"We're talking about almost a million games here, so not only is that a ton of Age of Empire games played..." And thus it was established that one Age of Empire match weighs one gram.
The Goths are such a great example of how our perception is flawed. They seem like archer civs can't beat them because when they do win it just *feels* so inevitable and horrible as your arrows bounce off the flood of huskarls. Meanwhile in another match your archer rush prevents them from even getting to huskarls and it is completely forgettable
It is also that champions + arbalest counters goths very well. Because goths lack the last defense armor they cannot go champions if there are even a few arbalest around, and because of the champions goths cannot go huskarls either. When I play britons vs goths on arena I just know I cannot lose because of this, nothing goths can make beats arbalest (or longbow) + champions. So the problem is people compare units (huskarls beats arbalest, and goths champions are cost-effective vs britons champions) instead of comparing the army comp (arbalest + champions > to any goth army comp)
@@simontheteacher1942 but the spamming power of goths infantry is unmatched, I hardly believe what you said, goths can win that army composition just by numbers and perfusion tech. And it would take just a hole in the walls to get your economy destroyed by a flow of huskarls.
@@renzoraschioni7954 Depends how you set up your map. I agree the spam potential is very good for goths, and britons aren't certain to win (especially at lower levels because it is easier to spam than it is to micro more expensive units). However we kinda have to assume that the game did not start in post-imp unless we are on arena or black forest (both maps where the spam doesn't matter because we fight in chokes and economies cannot be raided) ; which means that britons will be pushing your economy before you can get the spam going and you'll have to deal with the army of britons one way or an other if you want to keep the economy to maintain the spam. But obviously goths can win this match-up, it is not completly one-sided (except on very closed where I think britons should just straight up win) But your point stands. What makes goths so scary for most players is that when the spam starts you need 500 apm to deal with it as the defender 11
The problem is at low elos the players are often unable to exploit that archer rush (if capable to perform one), so goths vs archer civ is an auto win for the goth player when he manages to get a castle. So, I firmly believe goths are a bit unfair at low levels. Moreover, the goth plan (spamming infantry) is too simple compared to the opponent player who has to think of a more complex strategy to perform with is civ. I think so because I reached ~1050 elo, starting from below 900, only playing goths, and I'm a complete noob, I don't even know build orders and refuse to learn one (aside the art of war missions in game which I don't consider). I repeat to be clear, I DON'T know build orders, and I made +100 elo points by playing goth, eventually becaming what I hate, 11.
@@simontheteacher1942 Exactly. The Goths are nigh-invincible if they can execute their spam game plan. You are (probably) not going to outfight the Goth Post-Imp army, even with Champ-Arb (also Goths have their own, more spammable, Champs to counter Champ). You beat them by denying them their plan.
The problem with having an obvious advantage is that it is obvious, so your enemy will tend to plan for it. This is why Spanish archer rushes and Franks using Sword-line are surprisingly effective, because people assume you won't use archers as Spaniards and will always use knights as Franks. When your enemy makes assumptions, they are foremost making an ass of themselves, and Sun Tzu says never interrupt your enemy when he is making a mistake.
When I play Spanish, I use villagers to combat to catch my ally off-guard. I load a ram of villagers into his base and start building houses. Suddenly, I change diplomacy & my villagers start attacking him!
Totally true. I'm a fairly begginner after 15 years without playing the game, but the other day I won a game as Franks without making a stable. The other player asumed I was going to make knights but I had a few spearmen while booming, his camels and rams were inefective and I could outproduced him and swarm with axe throwers.
While the statement would fit well to a number of Sun Tzu's tenets, the maxim "never interrupt your enemy when he is making a mistake" is not from _Art of War_ but a remark by Napoleon. His army engaged those of Russia and Austria in 1805 at Austerlitz, where he feinted his left flank to be weak. While Austrians proceeded to bite this bait, his corps urged N. to act. He urged patience by saying "when the enemy is making a false movement we must take good care not to interrupt him", which would later be rendered to this shorter version.
Best thing about Franks is their crossbows at low level. People prepare by making spears for the inevitable scoutrush. Then you have 20 Archers and they more often than not resign.
I feel like eco bonuses are the reason behind strange one-sided match ups. Ex : archer civs like Mayans, Chinese, Britons, Vikings, tend to have better (early) economy than anti-archer civs, like Goths and Vietnamese, so they can quickly get the upper hand on those "counter civs"
That's what I was thinking too... and those eco advantages allow them to get ahead and end the game sooner while their opponents need to hit Late Castle/Imp to start countering them... so they are behind most of the game and can't actually counter the civs.
