Fill The Grid
Nine cells. Name a musician who fits both the row and the column. One guess per cell, and a new grid every day.
Loading today's grid.
How Fill The Grid works
Each grid has three constraints down the side and three across the top. Every cell is the pair of the two that meet there, so the top left cell wants a musician who satisfies the first row and the first column at once.
A constraint is one of four kinds: a genre, a country, the decade a musician debuted in, or whether they are a band or a solo artist. They come from the same MusicBrainz data the main Musicianle board is built from, so a musician counts for a genre only if their own release history is tagged with it.
You get one guess per cell, nine in total. A wrong guess closes that cell for the day. A name we do not recognise is not a wrong guess, it just does not count, so a typo costs you nothing but the retype.
Every cell has at least one valid answer. That is checked by the generator before a grid is ever scheduled: a grid where any cell was unsolvable is thrown away rather than published. It does not mean every cell is easy. Some have one valid answer out of thousands of musicians, and those are meant to be the ones you lose.
The percentage on a filled cell
When enough people have filled a cell, each filled cell shows the share of them who named the same musician as you. A low number means you found someone most players did not think of, which is the part of a grid worth comparing: two people with the same score can have filled the board completely differently.
It is often not there, and that is deliberate rather than broken. A percentage worked out from two or three plays is a number that looks authoritative and tells you nothing, so none is shown until a cell has enough fills behind it. Today's grid usually shows none. Grids in the archive keep gathering plays because they stay playable, so they fill in over time.
Is there a right answer?
Usually there are many, and that is the point. Two people can fill the same cell with different musicians and both be right, which is what makes the grid worth comparing rather than just scoring. The share text records how many cells you filled and nothing about who you named, so posting a result never spoils a cell for the person reading it.
How a grid is proved solvable before you see it
Six constraints crossing into nine cells is easy to generate and easy to get wrong. A grid is only published once every one of its nine cells has been checked to have at least one musician in the corpus who satisfies both the row and the column, so a cell you cannot fill is a gap in your knowledge rather than a grid that was never possible.
It does not mean every cell is easy. A pairing can be technically valid and still have only two or three musicians behind it anywhere in the data, and those are the cells that decide a good grid from a lucky one. What the check removes is the other failure, the one where a player spends their single guess on a cell that no answer could ever have filled.
Grids with too many of those narrow cells are refused as well: no more than three of the nine may have a single possible answer. A grid mostly made of one-answer cells is not a puzzle, it is a lookup, and the generator throws that draw away and tries another rather than publish it.
Why a wrong name costs you nothing
A name the game does not recognise is not counted as a wrong guess. You are told we do not have a musician by that name and asked to try another spelling, and the cell stays open.
That rule exists because the alternative punishes the wrong thing. The data holds a little under seven thousand musicians, not every musician, and it stores one canonical spelling for each. Without this you would lose a cell to a typo, to a diacritic, or to knowing an artist by a name the data files under something else, none of which is the game being hard. A guess should cost you a cell when you named the wrong musician, not when you named the right one in a way the lookup did not expect.
Names are matched the same way the main Musicianle board matches them, including aliases and ignoring case, punctuation and accents, so an artist you can find in one game you can find in the other.
Where the data comes from
Genres, countries and debut years come from MusicBrainz and Wikidata, used under CC0. Nothing is hand-written for a puzzle, which is why a constraint occasionally looks stricter than you would expect: the tag on the musician is what counts, not the way they are usually described.