Small Steps
also Step-by-step Growth · Manageable Steps
Growing by tackling tasks only slightly beyond your current ability, so each step succeeds and your error rate stays manageable, rather than jumping into the deep end or staying in the shallow end.
Small steps is CF’s prescribed learning method: identify your baseline (what you can already do and check with confidence), then advance one increment at a time so each attempt succeeds. CF’s rationale is rooted in error management. A step is “small” if you finish it with only a couple of errors; twenty errors means it was too big. The point is to keep your error rate below your error-correction capacity — taking a step too large is exactly overreach, where errors accumulate into an unsolvable backlog and you get stuck.
CF stresses that difficulty scales worse than linearly: handling three new things at once is more than three times harder than one, because the interactions multiply. So you isolate a single new variable — like walking with one shoe, where most of your existing walking knowledge carries over. Elliot Temple urges attempting things at most ~20% harder than a previous success, and growing maximum project size roughly 10% per success.
Three things make small steps work: they must be of multiple types, build on each other, and include practice with objective success/failure judgment. Aim for a success rate above 90%; on a failure streak, drop difficulty exponentially (a third, then a ninth) to recover.
This connects CF to Theory of Constraints: just as ToC says do not run a balanced plant at full capacity, small steps leaves a buffer of spare error-correction capacity rather than maxing it out. The opposition is the common pattern of chasing impressive, advanced work to seem smart — which CF diagnoses as a recipe for getting stuck.