Because goths were kind of designed to be the ultra late game spam that your defensive structures can't stop fast enough before its over (proxy 4 baracks is devastating). And the archer rush was likely not a base design of the game at the time and was what players found meta, so the stats of the units don't represent the philosophy of the civ entirely and how the game was run
well, he got lamed pretty hard, most people would have resigned right there. Also the Korean player didn't put any pressure after laming, so it was kind of his fault for losing that match.
I think the problem with Lithuanians vs. Teutons is that an infantry civ is perfectly happy to spam halbs. Armor ignoring cav still gets bodied by halbs, and Teutons have extra armor on top of being fully upgraded; that makes the matchup even worse for anything besides Lithuanians unique unit.
@@ewheimdall it’s actually both. A smart Teuton player won’t be using the Teutonic Knights because the Leitis will just roll over them. The game very quickly turns into a knight/trash war between the two of them.
I almost threw a Lithuanians vs. Teutons game a while back despite having 5 relics and full map control because I stubbornly refused to make anything but skirms and paladins/Leitis, and nearly got overwhelmed by their halb and cavalier combo until I started adding my own halbs, they ran completely out of gold, and I gutted their farm eco with hussar raids all over about a minute of game time. In retrospect, I feel like I should have gone heavy cav archers instead of paladins, Teutons can't answer those as easily.
I see it that the teuton player has huge eco advantage. Then either outproduces the Lithuanian in castle and wins the knight war or the game drags into imp. Teuton usually reaches imp faster because of better eco and wins on tech advantage. Can't fight under 13 range teuton castles etc. And overall the lithuanian player shouldn't be able to produce enough leitis to keep up with the teuton better eco.
If you don't have camels, there isn't a particularly hard counter to scouts and knights. Spear line is too easy to out maneuver, especially with cheap castles for backup.
@@Meloncov , putting spearmen around places that knights or scouts would attack and then reinforcing those positions when they are attacked helps to discourage and/or punish such attempts. You also need a numbers advantage against cavalry to make it work, most of the time. Additionally, making towers to hide your villagers in and fight raiders is just good strategy. Granted, I've never played an online game of AoE2 in my life, so what do I know?
Not sure if you’d consider this, but how about a video: “The Curious Case of Byzantines”. They’re inferred to be a fairly beginner friendly civ, but they’re conspicuously abscent from top 5s of Top or Bottom tier civs. Is this a case of Byzantines being so incredibly mediocre that they’re in fact balanced? You could highlight the civs they have the hardest time against via Arabia win-rate.
Just because a civ is beginner friendly doesn't mean it has to be the strongest or even top 10. It just means it's straight forward to get the most out of it with limited game knowledge. It's all about being able to play your game to the fullest. So let me give you an example. Britons vs Goths isn't onesided to Goths because Goths are the weaker civ overall. But you need to be a good player with great knowledge on how the game works to be able to win that matchup. The goth's have tools to beat the britons but a good player doesn't need those tools, he can use his stronger advantage and win with it but he needs a deep understanding of how the game works to pull it off. So the lower the rating the higher the winrate for goths will be, but the higher the rating you have the stats shown in this video. The byzantine gives you the tools you need to win, but using tools isn't always straight foward as it might seem. A great player will beat a decent player even with koreans against franks, 29% of games went to the koreans and those 29% of players knew the game better than their opponents. The byzantines aren't a strong civ just like how britons are stronger than goths even though the goths have all the tools they need to win. But as a new player you will be able to play the byzantines to their full potential long before any other civ. So to conclude, beginner friendly just means it teaches you what's important in the game, not how to be the best. You need to learn the rules before you can break them.
Portuguese losing against Berbers is a bit funny Historically, the Portuguese were the ones who used Genitours to counter Berbers cavalry because crossbows and bows were hard to use
@@sharilshahed6106 invented (most likely) by Scythians, refined by Persians, perfected by Byzantines... and used by some other... so in reality they should be fielded by multiple civs of the Near-East area (and even further) instead of knights. Byzantine special unit would be likely Varangian Guard. But then the game would run completely different.
