Step one: the long list
Candidates come from three places: reader notes in the post tray, the newer arrivals on CrazyGames, and games one of us noticed a friend or relative playing on the bus. For the autumn run that gave us twelve titles. Before anyone plays, Anjali reads each CrazyGames page and writes down what it claims about controls, saving and supported devices, so we can check those claims later rather than repeat them.
We set a game aside at this stage if its listing points towards buying anything, if it borrows the look of real-money gaming, or if it simply lacks a clear description of how to play.
Step two: the pink pass
One player, one device, one evening, with no reading beyond the game page itself. The pink pass imitates the way most people meet a browser game: tired, curious and not inclined to study. The player writes three short notes before stopping: what they thought the goal was after sixty seconds, the first moment they felt confused, and the moment they first wanted to keep going.
Those notes are filed and sealed until the blue pass is done. We do this so the second player cannot be nudged by the first player’s mood.
Step three: seven days of rest
Riso ink smudges if you stack sheets too soon. Opinions do too. We leave at least a week between passes, and the person who did the pink round does not play that game again in the gap.
Step four: the blue pass
A second person plays across several sittings on a different kind of screen: if the pink round happened on a laptop, the blue one happens on a phone or tablet, and the reverse. This pass is where we learn whether progress survives a closed tab, whether the game tolerates a sudden pause, how warm the device gets after half an hour, and whether the loop deepens or just repeats.
Step five: lining up the layers
We meet, unseal the pink notes and compare. Each pass is marked out of ten on the same five headings. The overprint is a weighted blend, forty percent pink and sixty percent blue, rounded to one decimal place. If the two passes differ by more than a full point, the card says why in its verdict line.
The five marks
First-minute clarity, 25%
We want to understand the goal through play, not through a wall of text. Low marks go to games that open on a long tutorial, or that hide what the controls do.
Hands and controls, 20%
Does the input suit the shape of the device? Holding a mouse button to dig, dragging a hook, walking with WASD: each earns its mark only if it feels natural on the screens the game says it supports.
Progress that sticks, 20%
We close the tab, wait, reopen it, and note what came back. All five games in this run list cloud saving on CrazyGames, and we check that it held for us.
Easy to set down, 20%
A good browser game lets you leave in the middle of something. We deduct for rounds that punish a pause or nag you to come back.
Weight on the device, 15%
How long it takes to load on a modest connection, whether the fan spins up, and how much battery a twenty-minute sitting costs on a phone.
What we do not score
We give no marks for how famous a game is, how many people rated it, or how polished its trailer looks. CrazyGames shows its own player ratings; we read them, but they never enter our arithmetic.