Some stats of KOTD

First kudos an NerFox, who again made a public google doc here:

I myself collected data for when games Ended and when people got up to castle/imperial age.

My results:

123 matches played
117 of which reached Castle Age (95%)
61 of which reached Imp (49.6 %)

Average time: 0:39:12
Average Castle time (when reached): 0:21:42
Average Imp time (when reached): 0:39:49

Average time spent in:
Dark/Feudal: 0:21:42 (55.3 %)
Castle: 0:13:34 (34.6 %)
Imp: 0:03:57 (10.1 %)


Of all 117 matches that reached Castle Age, 80 were won by the player who reached castle age first. This includes all the games where it was very close who reached it first and also all the games where one of the players was definetely way too greedy with his uptime and threw the game completely unnecessary. This makes it 68.4 % chance of winning the game if you hit Castle Age first. This means a statistical correlation of at least 61.3 % with a probabilty test of 95% as the deviation of the 117 analysed matches is 4.3 %.

i.e. That if you hit castle age earlier you have at least a 22.6 % higher winrate than your opponent. The stats so far suggest it’s about 40.4 % in reality, but ofc we need to account for possible statistical errors.

But if this is true, I think 68.4 % : 31.6 % winrate just for hitting castle age first is too much imo. The powerspike seems to be too strong, if this is truely the case.

At the same time the games are way too repetitive. I miss a lot of strategic diversity. It’s way too agressive - that means that both players want to be the agressor which leads to extremely snowbally games, that are horrible to watch repetitively. I don’t like watching games that are decided at about minute 15-18 and then drag out til 40 minutes cause it takes so long to snowball it to a clear victory. I don’t dislike this in general, but it’s too much if it applies to about 2/3 (felt) of all the matches.

I prefer having games with agression vs defence/greed, that have a lot of back and forth. Both for my personal game experience but also to watch at pro level. It’s just way more entertaining and leads to way more strategic diversity, as both players have different win conditions.
Don’t get me wrong we had some very entertaining games this KOTD. But I must say that most of the games were quite boring. It often either it was a very long skirmish with one major engagament that decided the whole match (like we see it often in arena actually) or a snowball of an early achieved lead. Most games felt repetitive and unnecessarily dragged out for only a few “decisive” moments.

I hope we get back to what Arabia was before, this new one is way too strategically restrictive. It’s not arabia anymore imo. It’s something new, but it’s not the map I loved to play for years at is was so entertaining, so diverse in how people approached it.


I think 68.4 % : 31.6 % winrate just for hitting castle age first is too much imo. The powerspike seems to be too strong, if this is truely the case.

This is the wrong interpretation. The first player in Castle Age typically gathered resources more efficiently and/or inflicted more pressure on the opponent’s production. This is exactly how the game ought to be played. In fact, early aggression and large standing armies weaken the most significant “power spike”, which was braindead 3TC play immediately upon reaching Castle Age.


Or defended against the opponents aggressive play efficiently.


It should be the other way around actually. The one who is behind should actually try to hit that powerspike earlier. He needs to be greedy to have a comeback.
If the result is, and there I agree that it currently is in kotd, that the guy who is ahead also gets often the kickstart to castle age, this only leads to an even harder snowball effect.
Ther is no

how this should be played. But people who are behind need some kind of comeback strategy. So if the one with the lead has the timing advantage of one of the possible strats to compensate for an early disadvantage (getting tech advantage), then it’s no wonder we have these boring “snowbally” games all the time where one single advantage early on decides and the remainder of the game is just total domination of the player who got the advantage.

We need to have these comeback strats, it’s how the game thrives. It’s what makes it interesting.

Yes, that’s certainly more efficient resource use: less waste of resources in the form of military. Perhaps the first condition should include both gathering and use.

On the contrary, being “behind” in a strategy game by definition must mean having a higher probability of losing than your opponent. Rubber-banding assistance for poor play is fun in Mario Kart, but not a strategy game. Note that this isn’t to be confused with spurious comebacks due to the commonly known advantages some civs get in a future age, or comebacks due to human error by the opponent.