While I understand why SOTL used the method he did to get these match ups, I don't think its really a fair representation to use Arabia 1v1's as the baseline for all civs. Its no surprise that most of the worst match ups here are good late game civs pitted against good rushing civs. I think that when a lot of people think of good or bad civ match ups, they are thinking about the late game, when its each civ's full tech tree and bonuses versus their opponent's. I play a lot of private games against friends where we tend to favour closed maps like BF and Arena, and in those games that usually last over an hour conventional wisdom tends to ring more true. Civs like the Goths and Vietnamese are really difficult to fight if you are an archer civ like the Britons and so on. You also tend to see a drop off of some of the more conventionally strong civs like the Huns and Franks, since the Huns have one of the worst imperial age tech trees, and the Franks' Paladin spam is predicable and easy enough to counter with most civs if you prep for it. One of the most one-sided games I had with my friends was when I was Saracens and they were Spanish and Franks, for fairly obvious reasons since its a top 3 camel civ vs strong cavalry civs, which is funny since in ranked games one of the Saracens worst match ups is against the Franks. So yeah I guess my point is that the strength or weakness of civ match ups is also dependant on the map and game length, which is why over-relying on Arabia 1v1s as a baseline can sometimes distort civ performance.
Good comment. I wonder if data is what constrained him. I do the same, play with friends in private games and the games tend to go long. But do our private games get recorded? If they do are there enough to get meaningful stats?
You know, I've binge watched most your stuff, the civ overviews even more than once... The thing is neither do I play Aoe, nor do I have any intention to do so at the moment... I clicked on your videos out of nostalgia, because I played AoK as a kid 20 years ago, but I stayed for your content! And honestly, I noticed I'm happy whenever I see you uploaded a new vid, so you seem to do an excellent job, thanks for that!
Every time I look at Franks bonuses I'm just floored by how well they all synergize together. Other civs sometimes feel like a grab bag of randomly chosen bonuses but for Franks everything comes together in such a clear manner that you almost can't screw it up.
It's even crazier considering how much of a laughingstock they were in the AoC balance. Then they got the forage bonus, the HP bonus applied to scouts, and suddenly it all comes together so much more smoothly.
It’s not just that but the knight line is a core powerhouse unit. If all the bonuses just improved the champion they wouldn’t be as good. Hell they’d probably suck. But as they stand they will always be a staple of 1v1 and team games.
Hey Spirit! Video idea. "One trick civ" is a common critique for civs. So what about a video on which ones are most diverse? Most strats and least predictable? In your opinion.
Chinese, Byzantines and Malians are the ones most known for their strong diversity in options. That said, diversity means different things at different stages in the game. Huns on paper have one of the smallest tech trees in the game, but can pursue any Feudal rush strategy you can think of and find a way to keep at it in Castle, which is part of why Hun wars were the norm on Arabia for years.
A match up I feel is actually pretty one sided is Berbers vs Mongols, Berbers favor. Berbers get a massive power spike exactly at the Mongols weakest moment, they have probably the best camels in the game, and great tools for handling Mangudai. I don't have stats to substantiate that, it's just always seemed really good for the Berber player to me.
Berbers are 53.6% against Mongols, so that's a good instinct. Definitely an advantage, but it seems Mongol's better early game may give them a fighting chance
"It doesnt matter if your game plan is narrow, predictable and counterable, if youre that good at it." - Spirit of the law when talking about the Franks
The other part is that Frank game plan is not even that narrow and predictable with cheap Castles, Throwing Axemen, fully upgraded infantry and access to gunpowder units.
A friend of mine who has never played any RTS game in his life before (he is now 34) made it to 1400 solo and 2300 TG. He only plays Franks scouts into knights, nothing else. He even plays without control groups!
Hey Spirit! I really enjoy your vids, even though I haven't played AoE in ages. I think you have a great opportunity here to teach people some basic statistics (as you have done in the past!). In this video I was really hoping to see either confidence intervals or even a chi-square test. Still some great insights!
Nice video as always. Many you could do a series on each civs with each best and worth Mach up and strategic planning to improve. I for a long time hope for the civs that I play a video with how to a lots of things. And see this on other civ may be a good way to Pic new civs to try. And see the point of view of the civs that I win and loose against may me think strategically. Thank you for all the help and work you do on Hage of empire 2. Hope the best for you.
You have missed one point. The relative familiarity of people with a particular civ. Older civs have a higher chance of people who "specialize" in them to the point that their actual elo with the civ is higher. Newer civs will have less of those "specialists" or people losing a lot of games in the process of learning how to master that civ
Goths probably 60% win I’d say. Goths counter the entire civ rendering its bonuses obsolete. The only thing Byzantines have in their favour is the cataphract. It’s super expensive to train and upgrade and not worth using until they are fully upgraded. I think Byzantines will normally face a miserable death to castle age / early imp raids before the fully upgraded Cataphracts come out.
But Spirit, what is the name of the "Frank theme song" - I've asked before several times, but nobody seems to know. I even tried to find it on Shazam, but to no avail...
