About
Carry the One is a small, free mental arithmetic game for the web. It exists because most maths practice apps are either aimed squarely at young children or buried under accounts, streaks and upsells. This one is a timer, four buttons and an honest score.
What it is for
Arithmetic is one of those skills that decays quietly. Rounding a bill, sanity-checking a number in a meeting, working out whether a discount is worth it — all of it gets slower without practice. A few minutes of drills against a clock is a good way to keep it sharp, and playing against other people is a better way to keep coming back.
How the game is built
Every question is generated fresh rather than pulled from a fixed list, so the same question rarely comes round twice. In a room, all four players’ devices generate the identical sequence from a shared starting value, which is how everyone sees the same question at the same moment.
Scores in a room are worked out on the server rather than on anyone’s phone, because a wrong answer moves other people’s totals too.
A note on the leaderboards
Solo scores are reported by your own browser. Treat the boards as friendly rather than official — they are there to give you something to beat, not to settle arguments.
Who made it
Carry the One is an independent project, built and run by one person. It is free to play and is supported by advertising on the menu and results screens. There are no ads during a game, and there never will be — a timed round with an advert in the middle is not a game worth playing.
Get in touch
Bug reports, ideas and complaints are all welcome. See the contact page.
Why four options instead of typing
An early version asked players to type the answer. It was worse in a way that took a while to understand: typing measures your thumbs as much as your arithmetic, and on a phone under a five-second clock it measures your thumbs considerably more. Four tiles keep the test on the part that is actually interesting.
It also makes estimation a legitimate strategy, which turns out to be closer to how mental arithmetic works in real life. You rarely need the exact figure — you need to know which of the plausible figures is right.
Why the wrong answers look like that
The three decoys are not random. Each question generates near misses — the answer plus or minus one or two, plus or minus ten, and a value about a tenth away. These are the answers you would arrive at by making an ordinary slip: dropping a carry, misreading a digit, being one out in a times table.
Random decoys would make the game far easier, because the right answer would stand out by being the only plausible one. Making the wrong answers plausible is what forces you to actually finish the calculation.
Why the clock shortens instead of the numbers growing
Both would make the game harder, but they train different things. Larger numbers train method. Less time trains retrieval — turning a method you know into one you reach for without deciding to.
The game does a little of the first and most of the second, because retrieval is the part that transfers. Knowing that 15% is “ten per cent plus half again” is worth nothing at a restaurant table if it takes you eight seconds to get to it.
What it deliberately does not do
There are no streaks to protect, no daily reminders, no currency, no levels to unlock, and nothing to buy. Those mechanics are good at manufacturing engagement and bad at producing practice — a streak makes people open an app, not concentrate once it is open.
The one social mechanic is rooms, because racing someone you know is the only reliable way anyone practises anything, and it stops the moment you close the tab.