What’s interesting about a game that is decided at minute 15, but lasts 40 minutes as it takes so long to snowball the lead to a V?
Comeback strats have several purposes, one of them is also that they should lead to a way faster win if the one behind can’t execute them in the needed way. They are risky, that’s what makes them only “comeback” strategies.
If the one in the lead has the advantage to cutback your comeback strategies by just being faster to the timing, then he needs to go the full grinding way and slowly snowball his lead to the V in “total domination” fashion.
But that’s boring and repetitive. And if it is like this we don’t need ages, TCs and miltary. The only meaning these different advantage fill in the game is to provide comeback strats the game can thrive from. If they actually fill a purpose to snowball a lead, they are totally misplaced and unnecessary. It leads to repetitive and boring gameplay.
■■■■, we damn need comeback strats. It’s just essential for an rts to have them. it’s basically the one thing rts thrives from, what it make it interesting to play. If the game is decided by one early skrimish and the reminder is just snowballing that advatage to a V, it’s just total grind. And also not fun to watch.

In this kotd already about 50% of the matches were like this. Aou could tell like mid feudal “ok, that game is over, just waiting for the call”. And then it takes like 15 minutes until the lead has grown until it’s undeniably over. And that’s absolutely terrible, there is no life in it.

1 Like

Wow. I’m going to sound rude but this is one of the worst use of statistics I’ve ever seen. Short story: Correlation does not imply causation. Long story: Please read about confounding variables, you have done almost every mistake described on this page, and these are eliminatory mistakes if you work with statistics or plan to work with statistics.

In your case, the “castle age time” is probably just a meaningless factor confounded by the actual important factor “which player is in the lead when they age up”. The correct causal interpretation of your data is probably something along the lines “The player who is winning has 70% chance of going up before their opponent”, which I’m sure is a statistic that surprises no one and is nothing to worry about balance-wise.

Typically the way you avoid confounders is by recording more than just 1 variable. For example, something you could do is to record the score of both players in your data, so that you end up with 2 groups:
group A: player who is losing score-wise clicks Castle Age first
group B: player who is winning score-wise clocks Castle Age first

And now you could compute separately the impact of castle age in each group A and B, this time highly mitigating the confounding effects. I’m pretty sure if you do this you’d see little to no correlation anymore. Or maybe you would, I don’t know. But while you haven’t done that it is impossible to deduce anything from this data.

This math is useless and gives the impression that you’re trying to impress people with figures. The correlation is obvious. It’s the causation that needs to be proven here.


If you know anything about statistics you would know that it only finds use in cases where you can’t find a direct causation.
All things you can compute fall under the case of “stochastics”.
Statistic correlation tests are the base of all our science. Damn, the whole world of physics bases on that.

The thing is that you can’t make test cases here like you do in physics. You can’t make setups that negate all exterior influence. Damn in that case it’s even impossible to creatoe one as the timing advantage would be ofset by better eco/military númbers that would disturb any result.

So in all conclusion it’s impossible to even find any causation, it’s only possible to find a correlation.
Demanding a causation proof in this context is hypocrit at best.

Wait, have we been watching the same KOTD? I must say I’ve skipped some match-ups myself due to time or whatever, but about 60-70% of the games I’ve seen have been actually really really great lol. There’s been actually so many matchups that have been insanely fun that I’ve actually forgotten about them (in lieu of new ones) and when people mention them again, I’m like ‘oh yeah, that was a great set!’

So many of the games also were really dynamic, with comebacks and whatnot to the point you’d not really know who’d come on top unless it’s really obvious through match-ups (and even then there were upsets!) or prior mistakes made.

I don’t think this is always explained through statistics either, the games have been so different that it’s hard to tell what some things may be the result of. Reaching Castle Age first without any context is like… duh? Obviously you should win when you reach stronger units? Now let’s not ignore previous gameplay and look at what happened in Dark and Feudal age, what are the strategies, the maps, the playstyles, and let’s go with statistics again.

After all, the stats say on (this, I think?) Arabia Incas have top 2 winrates, but are one of the least picked civs - clearly every pro is just dumb and should pick Incas more! Right?!



Have you even read the page I mentioned ??

Of course you can use statistics to prove causation (“prove” in the physical sense, ie; no reasonable doubt). You just have to eliminate confounders. I explicitly proposed a way to do it (group players by “who is ahead in score” before looking at the result). In your specific case if you had grouped players by score you could PROVE that there is no cause-effect relationship between “castle age first” and “winning” and that the correlation was in fact purely confounded by “who is ahead”.

I think there’s a large enough margin between “negating all exterior influence” and what you have done, which at this point is just “pointing out one random correlation in a small dataset”.

Not only I know about statistics, but I also play AoE2, so I can confirm that your interpretation makes no sense in both cases: it doesn’t make sense from statistical point of view, because you haven’t even attempted to eliminate confounders, and it also does not make sense from a player point of view. 2x more chance of winning by being first to castle age? Incredible, why do we not see all pros go for FC then?

