One of my earliest designs, born from a brainwave on creating hexagonal cells from a square grid:
One uncut square, designed in 2007.
I was so excited by my idea and immediately rushed out a crease pattern based on it:
Turns out I wasn't completely right - I had allocated one unit square for each border face of each hexagonal cell, but the angles between them were not the ideal 120°. In fact, a strict folding of this crease pattern produces a "brick wall" rectangular tessellation, which can be obtained by compressing the folded model. I did not expect the resistance of the paper to expand the folded model slightly, correcting the angles towards the ideal hexagonal shape. The structure of the model, with its springy layers, made it very flexible. The slideshow on top shows various flexing positions.
An example of how a half-baked idea can lead to something unexpected and interesting even if it doesn't achieve its original goal.