How Black And White Actually Works on Coolmath
It's a logic grid puzzle where you're given a rectangular board and a set of clues. Your job is to shade each square either black or white so that every clue is satisfied. That's the summary anyway. The reality is a bit messier, especially once the puzzles start using diagonal connectivity and region-counting rules that aren't immediately obvious from the title screen. The game sits in the logic/puzzle section of the site. You click a level, you get a board with numbered clues along the edges and sometimes inside the grid. The interface is basic HTML canvas with mouse clicks. No animations, no loading screens between attempts. It just resets when you fail and lets you try again. Most levels I've run through have between 8x8 and 15x15 grids, with difficulty scaling mostly through larger boards and more overlapping constraint types. The core mechanic is straightforward. Each clue tells you something about the arrangement of black and white squares in a row or column. Some clues are simple like "shade exactly three black squares in this row." Others involve runs, where consecutive black squares must match a pattern like "a run of two, then a gap, then a run of three." The hard ones combine region rules with edge clues simultaneously.
I spent a weekend trying to beat the higher-difficulty packs last year. What I learned is that most people approach it wrong at first. They start clicking randomly until something sticks, which works for the tutorial levels but dead-ends around level 20 or so. The actual method is process of elimination using constraint propagation. You look at each clue and ask what configurations are physically impossible, then cross them out until only one remains. It's basically the same logic behind pixel art solvers, except you're doing it in your head on a browser tab. One edge case that tripped me up repeatedly involved what I'll call the "floating island" rule. The game doesn't explicitly state it on the first screen, but in certain advanced puzzles, isolated single black squares that share no edge with any other black square are treated as invalid by the check mechanism. I spent about 45 minutes on one particular level because my solution looked correct against every visible clue, but the game kept rejecting it. The workaround was testing every candidate solution by looking for single-square islands and filling them in. Once I started checking for that condition first instead of last, completion time dropped from roughly 20 minutes per puzzle down to around four or five. Another thing nobody mentions upfront is how edge clues interact with interior clues. Beginners treat them as independent constraints. They aren't. An edge clue about a row immediately restricts every cell in that row, which then constrains the column clues that intersect those cells. Working inward from the edges tends to be more efficient than starting from the middle, especially on boards larger than 12x12. I timed both approaches on the same set of ten puzzles. Starting from edges averaged about six minutes per puzzle. Starting from the center averaged about eleven. That's not a huge difference on easy levels, but on the maximum difficulty pack it meant the difference between finishing a session and quitting out of frustration.
There's no official download because the game runs entirely in-browser. The URL is just coolmathgames.com with the puzzle section selected. You can play it on any computer with a modern browser. No account required, no installation, no ads that interrupt gameplay. The only real limitation is that the mobile version is noticeably worse. Touch targets on smaller grids get cramped, and the hint system doesn't trigger reliably on iOS Safari. If you're playing on a phone, use landscape mode and tap the grid cells rather than trying to click the clue numbers directly. The hint system is functional but sparse. Each hint reveals one cell and costs you a star rating on that level. Most players max out their hints on harder puzzles before realizing that three hints on a single 15x15 board rarely gets you past the point where you'd stuck anyway. A more useful strategy is to use hints only when you've genuinely exhausted all deductions from the current clues and hit a genuine wall. That usually means after ten to fifteen minutes of working a single puzzle, not after two minutes of guessing. Some people recommend printing the grids out and solving on paper. That's actually valid advice for the harder packs. The cognitive load of tracking possible configurations across multiple overlapping constraints is lower when you're writing possibilities in pencil rather than holding them all in working memory while clicking. I switched to paper for the advanced challenge pack and saw my average solve time cut roughly in half within the first week.
Get the Full Details
The game doesn't track progress across sessions very well. Refresh the page and you're usually back at the level select. There's no cloud save or local storage that I could find. This means if you're halfway through a multi-step puzzle and your browser crashes, you start over. Not a dealbreaker, but worth knowing before you invest an hour into a particularly stubborn level.
