I'm very confused but impressed. well done aha!

How I actually check if the stack falls over:
- Find the boundary between layers where a layer rests on the one below (this is the 3D surface where they both have pieces).
- If the boundry is empty, the layer above isn't resting on anything, so it falls and give I an alert (the alert text may be wrong, now that I think about it).
- Otherwise, test whether the center of mass of the pieces above the boundary lies above the convex hull of the boundary.
- If the center of mass (projected down into 3d) is outside the convex hull, it falls and some triangle on the surface of the convex hull separates the convex hull from the center of mass - this defines a plane about which the pieces above can rotate (in 4d, things rotate about 2-dimensional planes, in the same way things rotate about 1d axes in 3d).
- I also assume that the tower can tilt a little, so if the center of mass is close enough to the convex hull of the boundary, it will fall.

I had trouble with the visualization; after a certain rotation point the graphics seemed to "pop" so it is difficult for me to wrap my head around. Perhaps I just need to think more about this, I dunno. Anyways, I just took the dark colored blocks and it seemed to work. Cool game

My brain is too smooth for this :(

What an innovative idea! I agree, my brain is definitely too smooth to properly figure out what's going on, but it's an interesting, mind bending idea. The visualisations maybe could have been clearer (a wider array of different colours so you can match up blocks easier, as well as maybe some depth in the 2d view?).

this is highly confusing. Nice job implementing it tho ahah

@natpat I didn't do any shading to make it easy for to tell what was clicked on (the program inspects the color under the cursor). If I make an improved version, this will be one thing I change.

Me: It's time to play jenga on paper.

Gf: I don't think that's how you play jenga.

Great concept, very mind-bending. After solving 3d tower on paper I managed to remove 2 tiles from each layer with some intuition on how to extend 2d case to the 3d case and I think this is the most pieces you can remove, but I don't really have a mathematical proof that this is optimal.

@domantas I'm glad you took it seriously, I think you're right about 2 tiles from each layer being the maximum that can be removed. I assume it wobbles by up to 0.05 radians from the vertical, which is why removing a pair of parallel pieces in a layer doesn't usually work, but with a smaller wobble, you could get the center of mass somewhere that would let you remove 3 pieces from a layer.

@tesseract Well shit, I thought that there must be a piece left in all halves of the cube for the thing to be stable, but from your comment it seems that I had a completely wrong idea.

This was a hard one to visualize, perhaps some shadows would help with this. This is a great puzzle, but really felt like more of a prototype. Still, required some brain power to figure it out, great job!

I don't understand. I tried to rotate the blocks around, remove them in different order. Sometimes it looked like the structure was about to fall, or fell, but there's no feedback really. I guess I just didn't get it :(

If this is the real 4D, it's cool. It's fun to watch the tower falling out of over dimension.

I must admit, it is confusing, but it is also quite impressive, what you created. Good job!