Where to start?

One suggestion would be to get a friend and explain to them what you think you know about the game up until this point. Get the basic materials you'd need to play the game (just the moving tokens, you can write pencil on index cards for any cards, and use a dry erase board for the board itself).

You and your friend will each have some conception of what the game "should" be. With this in mind, use your components to try to play through the game, figuring out (almost "Calvinball"-style, but in a slightly more structured way) what should go where as you go. Expect disagreements over what people thought the intent of the game was. These disagreements will clarify things for you.

Don't count on this experience being fun in the same way as you would expect playing a game to be fun. It will be fun in a completely different way.

Keep your mind open as you explore your way through your first run of the game. There will be some essential seed that the game must have for the game to be what it is; know what that seed is and communicate it clearly to your partners. But in all other matters, listen to and strongly consider everything your partners say. You can feel free to make any decision regarding the rules at the drop of a hat, as long as you are willing to change things on the next go around. Soon, after a few trials, the best way will become clear to you and your partners, and the rules will stabilize.

This isn't necessarily the best way to go about making a publishable game or anything but it should be fun, and it should get you off the ground.

I'd endorse what jpfed says there: get a friend and talk it through with them. Until you feel confident enough to bang the ideas out on paper (whether as a formal ruleset or merely a set of decision trees), trying to explain it to someone else is the best way of firming up an idea that exists in a nebulous form.

The drawback with that method is that you have to be sure you know what you want, and not get totally sidetracked by the really cool ideas your friend will contribute - and they will! After all, that's what the whole concept of "playtesting" is all about; all this is, is playtesting at what I would call the "alpha" stage of development: the pre-prototype stage when the ideas are still in need of nailing down.

Ultimately you'll have to get something down on paper, but I wouldn't worry at this stage exactly what it is. Once you get the hang of this design game, you'll evolve your own mechanism for doing it (I always start with a very formal ruleset for instance, although it's evident that most of the time it's only really useful to me, and not to anyone who actually wants to play the game!)

The important thing is that, as with any new art or skill set, it's really tedious to begin with as you simply don't have useful reference points to know if you're even heading in the right direction. All I can say is that there are lots of us here who've "been there, done that, got the T-shirt" and I don't think any of us would necessarily call ourselves experts (even the published designers!) - we're all still here learning along with everyone else...