Why was Mongolia so dominant for you?
honestly it was maybe a bad example, I rolled an oval map with a start surrounded by city states, which is maybe equivalent to 2 diffculty levels lover, the first round of tributes launched the snowball and I'm not even experienced in domination victory. I feel I could have won tradition 3 cities in deity with mongolia. Anyway, I still feel successful warmongering gives too much science, it should allow you to keep up, not overcome. There is also the difficulty issue, where the more enemy units the better for the yields on kill. As for the AI, authority AIs tend to fall off, but when they do work they are the most unstoppable snowballing foe.
This seem like a pretty big map issue. What maps are people playing on where you can get 9-10 cities without attacking someone?
this is a big factor. I love to play the communitu map for aesthetics but it really favors progress, as coastlines are more elaborate thus expansions are easier to defend with more chokepoints. I should try playing a smaller map with same number of players. But in the end, that should be they key element of picking tradition: great start, little room.
-----
More generally on the thread:
again, I don't agree with "you have to be wide at some point, but you can choose how you get there". Like, it's probably the current state in deity (though I have won pure Tall too without a single conquest), but for the sake of diversity it shouldn't be a design decision. It's perfectly fine in terms of design to reward tall, it's a way to grow as legitimate as others, what you don't invest in expansion you do invest in other ways. I've posted in this thread how it has the potential to be fun, engaging gameplay, oposite to what we imagine about turtling. A lot of playstyle balancing comes down to the diplomatic AI. For me the funniest way to play a small army is to manage the great powers. Get support from other players when attacked, not out of charity but because they don't want your wonders to fall in the hands of their rivals. Delay your big push for victory until you think you can hold an all out assault just long enough to win. I hope to provide more detailed input if I find the time to understand the AI code.
In a related way, tradition providing a stronger start but falling off later is elegant and fun in terms of immersion with policy title, but I don't think it works in terms of design. Wide empires have all the means to take down tall ones by definition, if you remove late game scaling from tall (and there is already very little) there is no point in that gameplay. Or at least it shoehorns you into leveraging one timing conquest by medieval (and science is key for timing attacks, which tradition is not particularly strong at).
Hapiness:
As others have said, happiness is a big issue too, which means satelites are hard to grow, and are thus a 5% burden on your science and culture (S/C) for a long time. I don't know if Tradition needs a hapiness scaler with the capital population, or if progress hapiness needs a nerf, but managing hapiness should be a headache for large empires not small ones. Or maybe tradition needs to get back to the old "no S/C penalty until 3 cities, 7% above". I think there is also a design conflict where some want Tradition to be the Tall gameplay, while others want more flexibility with Tradition being the Capital gameplay but not necessarily smaller empire. There is another conflict in the definition of Tall: big capital and crappy expos, or (like in vanilla) big 3/4 core cities?
Medieval trees:
Tradition power is closely tied to its available follow-up. The Great Unstacking of artistry was good in terms of opening up options, as it is now an interesting tree for progress and authority. However it nerfed pure tall play, which is one step forward one step back in terms of gameplay diversity. More generally, I wonder if all the nerfs to CV were correctly targetted. Was Tall CV the issue before, or just that CV was too accessible on your way to SV or DV? The nerfs to CV may have balanced victory types in terms of aggregate results, but I'm not sure they have addressed that issue of balancing accidental CV vs intentional CV. As for other trees, I'm not sure how they work with tradition but both of them imply widening the empire I guess, they are not for pure Tall.