I've honed my idea on this more, and while this first "mockup" of sorts (using a few of the ramp pieces I've quickly drawn up in Paint to demonstrate) is missing key elements of the game, such as enemies, items, the holy grail, castles, cooler terrains, etc. it can at least vouch for what the general look of the game will likely be like (minimal color counts and sharp contrasting colors to give a look of what could be found on an Atari arcade table.)
It's admittedly hard to tell different "levels" (the height of the block patforms) apart at times, so I'm thinking of keeping the wall colors static like the are now, but change the colors of the floor patterns as they move progressively higher in the Y direction. The test floor pattern may not be the most impressive, but eventually there will be animated, thematic patterns in most areas of the game.
Not a huge amount of progress code-wise, but in a few weeks I should have a map editor (which WILL ship with the game at a later phase) and a rudimentary physics engine for the ball's movement on the maps, which will be stored mostly as a 3-dimensional "tilemap" of sorts, as I plan to allow for tunnels and double/triple-decker paths, which the original Marble Madness didn't. (It used a 2-dimensional tilemap and drew each tile based on the height attribute and the ramp type.)
EDIT: and due to the more advanced engine I'm planning, it would be excruciatingly hard to cast realtime shadows like in Marble Madness, so that's not going to be present. I could always draw them in manually onto the maps, but I plan to allow for user creation of maps at a later time, so that would require realtime projection, so unlikely to happen.