Play around with Construct. The only way to gather enough experience to make something worthwhile is to sit down and experiment. Also, search the forums for simple questions.
1.No way to put movies, AVI plugin sucks.
2.Question makes no sense.
3.Make it yourself with events
4.You don't need a behavior for everything, events can do practically anything. Hint, Use variables.
5.Events, math. experiment. search forums.
6.Construct isn't very documented or user friendly, you have to experiment.
7.No, not without heavy eventing or a custom workaround
8.it's probably too big, adjust scale
9.Draws it's texture at runtime. You can paste other objects or the entire screen into the texture, or draw lines. that's it. Very versatile.
10. UAYOWD (use at your own discretion) Not widely used, not fully tested/stable.
11.Search forums
12.Avoid at this point, too complicated for now. A layout just holds object positions and objects to be initialized, they run one at a time.
13.Just define the water as solid, or make the movement stop if it's not on land.
14.Don't use it.
15.No, use the wiki.
16.Variables and files.
17.mouse&keyboard action ->hide cursor
18.Use animated sprites.
19.Many ways of doing this.
20.Don't zoom out, or could be graphics card bug
21.Forummmmmmmmms
22.No, just add it to a sprite. Some effects are special and don't display anything like mask and erase. You need to experiment.
The best way to learn is to experiment. Open construct, and play with things. You won't be able to make a complex game as your first game. It doesn't matter how motivated you are, it is just out of your skill level. It's like trying to build a plane without even knowing how a wrench works. Forget about multi layout persistence, forget about massive stories or battles. Make a sprite move around the way you want it too, write simple output to a file, and get accustomed to how objects are picked and used in construct. Learn how layouts work, learn how sound works and learn how everything you think you will need works. I guarantee you that your core game idea will change based on what discover about Construct's inner workings, and It will be for the better. To use construct effectively one must work around it's faults.
Making a small game doesn't mean that it won't be a good one. Every successful game has been a small test (or part of one) which evolved at some point or another. All of my games have evolved from experimentation or testing of Construct's features. I've been using Construct for years, and still haven't looked at many of the objects. Don't plan out your game fully before you start making it, because it's design will change based on the tools you use. Play with the engine and let the game come to you. You can't rush game development. Get past the learning curve, (you asking these questions) and then you are ready to start work on a serious project.