Folks ask me all the time where I get game ideas, whether it's mechanics first or theme first. Sometimes it's a little of both, as we'll see here.
One of my favorite recent mechanics comes from Doug Bass' Garden Dice. In that game you roll four dice to plant crops on a 6x6 gridded plot of land. The dice tell you the coordinates of where you may plant. You can do other actions based on the remaining two dice results. Choosing which dice to use in which capacity is a big part of the long-term strategy.
So I spent yesterday thinking a few ways to use this basic skeleton for other purposes, the first of which is a dice-based resource acquisition game. This begins without a theme, but in exploring the mechanics, we start to see how a theme naturally emerges.
Play centers on a 6x6 grid from which you can acquire resources: A, B, C, D, E, and F. The intersections of each row and column show combinations of two resources and double-resources along the diagonal from top left to bottom right.
On your turn, you roll three dice and choose two of those results to be the coordinates from which you will acquire the noted resources. The third die shows how many of those resources you will acquire.
For example, you rolled 2 5 4. You chose to harvest from 2/5, which means you get 4 of resources E and B. If you rolled 4 4 3, you could choose to harvest from 4/4 where there are two Ds. This means you acquire resource D at twice the rate as normal. So, instead of just 3 Ds, you acquire 6.
But towards what end? I'm not sure. Perhaps you are trying to purchase advancements that require a specific recipe of resources, Waterdeep-style? Whatever the case, there are interesting permutations in this system.
1/1's resources can be a little more common than 6/6's resources. The likelihood of rolling 1 1 1 and 6 6 6 are equal. However, a roll of two matching numbers and a non-matching number is much more common. Thus, on a roll of 1 1, it is much more likely that the third result will be greater than 1. Conversely, on a roll of 6 6, it is much more likely that the third result will be less than 6.
Granted, it's a small statistical difference. (EDIT: And, as Levi Middleton points out, D ends up being the more rare resource.)
This still gives me some sense of structure for a theme. Perhaps the A resource is a common ingredient in the game's recipes whereas the F resource is something more rare but valuable, like straight victory points or perhaps wild resources that can be used as placeholders for other resources.
The other interesting facet of this system is that each combination of resources has a twin on the opposite side of the board. 5/3 gives the same stuff as 3/5. So, perhaps there is room for adding another type of resource to acquire, based on which side of the diagonal you choose.
Indeed, this comes to resemble the banks of a river. The river itself is abundant and fruitful. Its banks are blessed with useful combinations of resources while the far corners are dry prairies and deserts with less useful combinations of resources.
When you acquire resources from a space, so you also lay claim to it. In choosing a space that is occupied by another player, they may ask for a "tax" to give you permission to use that space.
Thus, our old friend the area control mechanic plays a significant part in this game. Those recipes I mentioned earlier? Those may be used to purchase advanced settlements that levee taxes on neighboring spaces; or award points to occupants of neighboring spaces; or renders a space unusable thereafter. Who knows?
Anyhoo, this is how my game design process usually begins. I'll notice a curious wrinkle of probability that makes a decent metaphor for a real-world phenomenon. Of course, it's usually at this point that someone will point out a game that has already covered similar territory, usually designed by Reiner Knizia!
But I hope that documenting my thought process is at least somewhat enlightening. Because geez, I just love designing games.