`Array.includes` is ~28× slower than `Set.has`

If you test membership repeatedly, convert the array to a Set. Over 200,000 lookups the gap reaches 28×.

Result

Run	`Array.includes` (100k array)	`Set.has` (same contents)	Ratio
Run 1	462.040	15.700	29.43×
Run 2	446.242	14.833	30.08×
Run 3	460.714	17.797	25.89×

Median spread: `Array.includes` (100k array) CV 1.6% / `Set.has` (same contents) CV 7.7% / Ratio CV 6.5% (computed from the 3 runs)

Conditions

200,000 membership tests against a 1,000-item set.

Where this number does not apply

For a few dozen lookups, building the Set costs more than it saves. Keep the array for one-off checks.

Reproduce

cd bench
node --expose-gc micro-fixed.mjs --out ../data/run.json --samples 21

Run it three times on the same machine; if the spread of median stays under 10%, you should land on the same conclusion. If your numbers disagree, tell us and we will retract.