skills as those in the highest ranks, are able to heavily abuse mechanics that high tier players know to recognize and actively counter, but which are passively extremely powerful.

Spirit is the only RU-vidr I know that can make data collection, comparison, and analysis be entertaining...

In basically every competitive game, the higher the skill level, the more your choice of character/class/faction/whatever matter, because players are better at making full use of their resources and strengths, as well as at punishing mistakes and weaknesses.

It took me a solid like 2 years of playing as a child (i played young. Like 14 i think?) To realize i could PICK my civ. I wouldnt even notice my civ until i got to feudal age and the architecture changed. When I finally figured it out, I'd pick my civ based on the architecture and the was a their pre-game splash sound. Such innocent memories. I also distinctly remember being scared if the 'you have been defeated' sound. I remember describing it as "evil chickens laughing at me."

Spirit, I have an idea for a series that would help newer or struggling players: How to Use ______. For a long time, I ignored the mangonel line. I always figured that it was too slow, too difficult to hit with, too low of range, too expensive, too squishy, too dangerous to my own units, etc. I only began to use them after someone mentioned it in a forum for a mission that I was struggling with. I think a lot of other low elo players might be suffering similar unit paralysis. A video series on confusing or daunting units might be really useful.

In my experience, the thing that mattered most for me about civs is how easy/clear their gameplan is. E.g. Franks, Britons, Ethiopians are pretty straightforward, so I'd play better. Playing this game with or without a clear plan in mind I'd say is a difference of maybe 200 elo or something. When you learn how to use a civ, that changes things. So now at ~1250 elo, it's civs like Bohemians, Malay or Malians that I struggle with, because I find it hard to say what their gameplan is and how to use their boni.

0:58 i have a counter argument, since timing can vary a lot deppending on somme civs, picking one and developing good habit with that particular civ is even better and will make u progress more quickly and in a healthier way. the real thing veteran should say is wich is generic enough to help new player developing theyr basic habit, without making them play a playstyle that they dont enjoy. really i'll say the same thing as in any other game, play what u love and learn the basic with it. be self critic before looking what the other does, and dont tend to learn of ur error.

Come here just for Sotl voice. Playing very rarely AoE2 these days (just casually vs AI). Still waiting for Age of Mythology DE or 2. Settings have great potentional with completly different civs/gods and monsters and its timeless RTS. I would even take new AoE set in ancient. Just dreaming.

Well for sure, when i started i could forget playing Persians, American civs or elephant based civs, playing as Teutons, Byzantines and Britons is how got started. And i would say the European powers are generaly the best for new players given they have decent tech tree, can always use cavalry which is way simpler to understand than eagle warriors and have somewhat simple to understand bonuses to boot. Defensive civs are also quite great for new players who wont be as agressive as higher skilled ones and really solid and mainly, cool unique units also help. Who wouldnt want to see teutonic knights mowing through entire cavarly and infantry armies ? Blobs of longbows melting everything from safe distance...

You know you've done something right when SOTL references your work. Well done coolios. Always appreciated your work

I feel it sort of matters. What matters more is how familiar the player is with the Civ. If they keep picking the same Civ, regardless of how good it is, they'll get better with that one. So I think picking the same one is more important than overall choice

I really liked Meso Civs, specifically Mayans and Incas when starting. Very simple reason here - Eagles do fairly well against most types of units and are resistant to arrows. As a consequence of this, I could make eagles, make halbs and not worry that much about microing them because I knew that in most cases they can be sort of left to do their thing for a while while I fix my eco. In general, civs with low maintenance units were my favourite. And still are. Incas, Mayans, Malay 2hs, Bulgarian Konniks, Goths. I know Franks and Britons are general favourites but they can be high maintenance untill serious mass.

I think I'm a part of a majority where as kid, I always played Teutons and Byzantines. Teuton's architecture + the UUs of both civs just kept me coming and coming back, much as I liked every now and then playing some other civs (Usually Mayans or Huns because I also thought their UUs looked cool). Now as somewhat of an adult I took a more "informed" approach to the tech trees and mixed that approach with my own "historical preferences", and I can say I've settled on playing Slavs, Magyars, Poles and Cumans. I can't tell you why I've leaned that way, as I'm Portuguese.

