How falsewrk works

falsewrk trains one thing: the engineering judgement that tells you a number is wrong before you have checked it. Not by teaching you more theory, but by giving you the reps that judgement is built from — on purpose, every day, in a few minutes.

The loop, in one paragraph

You are shown a structural problem — a slab, a beam, a column — and asked for a rough number, fast. You answer, and the reveal shows the true value, how far off you were, and the method that gets there. That is one rep. Each problem type comes back on a schedule that stretches as you get it right and closes up when you don’t, so your time goes where your judgement is weakest. When a type comes back after a week or more away, its first rep is a cold read — a measure of what you kept, not just what you can do while it is fresh.

Where to go next

• Why estimate? — what the reps are for, and why speed matters.
• Reps and grading — what counts as excellent, acceptable and a miss, and why the sign of your error matters.
• The science — the research on forgetting, spacing and testing that the schedule is built on.
• The schedule — exactly how falsewrk decides what you practise today.
• Practice and Retained — the two numbers that tell you whether it is working.
• Compared with Anki — what falsewrk borrows from the best-known spaced-repetition app, and where it parts ways.
• Reading your stats — the Stats page, chart by chart.