1 Like

Drafting has had the biggest impact on the civs and several outcomes.

That’s not how physics work. Sorry, but the whole physics is based on confidence. There are people who make Hypthesis, then we make experiments to check if they are “correctly representing nature in their respective boundaries”. Like we still use “classical mechanics” cause besides we know it’s actually “wrong”, it’s good enough for basically all the applications we have on earth, the margin of error is negatable. Also we knwo all the boundaries of all the standard models we use.
And yes we know these boundaries and we know that the theories we use are only working in these boundaries. Which is actually an indicator that all our theories are still actually “wrong”. They are just good enough in the defined boundaries to have extremely high accuracy, like not imaginably high accuracy. But only inside the boundaries of application.
There is no “proof of causation” in physics. It’s all confidence of accuracy inside defined boundaries.
So please stop spreading bs here. Something like “proof of causation” isn’t possible with anything real. You can only proof causation inside of theoretical constructions. Everything else is always just correlation.
Of course there is the difference of me not having eliminated all other effects, like it is done in physical experiments:

Which isn’t possible here cause we can’t create an experiment that eliminates them. It’s that easy. It’s just not possible here. So please stop demanding something which isn’t possible. If it was possible, btw, we would need to set up a whole “theory of the game” where we would need classify all possible influence factors. Then we would have “understood” the game. Which would make it redundand to play, as then everybody would know “how to win”.
So that’s total nonsense. Not only that it would be absurd amount of work, it would also, even if it was possible, destroy all the fun we have with the game. And yes, I think this whole “how the game needs to be played” bs is actually terrible for the game. The thing that makes games great and fun is the curiosity that you actually don’t know “what is better” in most cases and have to make decisions. At some time you learn preferences which is fine and gives you a nice revard for your resilience in trying to figure it out. But it’s a horrible idea to theorycraft a game cause of personal perception “how it shall be played to succeed” cause it takes away all the curiosity and learning process. At the same time it reduces the diversity of the game just because of some personal preferences. It’s just a horrible idea.

Since when is the score a good indicator who “is ahead”? Especially in the early game? You make things only more convoluted if you try to correct a disturbed influence factor with an even more disturbed one. The only thing you can possibly try to do is also make a confidence test with a score difference at defined timestamps. Then you would get a confidence check of the score difference and you could possibly state that score lead is more or less influential than the castle age timing advantage.

Sorry, but your way of correcting it is satistically wrong. Just wrong. You need to make specific “multi-factorial” tests there which need a lot of data cause you need to figure out the different dependencies of all your different factors and correct them. Like going up also has an influence on your score aswell as having a lead which influences the score can lead to a possibly faster timing. So you have double dependency which is actually even impossible to solve exactly cause of the noise. I mean you can get a result but because of the nature of double dependencies the smallest noise will lead to extreme jumps in the dependency factors. That’s the reality. Not only that it would be insane work, but most likely you will actually get bs results there.
The best thing we currently can do is try do define specific factors and make confidence tests how big their impact is actually.
So please be realistic here, that kind of confidence test is actually the best reasonable approach we have. If you like you can oppose my castle age timing confidence test with your own of having a score lead at a given timestamp or whatever. Then we can compare them to each other.
We can also make a double-factorial dependency test, but as I said: This actually won’t give good results cause of the noise.
What I can do, and maybe should have done, is defining some kind of “factor” of dependency. But I need to think about how to do it cause there are games where only one player reached castle and there is no measurable time difference to built that dependency factor around. (eg winning percentage over timing difference). I agree that this would be the correct way to apply it, but I actually don’t have a good idea how to define that dependency factor in a reasonable way with the given dataset.
But please if you think you have a more accurate dependency with your score lead, go ahead. Don’t demand from me to make your work. I mean I am pretty sure the score lead has currently even more influence on the outcome. But only because the games are so insanely snowbally. And if the castle age timing advantage is actually amplifying that snowball like you imply and I follow that actually, then we have a huge problem:

We only have 3 “difference makers” in lead in the Game:

