What a cross-math puzzle actually asks
A cross-math puzzle is a grid of cells joined by arithmetic. Some cells already hold numbers; others are empty. Around the board sit operators — plus, minus, times, divide — and equals signs that tie short chains of cells into equations. You are given a small tray of number tiles, and your task is to place every tile so that each equation on the board reads true at the same time.
The key word is "at the same time." A tile that satisfies one equation has to also fit every other equation it touches. That overlap is what turns arithmetic into a logic puzzle: the right number for a cell is rarely a matter of taste, it is forced by everything around it.
Step one: read the board before you touch a tile
Before placing anything, scan the whole board and find the equation with the fewest unknowns. A chain like 7 + ? = 12 has exactly one answer. Solving it first is free information — it costs you no risk and immediately constrains the cells nearby.
Lattice rewards this habit. Because every board is built to be solved by deduction, there is always at least one cell whose value is already pinned down by what you can see. Find that cell, and you have your foothold.
Step two: let each placement cascade
Every number you place changes the puzzle. The moment you fill a cell, re-read the equations that touch it: one of them may now have a single unknown that it didn't before. Solve that one next, then the equations it touches, and so on. Good solvers don't hunt all over the board — they follow the chain reaction outward from the cell they just filled.
If you place a tile and a connected equation becomes impossible — say it now demands a number that isn't in your tray — that placement was wrong. Take it back and try the alternative. On a well-formed puzzle you will never have to do this blindly; the contradiction tells you exactly which assumption to undo.
Step three: use the operators as clues
Operators narrow the field faster than the numbers do. If a chain multiplies to a small product, the factors are limited. If a difference is large, one of the two cells has to be large and the other small. Division is the strongest clue of all: a ÷ b that must come out whole tells you b divides a, which often leaves only one candidate in your tray.
Train yourself to ask "what does this operator forbid?" rather than "what could go here?" Forbidding candidates is faster than testing them, and it is exactly how the puzzle was designed to be read.
Step four: count your tiles
Your tray is finite, and that is a clue in itself. If three cells still need filling and your tray holds a 3, a 6, and a 9, then those three numbers go into those three cells — the only question is which goes where, and the equations decide that. Keeping an eye on what's left in the tray often collapses a daunting board into a handful of forced placements.
The one rule: never guess
If you find yourself thinking "I'll try this and see," stop. Every Lattice puzzle is machine-proven to have exactly one solution, which means a guess is never required — the next move is always deducible from what's on the board. When you're stuck, you haven't run out of logic; you've missed a clue. Re-scan for the equation with the fewest unknowns and start the cascade again.
That discipline is the whole pleasure of cross-math. Solve by deduction and every win is earned, not lucky. Ready to put it into practice? Play today's puzzle, or read why every Lattice is guaranteed to have a single answer.