The trebuchet is an old art. Last time it was seriously used, people were assaulting castles and pillaging the countryside (for the most part). But oddly enough, for a time period associated with such little scientific innovation, is the trebuchet really so complex? It isn't hard to think why it works: it pulls a sling, that sling releases, and so away it flies. But for accurate predictions and optimization, it is actually a fairly complex beast.
You may expect me to explain how it works. Well, I won't. It's complex enough where few definitive outlines exist for their ideal mathematical operations, and presently I don't have the time to take a crack at all that math. But I mention it because I feel as though it is a good example of complexity generated not by subject matter but rather the amount of subject involved. You have the sling, rotating around the end of the lever arm. You have the counterweight and beam, a first class lever, and in the real world, that beam has a linear mass density as well. You have friction around the fulcrum, friction in the chute, and a need for precise timing with the release of the ring from the end of the payload side of the beam (NOTE: if none of this makes sense to you, I would advise checking out http://www.redstoneprojects.com/trebuchetstore/how_a_trebuchet_catapult_works.html - it gives you a basic overview of how it all works). While this is all stuff we've covered, concepts we know, the fact that they are all combined in a complex system makes it much more difficult to be precise with.
So just take note that, while solving problems using these concepts may seem easy, similar problems can become much harder with just a few things added in. Me, personally, I like this. It means there are always new things to try and explore.