Three assignment matrices act on the same true state to separate calibration from state preparation. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.

highlighted = computed this step

The baseline matrix predicts fifty-seven zero reads

With true probabilities three fourths and one fourth, the baseline assignment predicts 57 zero reads over 80 shots.

Nzero=57,None=23N_{\text{zero}}=57,\quad N_{\text{one}}=23
Baseline readout matrixThe first row is the strongest assignment matrix.true probabilitieszero 3/4one 1/4assignmentobserved probszero 57/80one 23/80over 80 shots57 zero reads23 one reads

Assignment contrast changes the observed counts

Each row keeps the same true state. Only the calibrated assignment counts change.

rNzeroNone157232552535327\begin{array}{c|c|c}r&N_{\text{zero}}&N_{\text{one}}\\1&57&23\\2&55&25\\3&53&27\\\end{array}
Middle contrast matrixThe second matrix predicts the middle count row.true probabilitieszero 3/4one 1/4assignmentobserved probszero 11/16one 5/16over 80 shots55 zero reads25 one reads

The weakest matrix shifts more shots into one

The third row still closes to the same shot total. The readout matrix changed the classical observation, not the prepared state.

Nzero+None=80N_{\text{zero}}+N_{\text{one}}=80
Weakest contrast matrixThe third row is rendered as its own checked prediction.true probabilitieszero 3/4one 1/4assignmentobserved probszero 53/80one 27/80over 80 shots53 zero reads27 one reads