My findings:
1.)Liberty and Tradition are two paints on your palette to create your masterpiece, your game. Just like paints on the palette, would someone say, "Black is universally applicable whereas red is only beneficial at times, therefore red sucks, I'm throwing it off the palette and forever will only make black/white art"? To further the analogy, black and red can be quite complementary, just like liberty and tradition.
2.)If you are struggling with Liberty, a good "training wheel" scenario to try is Indonesia on a archipelago/small continents map. The bonus luxuries cover the happiness cost due to expansion, and the redundant unique lux means that the first few cities settled immediately add 6 GPT, addressing another BNW growth stunt, early game economy. Finally, in addition to the UA lux which covers the initial expansion happiness cost, each city that provides access to a new luxury (on the map), adds 4 happiness to the empire, making early Indonesia the antithesis of India (which is probably a good trait as noone except me seems to care for India
) early expansion provides extra happiness.
3.)I find Liberty and Tradition to be opposite not because one favors short/wide and the other favors tall/narrow, but because one provides many benefits that are critical early game but negligible late game, whereas the other has benefits that are game-breaking late game and more optional early game:
Liberty: crucial early, marginal late:republic: early game shaves off 2, 5 or 10 turns from construction times, late game maybe 1 turn
collective rule: an extra settler early doubles your empire and adds 10-25% to research, late-game it adds an insignificant contribution to your empire and can actually be harmful to research.
citizenship: while there are better options for getting early workers, an early-game bonus worker WILL find work and the faster build times are helpful, especially when added to pyramids making roads 2 turns instead of 3, late game much of your territory is improved and the best option for an extra worker is to disband him for the nickel that you get.
Tradition: optional early, game-breaking late: aristocracy: If the game is played at a challenging difficulty level, early game wonders are not an option to build with or without this policy with few exceptions, and the happiness bonus yields 0, maybe 1. late-game: with a tech lead and other build/buy options for the essentials, wonders can be considered and the policy helps. The happiness bonus is 3 or 4 in the capital and 1-3 in every other city.
Legalism: the later you get it, the more hammers you get out of it
Monarchy: 2 or 3 happiness and gold early, 20 or 30 happiness and gold late.
Landed Elite and Finisher shaves a turn or 2 off of a small capital (or other city) growing, a half dozen turns off a large capital's growth.
This doesn't apply to every policy in both trees, such as representation and the Liberty finisher being more beneficial late and monarchy being able to fuel an early game economy/happiness bank, but the benefits of Liberty can be game-changing early and the benefits of Tradition likewise for later game.
4.) One of the benefits of Liberty is that it is a much more flexible policy tree than Tradition. The three gems within the Liberty tree are collective rule, a critical early game acquisition, representation which has more noticeable cultural effect later game and the golden age obviously provides more gold, culture and production later in the game, and the finisher which is better later (would you rather have a great scientist instead of your 100 GPP great person, or your 1400 GPP great person?) Therefore a suitable strategy for Liberty is to get as far as the free settler and then enjoy the benefits of pursuing piety or Honor (or even Tradition). Later in the game, when golden ages are moving policies along, great writers can write you half of a policy, and certain wonders give you a freebie, you can quickly knock off the remaining three polcies which becomes a worthwhile tangent adding tons of happiness, a golden age, hundred of culture saved on policies, an expensive GP, and the ability to quickly improve newly conquered lands. The three gems in Tradition are landed elite, monarchy and the finisher, forcing the Tradition player to not divert from start to finish.
5.)I will digress that if I had to choose sides in the argument of SP tree vs. SP tree, I'd favor Tradition 10 out of 10 times. I played a game as France, a civilization that is well-suited towards liberty and pursued the liberty tree. Around turn 120, my five cities were struggling to maintain positive happiness, and the economy was just improving from negative to barely profitable. I needed to deploy 2 divisions of my military to ward off barb threats on different sides of my empire and then gave up when 2 battering rams and 6 horse archers marched onto an under-defended area of my land. Rather than rage-quit, I reloaded the initial auto save at turn 1 and tried again with Tradition. By the same time, my 3 cities were maintaining around +10 happiness, they were making about 30 gold/turn partially due to monarchy and the reduced maintenance but also due to the colossus's 5 GPT and the additional trade routes's 12 GPT, and the three cities were almost as good researching as the 5 city liberty. While I did lose both the fourth and fifth city sites, the extra production from population and gold enabled me to take it back aggressively. I had a similar experience and retry-with-tradition results with the Persians. I'm currently playing as the Maya using Liberty without tradition and it's going very well, but I think it's less because I'm playing Liberty well and more because Pacal is incredibly good. I imagine if I re-started and pursued Tradition instead, the results would be just as successful, probably more so.