When does a review add the MOST storage strength?

When you've partly forgotten it: a successful, effortful retrieval at LOW retrieval strength produces the largest gain in storage strength.

This is the counter-intuitive core of Bjork's framework. If you review something you already recall fluently (high retrieval strength), you add almost no storage strength — there was nothing to struggle for. But if you've let a memory fade and then manage to pull it back, that hard-won retrieval wires it in deeply. As Bjork's work implies: the struggle is the mechanism, not a side effect. Difficulty during recall is the price of durable learning — so a little forgetting before review is a feature, not a bug.