Playing Chain with a child

Chain works well with a child sitting next to you, and badly with a child watching you solve it. The difference is not the puzzle, it is who is allowed to touch the tiles and who is allowed to be wrong. Below is what we have found actually works: when a child is ready for it, how to run the ten minutes, what to say when they stall, and how to tell the difference between a hard day and a bad one. If you have not played it yourself yet, the rules take two minutes.

When it starts to work

Age is the wrong question. The thing that decides it is whether two-digit addition is comfortable — not fast, not automatic, just something the child will attempt without dread. A Monday puzzle asks for three moves on numbers like 79 + 54, so a child who still counts on fingers past ten will spend the whole session on the arithmetic and none of it on the puzzle. In practice that puts easy days somewhere around eight or nine years old for many children, and it varies by a year or two in both directions, which is another way of saying: try one and watch their face.

There is a second, quieter requirement. The child has to be able to hold a wrong answer without it meaning something about them. Chain resets the board when a chain misses, and that happens several times per puzzle even for adults. A child who treats the reset as a verdict will not enjoy this, and no amount of encouragement changes that in one sitting. Start on a Monday or Tuesday, when the puzzle is three moves, and let the first few days be easy on purpose.

If arithmetic is the barrier rather than the puzzle, there is nothing wrong with doing the sums for them out loud while they choose the tiles. "You said plus fifty-four, so that is one hundred and thirty-three" keeps the thinking in their hands and takes the mechanical part off the table. You can hand it back a few weeks later.

Sit beside them, not behind them

The failure mode is obvious in hindsight and very easy to fall into: you see the answer in fifteen seconds, and everything you say after that is a hint with a lid on it. The child can hear it. Even "hmm, interesting" turns into information about whether they are on the right track, and once the child is reading your face instead of the board, they have stopped solving.

What we do instead is play out loud, both of us, as two people who do not yet know. Say your own half-thoughts, including the ones that go nowhere: "I keep wanting to use the sixty-five, but then I am at one hundred and forty-four and I need sixteen more and there is no sixteen." That last part matters. Children rarely hear an adult narrate a dead end, and hearing it makes dead ends feel normal rather than shameful.

Read the board aloud at the start, every time. Start seventy-nine, target one hundred and sixty, three moves. Saying it converts the screen into a problem you are both facing, which is a different room than one person being tested.

Let them drive

They hold the mouse, or the keyboard. Every time. If you take it "just to show you something", the session is over even if it continues for another ten minutes.

This is easier than it sounds because the board makes wrong moves cheap. Undo takes back the last move and costs nothing. Reset puts everything back and costs nothing. A tile that cannot be played — a division that is not exact, a move that would take the value out of range — is simply refused with a reason, and it does not count as anything. The only thing that costs a try is playing the full number of moves and not landing on the target, and a try costs nothing either except a number on the screen.

Tell the child that. Explicitly, at the start: nothing on this board can be broken, and pressing a tile to see what happens is a legitimate way to think. That single sentence changes how most children play. Trying a tile because you are curious is a real strategy — the running value tells you something you did not know, and undo puts it back.

The keyboard is worth showing to a child who likes keyboards: the number keys 1 to 6 play tiles by position, and Backspace undoes. Some children who freeze at a mouse will happily hammer number keys.

What to say when they are stuck

The instinct is to give the next move. Nearly anything else is better. What has worked for us:

• "How far are we from the target?" This turns a stuck feeling into a number. From 79 you need 81 more, and now the hand can be searched for something specific instead of stared at.
• "What is the biggest number we could possibly reach?" On the easy puzzle above, the hand is +65 −28 +55 +54 −36, so the three largest additions give 79 + 65 + 55 + 54 = 253. The target of 160 is inside that, so no trickery is needed. On a day when the target sits above everything the additions can reach, this question is the whole solution: a multiplier or the swap tile must be in the set.
• "Which tile do we definitely not need?" Ruling one out is much easier than choosing one, and it is the same work done backwards.
• "What number would have to be on the board just before the last move?" For a +55 tile that is 160 − 55 = 105. This is the single most useful habit in the game and it transfers to real maths; the strategy guide works through it properly.
• Nothing at all. Count to thirty in your head before speaking. Most stalls resolve themselves inside twenty seconds, and the child gets to keep the solve.

If none of that moves it, narrate one fact rather than one move: "we have used two tiles and we are at one hundred and thirty-three." A fact is not an instruction, and the child can still be the one who sees it.

Why "exactly three moves" is the good part

Adults sometimes read the move count as a restriction bolted on to an arithmetic drill. It is the reverse — the move count is what makes it a puzzle at all. Without it, you would grab the nearest useful tile and stumble to the target, and there would be nothing to think about.

Watch what it does on the easy puzzle. Start 79, target 160, three moves, hand +65 −28 +55 +54 −36. Playing +65 gets you to 144, which feels like progress: you were 81 away and now you are 16 away. But there is no tile worth 16, and you have two moves left, so 144 is not progress, it is a trap made of a good feeling. The right chain goes further away first — +54 to 133, +55 to 188, which overshoots by 28, and then −28 lands exactly on 160.

That is a genuinely hard idea to teach with a worksheet and an easy one to feel here: sometimes you have to go past the thing you want. It is worth naming out loud when it happens. And it is worth trying +54 +65 −36, which ends on 162 — two away, three moves used, and still a miss. Chain does not give credit for close, which is either the cruellest thing about it or the point, depending on the day.

The swap-digits tile has a similar flavour on harder days. On one September puzzle you start at 41 and need 386 in four moves, and the four addition tiles together only take 41 as far as 202. No amount of adding gets there, so the digits have to be turned around: +92 makes 133, the swap makes it 331, and +40 and +15 finish. Children usually spot the swap faster than adults do, because they are not yet convinced that arithmetic has to be sensible.

When to stop

Frustration is not learning. It is worth being unsentimental about this, because the temptation to push through one more try is strong and it is usually about the adult.

Signals we take seriously: the child stops talking; they start playing tiles fast and without looking, which is the board equivalent of shrugging; they ask you to do it; they get the arithmetic wrong on sums they got right ten minutes ago. Any one of those means the session should end within a minute or two, and it should end without a lecture about persistence.

The graceful exit is the archive, which holds every past puzzle. Say "this one is a Friday, they are meant to be horrible — let's do Monday's instead", open an easy day, and finish on something they solve. Ending on a solve is not a participation trophy; it is the thing that makes them willing to open the site tomorrow.

Also fine: solve it together, openly, with you doing most of the work and saying so. "I will find this one, you watch, and we will do tomorrow's properly." Children are not fooled by a fake solve, but they are perfectly happy with an honest one.

A shape that fits ten minutes

One puzzle a day, on a day that suits the child rather than the calendar. Monday and Tuesday are three moves, Wednesday, Thursday and Sunday are four, and Friday and Saturday are five and genuinely hard — that is the same for everyone, so a child playing at school will have had the same board as their friends. The timer runs, but nothing depends on it, and it is worth saying out loud that the clock is only there for people who want it.

There is no sign-in. Progress is kept in the browser on that device, so nothing about the child is collected in order to play, and there is no account to make before you start. If a puzzle looks wrong, the feedback button on any page reaches us, and the FAQ covers when the day changes and what the site stores.