A) Eco
B) Military
C) Tech

If the Tech advantage cannot be used anymore to compensate for “falling behind”, the only two remaining factors are eco and miltary. But as we all know the one with the miltary advantage always can add eco behind his agression, which means in this case he can safely snowball the game slowly. He just needst to stay ahead slightly militarywise. And that’s exactly what we have seen in most of the games. A boring slow snowball with the one getting an early lead adding eco behind his military dominance. Ofc that is “good play” and also needs a lot of skill. But it’s boring and repetitive.
The tech advantage as one more element of “difference maker” just enables that kind of trinity where the player behind goes “all-in” on one of the three factors. We need to keep this alive otherwise the game becomes full boring grindy repetitive “meta” mode all the time. And annoying cause the player behind is just “dying slowly” and the player ahead must grind him to death. That’s horrible experience for all.
It’s just better to have this comeback potential in the game. It also makes games not dragging out unnecessarily. Cause the “all-ins” are risky and if the player behind fails with them he can instantly resign. But if there aren’t these comeback plays possible, games need to be grinded till the unavoidable end, just waiting for a “major mistake” of the one with the lead (which barely occur anymore cause we know the game so well).

It’s only in the interest of all Players that the tech advantage stays as a potential comeback factor. And currently it looks like we are indeed losing it, besides the powerspike is actually too strong for the investement. And that’s concerning. We need to understand that this kind of trinity is just essential for a varied gameplay and if we lose one of the difference makers cause it is dominated by another one (in this case military advantage) we get repetitive and grindy gameplay.
And I am atm already annoyed by the amount of “total domination” and “full snowball” games I see in kotd. The best game I’ve seen was already Hera vs JorDan, which only was “different” cause both players didn’t wanted to take a possibly deciding engagement till the very end. So ther was a stalemate until there was the one big engagement in the end. But there was no real “back and forth” in the game. Just very little skirmishes and nobody wanted to go the risk of overextending. This was the best game in the tourney so far, In every other kotd we had like at least 3-4 matches that were about the same length but a constant back and forth of the players. This is missing totally.
Maybe we had a few games where the lead changed in midgame at some point, but it’s just not comparable to the games we had in previous kotd tournaments. They were so much more engaging as there was this back and forth all the time. This new kotd may be “clean” but it’s also “sterile”. It’s repetitive meta play and everytime somebody tries something “off meta” he gets punished heavily for it.
The only thing we can witness is the actual meta shift that happens and the people who figure the perfect meta out more and more. Like the Viper with his 4 ranges strat. But do we want to see this being the only “viable” metaplay or do we want to see diversity?

I prefer diversity. I prefer to see different approaches, strats and playstyles crash. And none of them being “clearly superior” to one of the others. That’s what I want to see.
I don’t want to see 123 of which about 70-80 were basically the same game. I don’t want to watch the same game all the time again. That’s not entertaining.
And I also don’t want to play the same game all over again. Yes I am not on pro level. But it won’t take long until people here adapted the pro meta. And then we will have the same repetitive game all over in ranked as we currently see in kotd? Hell no.

If this happens I will look for a different game, some game where I am not forced to play the same game in repetition all the time. I don’t play games for repetition, I play them for their diversity. RL is repetitive enough to me.
And please stop that narrative of “how it is supposed to be”. Thats bs. It’s only your personal preference and you will figure out once you nerfed everything you “don’t like” to the ground you will miss it. Take it as motivation, try to beat the strats you don’t like in the game. Play the game fair.

This whole “arabia ist too wallable” or whatever narrative is just the hypocrit attempt of people to win a fight outside of the arena on a green table. That’s not how a game should be played. Learn to beat your opponents on fair conditions. And you even don’t know what you are doing there cause currently you are destroying the game with it. The fragile strategic balance that kept this game alive for 20 years is already falling apart cause of that dumb narrative and games becoming repetitive.
BTW what also need to be told that after one general strategic approach is nerfed (if there are still players just prerferring to play like this), it will lead to only a few wins for the other approach. Cause aftert the defensively minded players lost a few games they are then paired with “weaker” opponents. In the aftermath a player that only plays defensive will ALWAYS win 50 % of his games and a player who only plays agressive aswell. That means regardless how often you nerf defensive play you will ALWAYS win only 50% of your matches against the defensively minded players. Not because they are “one the same level” but because you are matched with players that would actually stomp you if they would play agressive. Players that are actually better than you. So the whole story of “defense is too strong” in arabia is actually also just a biase. At any given time defensively minded players will lose or win 50% of their matches against offensively minded players. And if the offensively minded players can’t bear losing 50% of their matches against defensively minded players they will just repeatedly complain. But the reason they lose isn’t that the strat would be too strong, the reason they lose 50% of their matches is because they are matched with players that are actually better than them. That’s already the case. Except for the very, very best players in the world, agressive play is actually the way easier path to begin and climb the ladder. Playing defensively on arabia is the way harder path to take. And if you don’t believe me, ask Viper. He will confirm that, as he already did several times.
Again the reasoning you only win 50 % of your matches against defensively minded players is just the matchmaking, not the “balance” between the two general approaches. And this 50% will stay as long as there are still defensively minded players playing that map. Every time you nerf walls or make arabia more open you will have short term success cause then the defensively minded players will lose some elo till they are on their new “spot”. But once they reached that they will again win or lose 50 % of their matches. And then you will complain again: “Walls are OP!”??? Like already so many times?
Learn to accept that you will always win or lose 50% matches against defensively minded players - no matter how much you nerf walls. It’s how the matchmaking works.