@@SpiritOfTheLaw on a side-note, you have great taste in background music! It's really immersive and I appreciate it a lot! Thank you far all the videos over the years. The civ overviews have been an amazing help into picking my main and pocket civs!
@@SpiritOfTheLaw indeed, I like the choice of your background music too. Is there a list of the songs you use? I asked two times under some videos but I guess answering thousands of comments every day is too much.
I would be interested to see a graph of deviation of the actual winrate vs the expected winrate (calculated from average winrate of both civs). Maybe there would be some surprises.
one thing to keep in mind with the Portuguese stats is that some people will be trying to force the Feitoria meme instead of just trying to win normally.
Man, I love this channel. I haven't played a game in years but the dedication to math and statistics to figure out a system is both relaxing and fascinating
If i remember correctly, they once used small to moderately sized squads for skirmish purpose and a weapon they would of been carrying as a sidearm was the francisca , sometimes used to throw at an ennemy to create a window to retreat if need ever be
Can you adjust these numbers for average civ winrates? It would be interesting to see how much those matchups are decided by the interplay between civ bonuses vs how much they are carried by the relative strength of the civs. Basically take the expected win rate by taking the average of the civ's expected win rate and the opponent's expected defeat rate (100% - average win rate of opponent) and compare that value to the win rate in the actual civ matchup.
The other day, I had an Italian vs Mayans match-up on Arena on the ranked ladder. I won it without struggling much. I then checked for curiosity what chances statistics gave me and noticed that I was at the worst possible spot for Italians, as they have a 36,86% win rate with Mayans, the worst against any other civ. Surely, Arena was a factor, but the win felt pretty good after that.
To be honest I hardly see what Italians are supposed to do vs Mayans, especially on arena. I guess there was a huge player mismatch because I literally don't see the comp.
@@MadMalMan Condos? They do absolutely nothing vs Mayans, they are shredded by plumes. Plumes counter everything the Italians can make on Arena, that's the issue. Italian cavalry is pretty bad and you don't want to go full skirms on Arena. Siege is weak to eagles. Hard to see options.
I recently played a match in AOE II DE i was never happier with my fast advance times as Mongols but in the end i cried when i realized i was against Burmese, Vietnamse and Teutons 11
On the Lithuanians Vs Teutons. Teutons just need to pump out Halberdiers and they can avoid dealing with the Leitis and can cost effectively defeat the Super Paladins.
It would be more interesting to look at the difference of winrates. So portuguese have a bad winrate against basicaly erveryone. It would be more fair to look at the value: (winrate of huns vs portuguese divided by winrate of huns vs everyone) divided by ( winrate of portuguese vs huns divided by winrate of portuguese vs everyone). This way the low winrate of portuguese and the high winrate of huns wouldnt factor in so much.
Is it possible that what we're seeing here is just the effect of the small sample size on some of these match ups? How much does the sample size of the match up correlate with how lopsided the outcome is? If the correlation is very high this could be a problem related to regression to the mean - if more games were played perhaps we might end up reaching a more even outcome in the end.
Assuming that the matchups are relatively even, even the small samples of 200 games is a decent sample size to work with if you are careful in what methods you use. It could be tested by randomly taking 50 of those 200, calculating the ratios, and then comparing to the total size.
@@ProperlyBasic11 There are 595 possible civ match ups. I used python to generate 595 sets of 200 coin flips. The lowest "win rate" was 78/200 = 39% and the highest "win rate" was 117/200 59%. I don't think your strategy of resampling a population with a known win rate would help at all, but I'm not very good at stats. I think much of the variation seen in this video would be expected by chance even if the civilizations were in fact perfectly balanced. Not to say that civilisations are perfectly balanced... As mentioned above these numbers could also be skewed by a small number of players always picking the same civilisations. All that is to say that people probably overstate the importance of civilisation vs skill level...
An interesting point will be to run a logistic regression on ELO(winner)+ELO(loser)+Civ +Map against win/loss as a binary variable. That way the coefficient of the Civ can give the average odds of win while controlling for skill and map.
Your honorable mentions are not actually all that surprising when you look at feudal/early castle instead of postimp. Eg lith vs teutons: teutons have the better eco bonus and the better kts (until 2-3 relics are collected). So its no surprise they have an edge. Same goes for vietnamese: Imperial Skirm is nice and stuff, but not if you lose the game in feudal.
I haven't watched it yet, but from my experience I expect Teutons vs Goths, at least on closed maps like Arena. I went random and got Teutons against a Goth player and even though we both boomed to Imp, I just went Teutonic Knights and Scorps and rolled on through.
