Preparing the one state supplies the second calibration row, so both truth labels have their own observed-count ledger. The row closure is exact only for the stated counts and assignment convention; real readout calibration must track state-preparation error, drift, leakage, threshold shifts, and statistical uncertainty.
Preparing one supplies the second exact row of the assignment matrix. Exact arithmetic here means exact results for the stated model inputs; measured inputs still carry uncertainty and significant-figure limits.
highlighted = computed this step
Prepared one gives a second row
When one is prepared, the readout counts are 2 zero decisions and 8 decisions of one.
Nread zero=2,Nread one=8
This row also closes to one
The second row is 1/5 and 4/5. These fractions come from the same total of 10.