The people who have bear the burden in the end are these who like to play both styles. Cause they then will win like 3/4 of their matches if they play agressive and 1/4 if they play defensive or what so on. That’s the worst thing about that whole story, that they have to bear it. They are then basically forced to chose one “side” to not longer have that discrepancy and get an “equal” match. And for absolutely no reason cause the offensively minded players don’t get a higher winrate against defensively minded players with these changes. And that’s the players the devs should try to keep as they enjoy the game for it’s diversity, for the whole thing. If the devs lose them cause they feel forced to chose a side they don’t wanna chose, it’s the worst possible outcome. And this includes most content creators like sotl, t90, viper… When they go cause they don’t have the same diverse experience anymore, the game will die. And again, for absolutely no reason cause the intended highering of winrate against defensive players won’t happen. It will aways stay 50:50 cause of the matchmaking.

If this continues, the game will be dead soon cause first we will divide the community into two factions “offensively minded” and “defensively minded” players. Then we won’t have any maps they can both agree on so they never clash. Then they will get bored because of the repetitiveness of the game. Then they leave. If we don’t stop with that dumb narrative of “how this should be” and also revert the last arabia changes we will enter that path of no return and slow decline of the game.
We still can turn back, but that need to be done now before the division of the community has gone too far. We can already see that division happening - this should be a wakeup call for all of us.

IN all conclusion I just again state: Please revert arabia!
It is fine if devs decide they wanna add a more “agressive” map type to the pool. I totally subscribe that. But Arabia is the bread and butter of the game and the changes don’t even change a thing in the Players experience if they have a clear preference “how to play” the game.
Offensively minded players will always have 50% winrate on arabia, the same is for defensively minded players.
The only thing that changes with this kind of changes to the “agressiveness” of arabia is that defensively minded players will lose some elo and agressively minded gain some. It doesn’t change anything in the game experience of either type.

The players who suffer are those who like to try different general approaches cause they then are always matched with “unequal” opponents and either have it way too easy or way too hard to win. These diverse players are then forced to “chose a side” to get equal opponents and have a nice experience.

So in all conclusion this whole “change arabia to be more agressive” is just a stupid idea. It only divides the community, but changes nothing in the experience even of the players who always demand these changes.

And we already seen that multiple times. Devs can make arabia more and more agressive - as long as we still have defensively minded players playing the map game experience for the agressively minded players will return to the same state as they again will lose 50% of their matches against defensively minded players. We are moving in a circle there and the only players it affects are those who actually enjoy playing either way - and in a negative way: They have unequal matches.

We need to break out of that loop.

Anyway, I analyzed your dataset, and found out another troubling correlation, hear me out.

It seems that when a player has the letters " V I P E R " in their name, they have a higher chance of winning.

Even more incredible, when the player plays SAXOPHONE, their chances to win can reach up to 68.4%. (Especially if they are Norvegian.)

I ran the analysis over 117 randomly picked matches of KOTD1, KOTD2 and KOTD3, that is hundreds of games, and the saxophone player almost always won, in fact 2 of the KOTD winners were saxophone players. Hence I can say with very high confidence that there is a correlation between playing saxophone and winning at AoE2, the statistical correlation is of at least 61.3 % with a probabilty test of 95% as the deviation of the 117 analysed matches is 4.3 %.

ie. That if you are a Saxophone player you have at least a 22.6 % higher winrate than your opponent.

If this is true, I think 68.4 % : 31.6 % winrate just for being a saxophone player is too much imo. The powerspike seems to be too strong, if this is truely the case.

Developers need to do something now, for the sake of balance. For example, make known saxophone players start with one less villager. The game should be about strategic diversity and not about who’s the better saxophone player. Recently tourneys have become very stale and repetitive due to this, because when you spectate a game with a saxophone player everyone knows he’s practically guaranteed to win.

