Exercise 3: Finding the 90% Threshold — Possible Solution
====================================================================

METHOD
------------------------------
Using the chapter's own simulation logic (known velocity = 20.0,
10-day sprint, real throughput fixed at 2.0 points/day regardless of
what's committed), incrementing the committed-points value from 20
upward and checking the completion rate at each step.

RESULTS
------------------------------
    committed=20: completed=20.0, rate=100.0%
    committed=21: completed=20.0, rate=95.2%
    committed=22: completed=20.0, rate=90.9%
    committed=23: completed=20.0, rate=87.0%  <-- first below 90%

The completion rate first drops below 90% at a commitment of 23
points - just 3 points (15%) above the team's own real velocity of 20.

WHY THIS WORKS AS AN ANSWER
------------------------------
This is a genuinely useful, actionable number: it shows the
completion-rate metric is quite sensitive to overcommitment - a
relatively modest 15% overcommitment (23 vs. 20) is already enough to
drop a team's own visible success rate below the commonly-used 90%
"healthy sprint" threshold many teams informally track. Since actual
completed work never changes in this simulation (it's capped by real
throughput, not by what's written on a board), this also reinforces
the chapter's own central point from a new angle: even a fairly small
gap between committed and actual capacity is enough to make a
perfectly normal sprint look like an underperforming one, purely as an
artifact of the number chosen at planning.