I'm a big believer in learning-by-doing, so in the face-to-face video classes I've taught, I have a few custom projects I give that I think are very valuable. (The online classes I teach are usually more regimented in syllabus and lessons -- they are also good classes, but the course material is more locked-in and there isn't time for this type of assignment.)
The project almost everyone learns a good deal from involves cutting a chase sequence. It sounds like the easiest thing in the world, but it really gets people engaged in understanding the connection between editing and understanding -- the fact that every editing choice should be based in a deep understanding of how we comprehend what we see.
The actual project is much more direct: the goal is usually a 90-second chase video. There is some type of scenario set up -- for example, an action that motivates the chase or just the appearance of the characters -- then a chase begins.
That really sounds like the dumbest possible scenario, but here's how we proceed: we start by watching scenes from a range of chase films and trying to understand the visual systems at work in those. For example:
In the midst of that immersion, we start to get the traditional techniques. For example:
Use a landmark, so when we show one character go past it and later another we will have a sense of how far apart they are.
Frame a shot with the escapee and the pursuer together so that one is perceived to be gaining on or receding from the other.
Build up a language of looking out of the frame and then revealing a point-of-view shot. This lets us use reactions from the character in the chase language.
Use a directional language, based on screen motion left-to-right or right-to-left, then develop a technique for "turning around" the screen motion when needed.
Find ways of using shots that communicate speed -- for example, a camera close to the road.
From what they've learned, each small team of students (usually four works well) storyboards their own chase, and then they tape it.
Generally, the basics go very easily (and it can be quite rewarding to see these come together) but for almost every project an interesting thing happens: there is almost always one point where something went wrong, and a crucial shot can't be used or just won't work. Having to solve this problem -- which can range from easy to very tough -- brings out some serious thinking about editing and how to reconsider / rethink a project in the last stages of the process.