I hope you agree with my analysis @casusincorrabil since it is the exact same methodology as yours, I even kept the phrasing but that was more for the sake of entertainment. Feel free to include it in your OP, I release this discovery into the public domain, although when this correlation becomes widely known it may cause a severe crisis for AoE2 because many players will start buying saxophones and gain an unfair advantage. Note that my reasoning is impossible to disprove due to our fundamental lack of knowledge of physics as you brilliantly explained in your post above, so I hope you won’t try to engage into that, that would make you a hypocrit / bs spreader (using your own words).

1 Like

Nice try, but actually works against your case.
If your numbers you put out are correct, that would mean that getting castle age first gives you the same advantage as being Viper…
Isn’t that a bit too much of an advantage for just getting castle age first? ^^
Just gettin up earlier makes you equal to the best player in the world? ^^

Even with reductio ad absurdum you need to be careful. Sometimes you actually only give others just a cheap shot against your case. ^^

It depends on what you mean with this sentence. If you mean that, if we suddenly would spectate a random game where we know nothing about the players and one of the players clicks castle age first, then you would be correct saying that this player now has 68.7% chance of winning. That is what statistics tell us, that “being first to Castle age” has a link with “winning”. But the statistic does not tell you that “being first to Castle Age” is the cause for “winning”, and that fact alone makes it a completely meaningless information for any practical situation. For example, let’s watch the same game as above, but now you are allowed to look at the units on the field, player APMs, or the BO, then suddenly the 68.7% figure would no longer be true, it would become very close to 50% chance of winning when being in castle age first, because the previous factors were what actually controlled the game result, and not the castle age time itself.

So there were some factors A,B,C such that that (A,B,C) => (winning), and (A,B,C) => (castle age first), and as a result, there is a correlation (castle age first) => (winning) but it is NOT a cause-effect link. A,B,C are called confounding variables. Contrarily to what you think it is perfectly doable to find and eliminate confounders, but to find about this you’d have to at least record more than 1 variable in your dataset, I agree it’s more effort, but it’s you who decided to give an interpretation to this stat in your OP, so why did you share stats if you’re not willing to record them properly? Without analyzing confounders your study is blatantly wrong. Not inaccurate, or empirical: wrong. You are using a correlation indicator in a context where it’s not appropriate. You could have made the exact same post without sharing any stat and say “I think that being first to castle age gives too much of an advantage” and it would have the exact same mathematical value.

Yeah I didn’t want to point out that his data doesn’t account for who the better player is at all.

He just says he who goes first to castle wins.
He discounts literally everything else that could cause them to be the first to click castle age.


I agree. But KOTD (and in fact all of Memb’s tournaments) are meant to be aggressive. He prefers aggression after aggression over aggression before aggression. I personally don’t mind watching them at all. So I’m fine with the tournament and its map. But I also wish we had a different Arabia in the map pool and it becomes more wallable (especially after all the palisade and house nerfs).

1 Like

A strategy game can’t be fair while rewarding suboptimal play, even if people give it heroic underdog names like “comebacks” or “non-meta”. What makes it a strategy game is that some types of play reliably deliver better outcomes than others, and that these can be judged based on the map etc. As a matter of pure personal expression, it is fine to play any way you like, of course. But a 50% win rate independent of how well you play would end the game as a competitive platform.


A lot of people want to play fast-ish castle age 3TC boom every time without even thinking, which is absolutely doable in Arena and a whole culture exists around it, but it was a bit stupid that it became the dominant meta for Arabia that every pro recommended to every new player, every time, in their build orders, videos, etc.


Really? Usually it was 3 TC against 1 TC all-in. It dependet ofc, because if 50% of players chose the more greedy approach with 3 TCs and 50 % the all-in only 50 % are greedy vs all-in and the others are strategic mirrors.
I think we should try to get back to that state where it oftne was 1 TC vs 3 TC. I liked that state. I know some people see this as some kind of “meta”, but with the way TCs are implemented I don’t think there is any other way to ensure the strategic balance. If it becomes 4 TCs vs 1 TC surely greed is Overpowered and if it would become 2 TCs vs 1 TC it actually can’t even work as the initiative of pressure usually easily more than compensates for just 1 TC more.
Adding incrementally TCs befind is actually more a thing of an agressor trying to snowball a lead than a competitive greed approach. The benefit is too marginal for the risk taken.

I think the state of 1 TC vs 3 TC we often had like 3 patches ago was actually very close to the optimum for a “strategic balance” imo.