In Roll Through The Ages, there are two game-end conditions: all monuments have been built (collectively among all players), or any player has purchased 5 developments. I've played this game about five times, so I'm still pretty new to this. I usually play to build cities first, and then end on monuments and developments more-or-less simultaneously, but I wonder: If you're trying to finish the game very quickly on developments (while your opponents are still building up their cities), is it better to just try to buy a development every turn, or should you aim to build a city or two on the way there?
If you are trying to rush the game - which is certainly a valid strategy if your early rolls force you in that direction - just buy developments whenever you can. If you have fewer cities, every turn you spend playing catch-up is a turn your opponents with more cities can buy either more-expensive (and more VP) developments or monuments. You're trying to dash across the finish line before they can reap the benefits of their well-developed engines. You won't be able to make much use of your developments, so just get what you can afford and gives you points.
One advantage of this is that if you are threatening the end of the game, say with 4 developments, your opponents will have trouble committing to the big, efficient monuments (that aren't worth anything if incomplete) and will have to settle for the smaller ones that are worth fewer points per worker. This is particularly troublesome for them if you're later in the turn order.
If you or your friends finds that a rush strategy makes the game too short, you may prefer the official Late Bronze Age variant that user1873 alluded to.