So, I just completed the first 200 turns of a Liberty game as Denmark (Deity level), and I have to say I do think the ideal play has changed a lot from G&K to BNW.
First, let me share a bit about the map conditions. Initially I rolled Denmark with the idea of a Metal Casting-Beserker rush. I began on a relatively Hammer-poor coastal site for my Capital, opened Tradition, and a bit into the game I discovered a phenomenal single-tile choke to the East blocking off 3 civs on that sub-continent. Far to the Northwest, there was also a great double-choke around both sides of a mountain range, holding back Sweden on a Petra start. There was probably enough space on the right for 4-cities, and enough on the left for 4 more. Two mercantile CS's in range, each with unique luxes. More importantly, a Beserker rush was going to make for a boring game. So, I restarted with the aim of testing out Liberty.
The result was that I was impressed overall, but not until after a great amount of patience. I did manage to open two Culture huts(!!), but I don't think that quite invalidates the game. Then, I understandably went for CR first, and I just expanded as fast as my legs could carry me. Pretty fast, that is, considering that I was able to ally both Mercantiles, multiple luxury trades were open, and half the new cities had a unique luxury of theirs on top of that. At turn 200 now, I am at or around the tech leader. Population has been the main driver of my Science all game.
The main thing that I noticed about this approach was how back-loaded the tech tree is in terms of raw Beakers. It probably takes the same number of Beakers to research two or three Medieval techs as it does one Renaissance tech. The result is that being behind in the early game only puts you slightly behind overall, once you do catch-up. Looking at the tech tree in terms of Beakers needed rather than total techs reveals that you can sacrifice yield of ~10 Beakers/Turn from T60-T100 due to delaying the NC, and it only amounts to a fraction of a tech come Industrial Era. The main objective is to get to Education ASAP to get those Universities staffed. That goes with basically every other tech. It's not so much about slogging through them as it is being able to use their benefits right away, or not. And Philo is one tech that an expansive opening is not able to use immediately, most of the time. In theory, going to Education through Civil Service first rather than Theology should be just as viable, as long as you don't dally too much getting there and you make up the 10 or so Beakers for the NC elsewhere.
Speaking of that, the ability to select the optimum Science site for the NC while going wide often more than offsets the penalty for not getting it early, at least in terms of Beakers per game. In this game, I found a very nice Coastal city with plenty of Jungle as my second. The NC on top of the University and Jungle TP's there have led to really great yield. I imagine the same would apply for an Observatory city. The only trick is getting high enough production, whether through ITR's or tile yield, because the NC is much more expensive on 6-7 cities.
Balancing against vertical openings as well, there is a point in the game after you've hit your 2nd GP or so that you are far more limited in your Beaker yield by your Population and Land Area than you are your ability to dedicate the Capital to Science. Land is still power in that spot, and that's also the point where conquest becomes appropriate for a standard Trad opening, but there is something to be said about settling those sites yourself, if available.
If you do settle those cities, you definitely want to settle them early, notwithstanding ITR support. The reason is, certainly, that 5% static cost increase. For an empire with a Beaker yield of 20, that's break-even when you add the Beaker from that first pop, and then all positive thereafter. So in that stage of the game, settling helps you tech. For a post-NC empire with a Beaker yield of 100 though, you're only break-even after 4 pop and Library, and behind on every input other than tech to even get there. That's to say nothing about the T150+ empire with beaker yields twice or three times that. At that stage, it's probably just expansion through conquest, keeping prime city sites only. It's a bit of a canard to say that early cities cost you less, because it's all 5% across the board, but the issue is a city being able to contribute immediately to your tech progress. Building Beaker yield early-game then expanding is rear-end backwards, seen in that light.
Another thing that I noticed but did not try in this game, Liberty openings seem to hybridize with other SP trees much better than Tradition openings. With Trad, I always want to finish ASAP to get those Aqueducts. But Reformation and Meritocracy are rarely in full effect at the time they become available, and there's even a strong argument to delaying the GP until after the first natural Great Scientist, in order to get a 200 point gift instead. You could start either Piety or Commerce for standard play, left-side Honor for warmongering, or Aesthetics/Exploration on certain starts.
Overall though, I do think that wide Tradition is a safer bet on most maps. If you can settle half a dozen good city sites without angering any neighbors AND if you can have good ITR service to your Capital AND if you can keep happy up all the while, then you'll be able to hold on long enough for Liberty's long-term nature to kick in. But, that is a lot of if's. Tradition with all 4 cities is usually still wide enough for your territory size not to bottleneck your tech progress, with additional cities added through conquest or stacking multiple coastal ITR's. And the assumption at the time you have to make the choice is that you are not on the kind of map for 6+ cities, which puts you in Trad unless you restart like I did.
