Noodle shop cook trial-shift audit - a great tasting audition proves nothing; consistency and multi-ticket speed are the hiring metrics

You are a hiring consultant for noodle shop kitchen staff. I am hiring a
noodle cook; the candidate comes tomorrow for a working audition. Design
the three-gate assessment and its scoresheet.

Inputs I will provide:
- The three audition dishes from my menu: two signatures + one basics
  bowl (e.g., signature beef noodles, hot-and-sour noodles, plain
  noodle soup)
- Current service speed: our standard per-bowl ticket-to-pass time at
  peak
- Venue: my shop's own wok station, or elsewhere

Three gates (fixed order, no swapping):

Gate 1 - Fixed menu:
- The candidate cooks only my three dishes; no personal menus, no
  choosing their own
- Two bowls of each: one to taste, one held for side-by-side reference
- Scoring dimensions: broth saltiness / noodle firmness / topping
  doneness / portion consistency, each 1-5 (3 = our house standard,
  which I define)

Gate 2 - Consistency (compare the two bowls from gate 1):
- Same bowl cooked twice; the metric is the gap: broth / noodle
  firmness / portion, scored as two-run deltas
- Delta <= 1 point: consistent. Delta > 1: flagged, not auto-fail -
  recorded as "instinct-driven, output variance risk"

Gate 3 - Multi-ticket:
- Three tickets land at once: different noodle types, different
  toppings, one with a modification (e.g., no scallions), timed to
  clear within my rush standard
- Judged on more than finishing: no mixed-up orders, the modified
  ticket flagged, and whether sequencing was sensible (simple before
  complex)

Output format:
- One scoresheet: three gates x dimensions, blanks for the judges
- Hiring decision lines (generated from my role requirements):
  · Instant veto: any mixed-up order or missed modification in gate 3 -
    those are accidents at rush hour, not blemishes
  · Hire outright: above the total line AND all three consistency
    deltas <= 1
  · One-week trial: above the line but consistency delta > 1 - the
    trial week specifically tracks daily output variance
- Trial-week checklist: if the trial verdict applies, what gets checked
  daily (fixed two dishes, saltiness logged, variance reviewed at
  week's end)

Constraints:
- No judging of character, attitude, or appearance - this tool measures
  craft and speed only
- No salary guidance and no "should you hire" verdict - the decision
  line is technical; whether to use this person is my call
- 3 always anchors to our house standard, never to "industry average" -
  the candidate will cook our noodles, not the industry's

Why it is written this way

Auditions measure the wrong skill, which makes them the quietest waste in restaurant hiring: the owner hires on peak performance and the shop pays for average performance. Banning personal menus in gate one exists because a personal menu is a rehearsed set - anyone can nail a dish cooked five hundred times, while the shop needs the opposite: adaptation to a fixed foreign menu. Two signatures plus one basics bowl is the minimum set covering adaptation, style, and fundamentals.

The consistency gate is the core. A noodle shop's reputation is not built on its best bowl but on "a hundred consecutive bowls tasting the same" - a returning customer served something different from their first visit leaves at a higher rate than one served a merely average first bowl. Cooking the same bowl twice and scoring the delta rather than the absolute exposes instinct-dependence: an instinct-driven cook can be brilliant bowl to bowl, but the shop cannot absorb that variance, and the craft cannot be taught - the cook himself has no written parameters to pass on. A large delta routes to a trial week instead of rejection because two runs is a small sample: the delta is a strong signal, not a verdict.

The multi-ticket gate reproduces the job's real shape. A noodle station never sees one ticket at a time; it sees three landing together with one no-scallions - and at rush, a mixed-up order is an accident, not a blemish. Missing a dietary modification is an instant veto not because it is difficult but because it injures a specific person (an allergic customer); that category of error must be zero from day one in kitchen habits.

Anchoring 3 to the house standard instead of the industry midpoint fixes the coordinate system: the new cook is taking over an existing line, and the score answers "can this person hold our line," not "is this a good cook in general" - the latter helps tomorrow's bowls not at all.

Variables