So Representation says that "each city you found" will reduce the cost of policies. Is this only for cities I found after adopting the policy? Or is it retroactive, so that all cities I've founded benefit from the policy?
The true answer requires another term. The cost of policies is founded based on neither the current situation nor in a way that applies to all your cities. The cost of policies is calculated from a formula that depends on the number equal to 'the most cities you ever owned' (annexed or originated). This term introduces an extra cost as a function of that number, f(C).
Representation substitutes 0.67*f(C) in place of f(C) in that determination of social policy costs.
So as you can see, in some way, Representation reduces the degree by which cities increase the costs of social policies. The inaccuracy of that is just that cities do not increase the costs of social policies, because C is not a count of your cities, but is the number equal to 'the most cities you ever owned'.
Based on the wording I have been assuming it's not retroactive and only applies to cities founded after the policy. If this is the case, do you find yourself rushing to this policy when you go Liberty? I still don't, since I want the faster Settlers of Collective Rule, and I feel like I'll benefit from the free Golden Age later in the game.
Thoughts?
Representation should always be the last policy you ever take in Liberty. The math shows that not only does Representation -> next policy come later in total than just taking that next policy, but in fact Representation -> next policy -> N more policies cannot save up as much culture as the cost of the policy after that, meaning that cutting Representation out in the first place gets you the whole string sooner. And this is
all that Representation does. By cannot I mean practically cannot. If the climb were steeper it would be literally cannot; even the rising savings are not necessitated to catch up to an exponentially rising hump. As a simple example, ... the actual exponential curve 2^x is the sum of its entire history, 0 to x-1. Any P% of those costs summed up could only result in P% of 2^x, which is going to be less than the modified 2^x*(1 - P%) in every case that P is less than 50.
Delnar's estimates say an extra policy might turn up by the time of the Rationalism finisher. So into a tenet? So now you gotta ask if the conditions for that occurrence - the overexpansion early - costs you indeed more of an opportunity than the culture costs. Heck, it could be strictly worse than that, if your steady development and lack of emergency response owed to 5 terrible cities, no national wonders, and collosseum maintenance , produces better culture over that span of time anyway. And then making the policy pop into the ideology instead of too early is a challenge too; speeding up culture and slowing down Rationalism sounds like a formula for failure on that point.
And the variability owed to cities is a factor too! The cost determination results in the paradox that social development slows your social development. So adopting policies quickly just means they're getting out of reach faster. You're better off accepting high costs and
not adopting policies so fast, to let you have a boom of cultural output with favorable terms in the formula later.
The anthropology of Social Policy adoption is off.
Representation's side effect is one that is sometimes worse than nothing, because golden ages are always better to get later, sometimes pitifully ineffectual early, and the Golden Age it triggers is not free.
When you consider that Liberty track's only benefit is itself just an award of great people points in the Capital, which ALSO are awful early, it means that Liberty should be delayed finished, and Representation should come last.... and maybe only the four other policies in it should be taken ever.