Lithuanians spoiled me and I feel my best playing them. Bulgarians are probably my favorite tho.

This made me think of an idea of having a ~1000 rated player being told what to do at all points in time by a 2500 player versus a 2500 player limited to the in game choices of a 1000 player. Wouldnt be a very scientific test but would be interesting to see if micro and speed is as important as correct game choices. The 2500 player would que up villagers quickly but only if the 1000 elo coach remember to tell him. Flip side the 2500 elo coach would likely know to remember to tell his 1000 elo player to que up vills but he may take a lot longer to do so

Neat to see how you used stats to answer this question and explained all of the methodology. More content like this please!

As someone who is also a wargamer, (tabletop, throwing dices and all). I remember when I followed some of the competitive Wargaming scene. There were some Army/Civs that were considered "Gate Keeping" as in, an army which makes a difference between a "Good player" and a "Bad Player". In other terms, if you don't know your rules well enough and/or if your own army composition is good enough then you're gonna get stomp by this "gate keeping" army. As such I wonder If some AOEII civs could be considered "Gate Keeping" ? (Franks maybe ?) It could also be strategies that Low ELO players tend to struggle against such as : Scout Rush or DRush, Which would turn civs which excels at those strats the "gate keepers" ?

I feel like most brand new players will gravitate to civs they have a personal connection to, for example, if the player themselves identifies as German, they'll most likely play Teutons, French, Franks, Japanese, Japanese. Sucks for those of us living in countries not represented in the civ choice though!

When i first played AoE2, Celts was my favourite civ mostly because of the movie Braveheart. For new players, it doesn't matter which civ it is. Just go with what you feel is right, and then you can discover the most optimal civ for you.

This is a fascinating discussion ngl. I love meta discussions like this about the game and scientifically arriving at an answer rather than simply relying on personal intuitions about the game. It also goes to show that balancing the game by the top players affects lower level players quite a bit too, and therefore is a good way to ensure that the game maintains a healthy meta for all levels of play

Imagine doing this as an assignment in school. Not only he’s gonna get A+ but the teacher might as well let him pass the whole course. What a detailed analysis. 👍

For very new players, I think the most important thing is to improve by choosing a very standard civ with a clear game plan, following a build order and probably repeating that same civ over and over. The Franks fit that really well with a Fast Castle Knight Rush following a very clear build order with a focus on economy first then army, and if it works it's very rewarding. If you want to improve I think that's a good route.

I come from a totally different beast of game, where what you pick is only relevant in extreme cases where something is stupidly op or unplayably bad, and the important thing is how you perform across your own games rather than what characters are being played. With that background, I have a personal bias to pick what I enjoy rather than what's considered "the best", even by statistics. That doesn't mean I understood anything that was said here, but I enjoy the passion with which SOTL explains everything, so as always I carefully listened until the very end

That wasn't the result I expected, so I appreciate the work done to get this result.

Back in the day I've always been a casual player in this game and for me civilization choice was also always a matter of flavor. I picked teutons because the Teutonic knights had cool capes and felt "broken" due to their high melee armor or because I liked their building style and stuff like that. It was about tech and units too, of course. But stuff like that mattered too. :D

there is a catch in here. by exluding civ pickers, we only value people who play random, which is a small subset of the playerbase in lower elos. we essentially start evaluating how good a civ is in random play - as in, no preperation for specific strats. that inherantly seems to favour knight civs.

I’d say it is important in a sense of learning the standard villager start. Some civs are wildly different starts.

One of your best vids! I hope you're going to use it for your maths academic work! What a fun way to learn about stats and variability :)

At 1000 elo, I was already following builds, thinking about my civ's weakenesses and strong points, and even did a bit of micro from time to time. I wouldn't call that being casual player, since I did already think about how to play the civ I was playing. I already knew how to play the "different" start civs like chinese. I never considered myself even a mediocre player, but having played with friends with I would consider casual, they were nowhere near my level. I would say "casual" as something below 600. It was mentioned in the video that if it was checked below the 900 elo, maybe even lower than 600, then the average would become closer to 50%, and I believe that's true (specially if you remove outliers (bengalis, Gurjaras, etc), which I think wasn't done). Could use more research, as said.

