How do I account for unknowns in my estimates?
You estimate better than you used to, but unknowns and scope creep still catch you, because those are the two hardest things in the whole discipline. You are not failing at estimation. You are bumping into the part of it that humbles everyone: you cannot size precisely what you cannot yet see. The trap is treating your estimate as a promise about a known thing, when much of the work is still fog. The unknowns are not a flaw in your estimate; they are a feature of the work, and the honest move is to price them in rather than pretend they are not there. Scope creep hurts most when the original scope was never named clearly enough to defend. Try this. On your next estimate, name the parts you genuinely do not understand yet and add explicit time for them, and write down what scope you are assuming. Making the unknowns visible, on paper, is what protects you from the quiet gap between what you guessed and what was actually there.