I think matchups like Lithuanians vs Teutons or Goths vs archer civs has an element of the "advantaged" player getting overconfident. the Lith player will go "oh, it's Teutons, time to spam Leitis" but of course that's what the Teutons expect so they're not going to mass their TKs. the Lith player is not adapting their strategy, they're rushing to put down a castle, getting the Elite Leitis upgrade, get overconfident, don't bother playing perfectly, only to meet an army of siege, hand cannons and halberdiers from a player who's been playing and planning everything carefully.
First video to remind you that the cobra car outtro from your cobra car video of the first of april 2019 was awesome and should be your standard outtro.
I wholeheartedly agree that in the original Indians Overview vid he severely undervalued the Elephant Archer. And to think that it's since received several buffs since then as well. He had the tunnel vision of treating Elephant Archers like a worse War Wagon (if anything WW are vastly inferior imo) from the perspective of how quickly they both fall to Halberdiers, but against non-cavalry archer counters, they're a very strong unit! Turks neither have Pikes nor Elite Skirms as an effective answer.
My go-to unfair matchup for when I'm messing around with the AI is Berbers v Huns. They go cav archers and I have camel archers. They go foot archers and I have genitours. They go knights I have camels. Works well against other cav/archer civs too such as Tartars
In 1vs1 maya and brits are played by xbow micro nerds / hera lookalikes and thus make short work of your workers long befor you have acess to your castle units. And even if you reach your unique unit, the feather arrow archer has bonus damage to all infantery, and the britons have acess to -20food on your champions.
I know this has doesn't have much of anything to do with the subject of this video, but I wanted to say it anyway. It is really cool and heart warming to see how this game that came out such a long time ago still have a huge player base, and even more so have as much game support and updates. Very few games can make that claim, and it shows just how exceptional of a game Age of Empires 2 is.
First time i ve played aoe was when i was 9 . I remember running around the town to get the setup . I dont even play aoe now but i watch these vids for nostalgia at least weekly
As a Aztec main, fighting infantry in castle age is especially hard because there's no time to put up multiple castles and upgrade the jaguar warriors.
Yeah this is the one that I didn't find suprising at all. The best counter to aztecs in castle age (which is when they shine the most) is longswordmen most of the time, because they can deal with eagles and aren't hard-countered by monks, (unless they commiting to crossbow, which aztecs don't do especially well)
longswords losing to knights "cost effectively" is one of the BIG imbalances in the game. i.e., the swordsman line should always be a soft counter to the cavalry line, with the spear line being a hard counter. infantry > cavalry, cavalry > archers, archers > infantry - this is the way it should be. knights are OP in castle age. SOTL did this in an earlier video ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-jBuUieb_0_I.html - poor longswords (even burmese with +2 attack) lose to knights even without bloodlines.
Considering base Huskarls have 10 attack as well, and that the Swordsman line kind of mirrors the knight line for attack, a +1 to attack would probably help balance them. That, or perhaps an armour buff by +1 to Swordsman units from LS onwards (1 base melee armour for LS and 2 HS and 2 melee for Champs) Knights are overhyped and not difficult to counter but them outright winning against generic LS cost effectively is just nonsense.
Actually agree with it tbh i always found this awkward since i started playing aoe2 (17 years ago XD), longswords should win vs knights cost effectively. Its the reason why we not see champs no where near as much as arbs or paladin. Archers have there range, knights with there mobility .. atleast swordsmen should have something which makes them atleast good in close range battles.
I am not sure about other civs, but I know Franks and Lithuanians well. As for the Lithuanian vs Teuton matchup, there are 4 points. 1. Teutons have bonuses which last throughout the game. Cheaper farms, extra garrision, free MH, HM, extra armor. Lithuanians only have 1 bonus with +150f and another one situational being extra attack with relics. Extra speed on spearmen is nice. Lacking Plate mail is a deal. All others are trifle since Teutons do not go for ranged units. Also as for the Lithuanian army, Teutons won't go for Siege Onagers too. 2. Lithuanian monks are totally negated by the Teuton team bonus. 3. With the Leitis nerf, it becomes less incetivised to create and upgrade Leitis since they will be a defensive unit nonetheless. The Paladins will be the ultimate raiding units which do not counter any of the strengths of Teutons. 4. The Lithuanian options are very limited with situational extra attack on Paladins and UU, and faster moving trash. Extra pierce armor on Skirmishers is useful only against archers and Cavalry Archers, not other Skirmishers, infantry, Cavalry or siege. Lithuanian infantry, archers, siege are trash literally.