Everyone has a playstyle that will likely suit one faction better than another. I think that's important at lower elo to pick a faction that suits your overall game plan and follow it through.

My experiance coming from moba games, I personally believe civ choice is VITAL to new or casual players. No, not because of counter picking. But because of familiarity!! Sure, once you know how to play one civ and understand why you make the choices you make,you can play any civ. HOWEVER you will still be better at one civ then others due to amix of habbit and familiarity of playing that civ the most. Therefore, choosing your "favourot civ" for players who are still learning, be that due to casual gaming or being new to the game, is possably the most vital choice. Not because of game balance. But because of personal growth and learning in the game.

I think of "new" as meaning having played fewer than some number of multiplayer games. 900 ELO is nothing like that kind of new.

Can't bieve you're still churning out quality content of a game that's existed almost as long as I have. Thanks for making the aoe2 community even more special.

Anecdotally,when I was still playing Dota regularly,heroes like Slark,Blooseeker and Spirit Breaker were considered OP in average public games as they could easily pick off isolated players by themselves;and then snowball into becoming unstoppable. But in competitice and higher level of play these heroes would be considered terrible as teamwork would nullify their strengths. Interesting how balance perspective and reality shifts across skill level.

Hey man, love your work! I think the stats for this video are tough as devs will always want in a balanced world all civs to come around 50% win rate. However, the more interesting question is do certain civs counter others? Like, I believe Hera mentioned Bulgarians as a good counter to mezo civs. Are there matchups where the civs matter more? (More for you to dig in ;))

I'd love to see a video on how much better is the Extreme AI vs Hardest AI. Similar to your DE vs HD AI video. This isn't actually possible to do in a Random game, but should be possible in a Map editor.

Great and fun video as always! I have a methodological note: I don't think we are really comparing new players, just different levels. I have played for one year and I am at 1100, but hardly a "new player" at this point... As I did not play videogames consistently in the past, I bet a lot of people coming from other strategy games would surpass me in short time. And there are surely a lot of casual players who are not going to get much better even if they play for years, if they are not committed to learn more. A better estimate for new players would be for instance in the first 100 games played by a new account... But I guess those data are hard to get! :)

I haven't played AOE 2 in over half a decade yet I always look forward to every SOTL video

Civs make it so fun either way. Thanks for continuing to post great content!

:50 ah, yes, the classic "one on food, one on wood, lumbercamp in the middle of nowhere for 2 straggler trees" strat.

I do dislike how building more aesthetic cities, with housing blocks and specific sectors for lumber, mining, and production, is so heavily discouraged. I know, its for a good reason, but I have to wait until Manor Lords comes out to have anything close to what you can do on a more casual game of AoE2.

Great video!

I love how, in a parallel universe where all the civs are the same, Aztecs still have the highest win rate 11 Thanks for the video man, I love your work ♥

At 1080ish Elo, I found myself really appreciating Lithuanians. I could reliably perform build orders quite well, but there were generally a couple of slight inaccuracies that crept in, so the straightforward 150 extra food just helped me stay on track and account for those minor slips.

Hi Sotl Love your content and this one is exceptional awesome! 🕵️‍♀️Spoiler: Since we have learnt that Civ choice matters at any skill level 🤯 the questions coming to my mind are as followed: how much does the civ choice of my opponent matters too? Or in other words, are there any civs that perform exceptionally good against civs with a higher win rate (anti-cav civs come to mind)? Please keep going with your high quality content and your unique (mathematical) way of seeing AoE2! 👍

Hi Spirit Of The Law, firstly, excellent videos you make! I love the analysis you make, considering that it is so important in a game such as aoe2 when a lot comes down to very minute details in the gameplay of each player. Then I'm wondering, what is the minimum amount of deer that need to be in a flock of wild animals in order for it to make sense to build a mill and mill those deer, compared to farming with those villagers instead? On arabia, you often see flocks containing three deer. Are those worth going for?

