Exercise 3: A Delayed Discovery — Possible Solution
====================================================================

RESULTS
------------------------------
    Caught at 1st checkpoint (chapter's own baseline, after 5 features):
      5/20 rework (25%)

    Caught at 2nd checkpoint (after 10 features):
      10/20 rework (50%)

Rework doubled exactly in proportion to how much extra work was built
before the mistake was caught - the second checkpoint discovery point
sits at feature 10 instead of feature 5, and the rework count matches
that position exactly (10 rather than 5).

WHY THIS WORKS AS AN ANSWER
------------------------------
This confirms the real lesson isn't "have checkpoints" in the abstract
- it's "how much work accumulates before the FIRST checkpoint that
actually catches the problem." A team that checks in every 5 features
but doesn't ask the right question, or doesn't get real feedback, at
the first checkpoint gets no benefit from having technically "done" an
iteration - the cost is determined by when the mistake is actually
caught, not by how often check-ins are nominally scheduled. This
matters in practice: a sprint review that doesn't genuinely surface a
wrong assumption (because the demo shows the feature working in
isolation, not integrated with the real system) provides the
appearance of iterative feedback without its actual protective value -
exactly the gap between "having ceremonies" and "the ceremonies doing
their job" that later chapters on Scrum return to directly.