200 gold +2 gp + world map probably isn't a bad price for horseback riding if you figure out how much it will cost you to research it. If you do set your research slider to 0, you'll have a lot more money than that to give.
When buying a tech from AIs, the price depends on a lot of things. The most important factors are how much it costs to research it, what new things the tech gives, and how many civs know the tech that you have contact with. There's a good article about tech cost in the strategy articles forum. Basically, if you try to buy an important tech from the only civ that knows it, it will cost a whole lot.
The best time to buy a tech is when some AIs already have it (so it is cheap), but some AIs don't have it (so you can sell it to them for other techs, money, luxuries, alliances, or whatever). You can see why this works best when you're somewhat behind in tech, but when you can out-research your enemies, the strategy HawkWings described works better.
Buying has a number of other advanteages:
You get the tech right away, rather than selecting the tech you want and researching for several turns before you get it. This is helpful for trading so you can look at what tech other civs need, buy it, and sell it to them all in the same turn.
You get the tech right away, but pay for it later if you pay gpt. Paying gpt is also a good way to persuade that AI not ot attack you. And if they do attack, you don't have to pay them anymore.
Even though you're giving the money to the AI, they usually waste it by rushing stuff. They don't make very good use of it.
So yeah, you can get ahead by just researching up to regent difficulty, but above that, you should do some buying and selling keep up and get ahead.