I came here for AoE2, and ended up with something that could a qualify as a apply statistics project for undergrads. I really love this channel their uses of the maths

This made me think of an idea for a tournament: Reverse pickin - Your opponent choses the civs you'll use at each game.

No civic choice is important. Because players have “styles” and personality even if they don’t play well. A person may be more cautious and defensive, another very aggressive. Me I loved to raid and mongols with mangudai was perfect. Teutons on other hand didn’t really have good horse archers. If you are a defensive minded person, picking a defensive minded civ is a no brained and has a big effect on your win ability.

Great video! One potential issue though: how many games were played in each of the ELO brackets? The more games being played, the smaller the mean absolute deviation should be for the bracket. This may not matter here though, especially if the highest ELO players had the most games played

You should write a paper about this and send it to a scientific journal Your videos are like scientific papers even with discussion High quality videos, congratulations!

Incredible video as always. Insane math skills.

When the most important factor in choosing a civ is how pleasing the building aesthetics are and the cool factor of the unique unit 😎

0:03 subtitles: "playing age of vampires" I think I would love that game.

Yoo SOTL please could you make overviews of civilisations but from the perspective of countering and defeating them?? For AOE2 please!! Love your vids btw

thank you for the insight ! =)

> plays Teutons for the 5,987th time in a row Huh?

I love how auto generated subtitles make you talk about "Age Vampires 2".

Hey Spirit. Was wondering if you would consider a new series about which Civs might generally compliment each other in team games. That might be too vague since it's highly dependent on what you're trying to do strategy wise, but it might be more doable if it's just about covering each others weaknesses in general. Love your content, if you have a Patreon or something I'll definitely check it out. Sorry if it's rude or annoying for fans to give out potential ideas.

Thinking back, there are two things that impacted what civ I chose: 1: The Unique unit 2: The architecture of the buildings

I don´t play ranked games, but the first 3 minutes of the video sums up pretty well how i play and why those civs are way better in not professional games. I play with 4 - 7 friends, mostly the kind of game that last 4 - 5 hours and you end up having tons of resources. Micro in those games is almost useless, except some particular moments. So, civs with passive bonuses and better tech trees are far superior than the ones that are more micro intensive. A common strategy in the late game of those games, is to spam units on attack mode to the enemy and just see the fight develop itself. Targeting only important units.

As a low elo player, civs matter but knowing what your opponent's civ is more important. For instance, I played a lot with Britons. It is expected that I'd got for longbow. However, I don't know well more than 5 civs, so I have no idea if my opponent has a good counter to longbow or not, and if I should adapt first. So Britons can be his nightmare or can be my nightmare

Civ choice is important to new players, cause the most fun part of the game when you start is theory crafting with the different civs

I feel this. As a casual player for 20 years, I've always picked Japanese for the lower Wood cost and Fishing bonuses. Stacked with faster Infantry attack rate and the ability to just get more arrows from your Towers, it's always been my go-to.

5:21 that archer-micro :D

When I got introduced to the game, I was told you always had to have the "Unlock all techs" on. So for me and my childhood friends, all civs were pretty much the same with the only choice being the Castle Unit and the funny chant at the start of the match xD

I did not expect a Pearson's R in an AoE 2 video about civ picks lmao

I just clicked the like button and the number went from 2.1k to 2.2k.... what an awesome moment!

In any game. Whenever playing ranked it's important to pick those things you know best. And in general, for everything. Whenever the result is very important, you have to act in the most effective way available. For casual cases. You can experiment and fail, especially if the goal is to learn.

In my opinion the question should be more "what civ you should avoid as new player"? And answering depending on where you want to get.

@Spirit Of The Law Could you do a video on the different scorpion bonuses? Ethiopians, Khmer, Chinese, Celts and maybe Mongols

I remember when I started to play AoK then I played first Teutons (Teutonic Knights looked and look so cool) and later the Byzantines (defensive civ). When I learned the game a little more I began to play the Franks and after AoC came around, the Spanish.

