Some random notes:
My last few levels were 83, 89 and finally 81. Level 81 wasn’t really that hard, but I though it was going to take longer than it actually did.

IF I RECALL CORRECTLY, level 94.

I thought Level 100 was going to be the granddaddy of them all. I mean really… three pieces? It was no walk in the park, but it certainly wasn’t the most difficult.

I’d be interested in hear which levels others thought were the most difficult.

Brainstorms for Blockhouse2:
– Multiple disconnected areas like 83, but with weirdly shaped pieces, a la 100.
– A few levels with the exact same board and piece starting positions but different goal positions.

It does make sense to change orientations when you’re stuck on a puzzle. Changing your perspective helps you break out of ruts and find positions you haven’t found before.

Maybe it has nothing to do absolute orientation… just that it’s not the way I’ve been banging away at it for the last ten minutes.

Anyone else experience this?

Really great game. rec.*.i-f takes me WAY back, although I still hang out with some of those folks.

Hint for puzzle 83: The general progression of the solution is to put the blue block on its target, then the purple block, then the red block. But the blue block will need to come off its target once and then go back on it in order to make this work.

Let me know if that’s too obscure…

As my friend Dan said today, “This is pretty much the best dollar you ever spent, right?” Yep.

As far as the “themes” go, the puzzles just naturally fell into four categories (single-square, single-poly, segmented, double-poly), and each type felt different. As far as the individual puzzles go, it was sort of overwhelming to try to come up with 100 of them, so I made my job easier by coming up with a pattern which I more or less repeated within each quad. For instance, I tried lots of combos of shapes for the double-block puzzles, and I narrowed it down to 10 combos that I really liked. I repeat these 10 puzzle types in each quad, getting gradually harder. That structure helped me know what to work on next. I’d pick one puzzle-type (say, two interlocking L-shapes), and I’d design at least a dozen of them, and then I’d pick my four favorites, and order them by difficulty.

As far as designing individual puzzles, I tried to make them as simple as possible while remaining interesting. For instance, I’d take the two interlocking L-shapes blocks, and I’d try putting just a single square wall somewhere in the puzzle. In a 6×8 puzzle, there are only 48 possible places to put a single wall (and really fewer, when you consider rotations and reflections), so I tried them all exhaustively. Then I’d try exactly two walls, and exactly three walls, etc. Even here, I felt like I was almost exhaustively exploring the puzzle space, because my evolutionary process was trying thousands of wall positions, and only showing me the ones that were interesting. In the end, this turned out to be enough. If you go back and look at all the double-block puzzles, you’ll see that they mostly contain four or fewer wall blocks. That was all the complexity I needed.

The other puzzle types were a bit different, but followed a similar pattern. The single-poly puzzles were exactly like the double-poly puzzles, except I needed to add a lot more than four wall blocks to get interesting puzzles. With the segmented puzzles, I started with a simple wall separating the segments, and then had the evolutionary process add only a handful of walls. For a lot of them, I added the walls symmetrically, which created segments that were mirror images of each other. Those actually seemed to result in particularly interesting puzzles. The single-block mazes were the hardest puzzles to design, because my evolutionary process was not good at evolving them. In the future, I’d like to come up with a better process for this.