MF-01 / INDEPENDENT REPLICATION / DRAFT FOR SPECIALIST REVIEW
No agreement, qualification, registration or result is claimed.

A. INDEPENDENCE
Owned Metric Flight instrument is the baseline. Replicator uses a separate team,
instrument and location. Dresden is a prospective provider, not an appointed team.
Record funding, conflicts, shared designs, software and calibration dependencies.
Payment buys work, never a favorable verdict. Contract publication of null results.

B. FREEZE BEFORE CONFIRMATORY DATA
Each lead and independent reviewer signs serial-numbered apparatus, optical and
force calibration chains, geometry, bandwidth, acquisition schedule, covariance,
corrections, exclusion rules, control thresholds and stopping rules. Register both
protocols and versioned analysis before unblinding decisive data. Preserve raw
records, environmental logs, calibration certificates and deviations with hashes.
Linearity slope/intercept tolerances and effective sample counts remain unresolved.

C. THE ESTIMAND
Full signed beam-on minus beam-off force, after nuisance corrections. At normal
incidence with collinear reflection/transmission: Fpred=(1+R-T)*P/c. P is intercepted
power at the mirror, not source rating. At ideal 2 W Fpred=13.3425638 nN.
For an opaque mirror use (1+R)*P/c, not 2*R*P/c. Characterize clipping, scatter,
angles and the full vector momentum ledger. Source and dump remain off balance.

D. PREDECLARED CLOSURE
For each team residual r=Fmeasured-Fpred. Require |r|<=U(r) AND
U(r)<=0.15*|Fpred| at the primary point, plus every control/repeatability gate.
U=k*u; k=2 is proposed, coverage must be justified. Do not divide uncertainty by a
near-zero residual. Different measured P,R,T mean different expected forces.
Compare residuals rA-rB using u_delta=sqrt(uA^2+uB^2-2*cov(rA,rB)). Shared standards
can correlate results despite separate apparatus; estimate covariance before
unblinding. Comparison coverage and discrepancy decision rule need signed review.
Agreement of two inadequate instruments is not a pass. Publish discrepancies.

E. BOUNDARIES
This reproduces conventional photon pressure, not new propulsion. Alpha 1 stays
on hold until a defined coupling, source geometry/material response, prior-constraint
mapping and distinguishable prediction pass external technical review. A signature
without that analysis does not close the gate. Graphene is not required for MF-01.
