Every family in Ponder had to earn its place. For each one that made it, we tried and set aside others. Here’s what we look for in a puzzle, why some types didn’t make the cut, and which ones we’re exploring next.
What we look for
It’s fun from the first move. You should be able to learn the rules in under a minute and find something to do straight away. A puzzle that opens with five minutes of staring before anything happens loses most people before it gets interesting.
It’s quick to finish. A board should take minutes, not an evening, and a Daily set of six should fit comfortably into one sitting. We would rather you finish a puzzle and want another than abandon one halfway.
It’s challenging without being slow. The hard part of a good Ponder puzzle is seeing the right move, not holding a dozen possibilities in your head. Our hardest boards should still let you keep moving once you spot what’s going on.
It has a high skill ceiling. This one matters most for Race. If every player finishes a puzzle in about the same time, a race between friends comes down to luck. The families we pick reward practice: the more you play, the more patterns you recognise and the faster you get. That keeps a race interesting whether you started last week or last year.
It works on a phone. Each move should be a tap or a short drag, and the board has to stay readable on a small screen without zooming.
It can be explained. Every board needs exactly one solution, reachable by reasoning, and a hint that can explain the next step in a sentence or two. If we can’t write that explanation, the board doesn’t ship, and if that keeps happening, neither does the family.
What didn’t make it
Most of the puzzle types we’ve set aside fell into one of three groups.
Too close to something we already have. Each family in Ponder asks you to do something different: balance symbols, divide a board, trace a route, light a grid, place flowers, join pairs, connect islands, fill in numbers. A puzzle that plays like one of those under a new name adds a menu item without adding a new way to think.
Too much to keep track of. Some puzzles give each cell several things to manage at once: a number, a colour, a direction, a running total. That can be wonderful on paper. On a small board on a phone, it turns into bookkeeping, and the fun drains away.
Too much thinking per move. There are brilliant puzzles where every step means following a long chain of what-ifs. They suit a quiet afternoon with a pencil, but they stall a race and they aren’t what most people want from a short break. Ponder’s puzzles are meant to be challenging and still let you move quickly.
How a new family gets in
A candidate starts as a handful of small boards we solve ourselves. If it’s fun and passes the tests above, we build two things that work independently: a generator that creates boards, and a solver that checks each one has exactly one solution. Then come the hints, which must explain real deductions rather than just reveal answers, a Learn to play board and a guide.
A family only joins the Library once it has a full set of checked boards: at least a hundred at each difficulty. Bridges and Mini Sudoku came through this process; they were on our candidate list from the start and joined Ponder in October.
What we’re exploring now
Nonograms, where number clues along each row and column tell you which cells to shade, and the finished board reveals a picture. The question for us is whether boards small enough to finish quickly on a phone can still be interesting.
Slitherlink, where you draw a single loop around a grid of numbers, each number saying how many of its sides the loop uses. It has an enormous skill ceiling. The question is whether it can stay fast.
Something beyond the grid. Ponder doesn’t have to be all grids. We’re also looking at sequence, word and graph puzzles that would give the Daily set a different texture.
None of these is guaranteed. Each one has to pass the same tests as everything already in Ponder, and some of them won’t. You can follow along on the roadmap.