Building a big set, in stages.

The search is the cost. On a set of any size, the time is not going on assembly — placing a brick takes about a second and a half — it is going on finding the brick, and finding scales with the size of the pile in front of you. Which is why tipping a numbered bag straight onto the table is the most expensive thing you can do to an evening: you have just made the haystack as large as it can possibly be.

Sort by shape first and colour second. A tray holding every 1×2 in the bag can be worked by touch; a tray holding everything blue has to be read piece by piece, because blue is not a shape and your hand cannot feel it. The exception proves it — a roof bag or a baseplate frame is mostly one colour anyway, and there colour and shape are the same sort.

Sort no further than the bag you have open. The designer has already done the coarse sort for you and put a number on it; re-sorting the whole box into a rainbow before bag one is a different hobby, and a good one, but it is not this build.

Then build in the units you were given. A numbered bag is a sitting: it has a beginning, an end, and a model that stands up on its own when you get there. Stopping halfway through one is what leaves you, a week later, with a tray of unlabelled parts and a page you cannot remember starting — and that is the pause most unfinished sets never come back from.

The rest is small and unglamorous. Work over a shallow tray with a lip, never a carpet. Turn the model, not your head, so the page and the piece agree about which way is front. Check the step before you turn the page: a brick placed one stud out is always found four steps later, and it costs the four steps back, not the one.

And keep the spares. The factory adds an extra of the tiny parts because tiny parts go under sofas, and the day you want one is the day the set has been out of production for three years. They go back in the box, in their own bag, with the instruction book — which is the other thing worth more than people think.