Exercise 1: Raising the Fixed Wait to 80ms — Possible Solution
====================================================================

RESULTS
------------------------------
    80ms fixed-wait test over 200 runs: 156 passed, 44 failed
    Pass rate: 78.0%

    (30ms version, same random seed, for comparison: 47/200 passed, 23.5%)

WHY IT'S HIGHER BUT STILL NOT 100%
------------------------------
The simulated load time is uniformly distributed between 10ms and
100ms. Waiting 80ms covers roughly 78% of that range (from 10ms up to
80ms is 70 of the full 90ms spread, closely matching the measured
78.0% pass rate) - but any load time between 80ms and 100ms still
exceeds the fixed wait and still fails. Raising the wait narrows the
window of failure; it doesn't close it, because the wait is still a
guess at an upper bound rather than a check of the actual condition.

WHY THIS WORKS AS AN ANSWER
------------------------------
This demonstrates the core problem with fixed waits precisely: any
finite fixed wait either has to guess so generously that it slows down
every single test run to cover a rare worst case, or accept some
nonzero flake rate for whatever fraction of real-world timing exceeds
it. The chapter's own polling-based fix sidesteps this entirely by
checking the real condition rather than betting on a duration - which
is why it measured 100%, not just "a higher percentage," across the
same 200 simulated load times.