Would like your thoughts on an observation I made playing on Emperor: I believe that founding religions predominantly does you more harm than good, especially with regards to the game at large. Here's why:
1. Researching the necessary advance delays more relevant tech
Efficient use of tech is crucial in playing Emperor successfully. This means prioritising key technologies with regards to their immediate and strategic relevance. While I don't believe in an ideal generic tech-path, I believe that all religion-founding techs except one are always trumped by other available ones in terms of relevancy,so researching them carries a hidden "opportunity cost".
In particular, pursuing the three early religions delays much more immediately useful advances such as bronze working, worker techs and the pottery/writing/alphabet path. The one exception may be code of laws/confucianism, which is useful to have asap in all games where you have the opportunity to found/conquer a large number of cities early.
2. Spread of your own religion among your cities slows the spread of other ones
I believe having as many religions as possible in your empire is preferable to having just one predominantly as it gives you more choices in terms of state religion and the related diplomatic boni, allows you to build more temples and improves the benefit of the free religion civic. Early religions in particular tend to spread quickly through your empire, and religiously affiliated cities are less likely to acquire additional religions.
3. Making the religion you founded your state religion will spoil relations with other Civs
In most games, switching to a state religion works to your disadvantage in that it spoils relations with other civs, hampering trade and inviting aggression. The exception are island starts or games where your neighbors all share one religion. But a situation in which you manage to found and benefit from spreading your religion to other (unaffiliated) civs is much less likely. Thus, even if you manage to found a religion, it is usually unwise to leverage the larger part of its benefits (+1 happiness from state religion and access to religious civics).
4. Shrines are better conquered than buillt
Shrine benefits are a major reason for founding religions, but in most cases you are better off spending ressources at conquering them than building them. To benefit from a shrine you need to found the religion, spread it and farm a great prophet. The AI is much more efficient at this, but experienced players are more efficient in their use of military might, so letting the AI found the religion and build the shrine is preferable to doing it yourself.
5. Letting AI found religions is more likely to keep them hostile to each other
This is possibly the strongest argument against founding religions: Religions foster dissent among the AI, and dissent amongst the AI works to your advantage in slowing trade and inciting war. Any religion you found is a religion not founded by the AI, increasing the chance of religiously homogenous alliances which work against your interest.
As a bottom line, I would argue that on Emperor games, the dominant (more useful in the majority of settings) strategy is to abstain from founding religions entirely.
The most frequent exception may be island starts, where an early religion can be helpful. The best religion to found my be confucianism, as it arrives with a key tech and in the mid game, giving you time to adapt to the prevailing religious climate.
Opinions, objections?
J.
EDITS: Formatting.
1. Researching the necessary advance delays more relevant tech
Efficient use of tech is crucial in playing Emperor successfully. This means prioritising key technologies with regards to their immediate and strategic relevance. While I don't believe in an ideal generic tech-path, I believe that all religion-founding techs except one are always trumped by other available ones in terms of relevancy,so researching them carries a hidden "opportunity cost".
In particular, pursuing the three early religions delays much more immediately useful advances such as bronze working, worker techs and the pottery/writing/alphabet path. The one exception may be code of laws/confucianism, which is useful to have asap in all games where you have the opportunity to found/conquer a large number of cities early.
2. Spread of your own religion among your cities slows the spread of other ones
I believe having as many religions as possible in your empire is preferable to having just one predominantly as it gives you more choices in terms of state religion and the related diplomatic boni, allows you to build more temples and improves the benefit of the free religion civic. Early religions in particular tend to spread quickly through your empire, and religiously affiliated cities are less likely to acquire additional religions.
3. Making the religion you founded your state religion will spoil relations with other Civs
In most games, switching to a state religion works to your disadvantage in that it spoils relations with other civs, hampering trade and inviting aggression. The exception are island starts or games where your neighbors all share one religion. But a situation in which you manage to found and benefit from spreading your religion to other (unaffiliated) civs is much less likely. Thus, even if you manage to found a religion, it is usually unwise to leverage the larger part of its benefits (+1 happiness from state religion and access to religious civics).
4. Shrines are better conquered than buillt
Shrine benefits are a major reason for founding religions, but in most cases you are better off spending ressources at conquering them than building them. To benefit from a shrine you need to found the religion, spread it and farm a great prophet. The AI is much more efficient at this, but experienced players are more efficient in their use of military might, so letting the AI found the religion and build the shrine is preferable to doing it yourself.
5. Letting AI found religions is more likely to keep them hostile to each other
This is possibly the strongest argument against founding religions: Religions foster dissent among the AI, and dissent amongst the AI works to your advantage in slowing trade and inciting war. Any religion you found is a religion not founded by the AI, increasing the chance of religiously homogenous alliances which work against your interest.
As a bottom line, I would argue that on Emperor games, the dominant (more useful in the majority of settings) strategy is to abstain from founding religions entirely.
The most frequent exception may be island starts, where an early religion can be helpful. The best religion to found my be confucianism, as it arrives with a key tech and in the mid game, giving you time to adapt to the prevailing religious climate.
Opinions, objections?

J.
EDITS: Formatting.