frunobulax said:
Try to aim to build a 'Science City' - i.e. one that has a Library, a University, Copernicus's Observatory Wonder, Newton's College Wonder, and eventually the SETI Wonder. All these improvements will multiply together to give a very large number of science 'beakers' per turn - c.140-180 depending on city size.
Sorry, I think this is a fallacy. All the science multipliers from buildings and Wonders are applied to the base science income. They don't apply cumulatively. The only benefit to having a "science city" is if you are going for a 20K culture victory, and/or if you only have one city that has a high gpt.
In my experiece, the keys to any successful civ strategy, including one requiring high speed science, are mostly in the way you play the Ancient Age and the early Middle Ages. Here are some o fthe things I've picked up from the experts around here:
- Get your gross commerce - gpt - as high as possible by building roads everywhere your citizens are working, growing lots of people, making sure they work river-side tiles wherever possible, exploiting commerce bonus resources. This starts right at 4000 BC. Make workers to develop tiles to deliver food and gold.
- During the Ancient Age, don't feel bad about trading for techs, or running 10% research. Particularly at higher levels it's cheaper to buy than to research. Focus on getting your gpt income up and then choose research projects that you think will provide you with a trading edge in 40 turns.
- Use all the trading tricks. Contact trading and Map trading at the turns when Writing and Mapmaking respectively become available, can be huge income earners, and/or get you to tech parity. Check out Moonsinger's Trade Training thread where a single trading turn was played out out by some of the top players. It transformed my attitude to tech trading. Look out for "2-fers" - situations where one civ has two techs or more you want and another doesn't have one of them. If you can buy the unique one you may be able to get the other one cheap or free in exchange for it. You have to ne checking pretty the diplo screens very regularly to spot these opportunities. Don't imagine the AI civs are not doing this among themselves.
- Don't let city governors run your economy until you have an unassailable lead. They really don't know how to do it, and you don't want your civ to be run as badly as the AI civs run. They hire entertainers at the drop of a hat, don't know about taxmen, ignore high gpt tiles ....
- Get to Literature as fast as possible and then build libraries in all your core cities. They multiply base science in the city by 1.5.
- As soon as you can get Currency build Market Places in your core cities. These will increase your net tax income by 50%, allowing you to plough more into science. Also, if you have a few luxuries, they allow you to increase the productive population in your core cities without unhappiness problems.
- Build a second core as soon as possible, and read about corruption and techniques like RCP (Ring City Placement) to reduce it. Your first core is usually round your Palace. The Forbidden Palace provides another center round which you can build a second core of low-corruption cities. Various tricks are available to get a second core up and running fast. My favourite is to build the Forbidden Palace in a ciy local to the Palace, where it can be built in 20-30 turns during the Ancient Age, and then use the first Great Leader I get (I'm basically a war monger
) to rush a new Palace in a juicy captured AI capital city. That way the AI has kindly built your econd core for you
- Did I say don't use city governors?
- Republic is the way to go. Democracy is actually not that great in my experience. If you haven't reached a fast science rate by the time Democracy is available you are already in trouble. Republic gives an extra gpt on every gold-producing tile. Even for quite large military civs it still makes sense relative to Monarchy, as the Republic military upkeep cost is offset by the extra gpt you are earning, especially if you have built markets and libraries. Monarchy is effective if you want to engage in long slow wars, when it will give you zero war weariness. But who wants long, slow wars?
- Figure out what will trigger your civ's Golden Age, and make it happen when it is most productive - typically while you are at peace, in Republic, during the Middle Ages. Your Golden Age creates an extra gpt and an extra shield per turn from every tile that produces them already. During this 20 turn period you can build up your treasury, or use the extra gpt to fund a fast science rate. At 4 turns per tech you can research 5 techs in your 20 turn Golden Age. The extra shields mean you can build extra libraries or universities according to the time when it happens. It's a slingshot effect that can put you in the lead and in a position to leave the AI standing.
- Turn down the research slider during the last couple of turns of the project if you are doing high speed research. You often find you can finish it in the same time at a lower slider level and save the cash. If you run flat out to the last turn you may be throwing money away unnecessarily.
- Oh! And don't turn on the city governors until the late game.
Don't deny an old micromanager his pleasures 