Random variability in skill at lower ranks is the biggest culprit in these mmr systems. You have to go up a massive number of points to see a noticeable improvement so most of the range is pretty similar, but as you go you slowly start to hit skill cap an plateau. Once you start hitting skill cap, factors like map and character/faction choice become more relevant. But they always matter alot.

I feel that even though civ choice _really_ starts to matter when you get to high level play, it is something that should be addressed at low levels for one simple reason - choice paralysis. People like me, those who have been playing since the original Age of Kings and have learned each civilization as they've come out, might know how each civ generally plays, but the most amount of civs we had to learn at one time was when the original game released with 13 civs. All the other civs were added in batches of 2-5 with the various expansions over the years. Sure, we may not be an expert at all the civs, but we were able to at least get a feel for them over the years, both by playtesting as them and playing opponents online. However, a new player who wants to get into the game nowadays has _42 civilizations_ to choose from when they start a game. That's a lot to choose from, and a lot of time you need to put into the game to get a general feel of how each civ plays to find the one that fits your playstyle. So it's totally understandable for newbies to ask questions like "which civ is the best" or "what's the easiest one to learn".

@spirit of the law, I have a question. What number of villagers at which pop caps would you recommend in general? For example: say you play a game with a pop cap of 300. How many villagers should one generally have in the early, mid and late game? Hoe would that vary from strategy to strategy?

I mostly play Goths because I do a early Feudal infantry rush and just spam the enemy from every side until he resigns, free loom and cheap infantry helps in that. Also because that's all I know 11. I am 621 ELO player on 7 win streak after losing 14 matches at the start.

I stopped playing games like this. One of the reasons is I kept picking civs that were low-tier because I thought they looked cool, such as the Chinese, and then I got punished badly because I had no idea how to play the game and didn't realize there was a "correct choice".

I think for low skill level it is more important. High skill level can quickly adapt and they know what each civ is good at and know how to deal with it. The thing with high skill lvl is tho that they know what to pick and do with it. Even the more specific bonussed civs. But also what to do against which enemy civ, which is why civs are important at high lvl but more flexible. For a lower skilled player they usually dont know all civ play styles, all bonusses. So they need a civ that fits their own style. Do you prefer cavalry or archers? Do you prefer rushing? Or defending? Personally im more a casual player, i do like to pick random civ. But i kinda got one favorite way of playing, which just fits my thinking.

I think they need to buff Bengali, looking at the stats Bengali is noticeably the worst civ, the gap between Bengali and the second worst civ is huge. I hope they make Ratha takes lesser bonus damage from Skirms

I used to main saracens a lot in original game, and then i found out how op mongols is their cav archer combo with ram are unstoppable really fun to play them. Also try byzantine because of their full tech tree.

The perfect mangonel micro messed up my focus during the methodology 11

Good vid. More reasearch is needed....

I think at low ELO civ choice may matter more when players are used to 1 or 2 strategies and they accidentally get a civ that cant pull of the strategies that they are used to while high level players could work around a civ's weakness much more easily

My favorite faction is Portugal. Ballistics on gunpowder units with organ guns. Not to mention that fetoria

Perfect video

I am not sure if I follow correctly your calculation. Does the final average number relates to a statistical test, like a chi^2? If so, what confidence level does 2 or 3% refers to in relation to the null hypothesis?

Good video

This needs to be seen in relation to matchups, If a civ is over picked (Franks, Huns, Aztecs) civilizations that have good matchups against them will have inflated winrates that aren't representative of civ balance but rather player choice. You need to compensate for civ pick rates to control for disproportionately high picked civs and their matchups.

I love how he just shuts down people saying "Civ pick doesn't matter at a beginner level" with facts and logic.

Very sound scientific study! Where is the arXiv link? :-P Now I can rest assured that all the time I spent mulling over the tech tree was worth it, even as a noob. Would have been worth more as a pro, but it seems like it does matter for both.

I do also think that at lower elos it matters more to know your civ than what civ it actually is, and since many new players look at the pros and what civs they recommend ,people will learn those civs better which leads to the correlationof top civs

2:30 before watching rest of video, I want to say that I think both arguments are true: it is less impactful if new player chooses more generic civ, but it is very impactful if he chooses more unique style civs like Khmer or Saracens for example.