01
Calculate every row before describing the pattern
For the first pair, twelve minus ten equals two. The next pairs give twenty-two minus twenty equals two and thirty-two minus thirty equals two. The supplied differences are therefore all positive two. The readings in this table exceed their paired references by the same amount. This pattern is consistent with a constant additive offset in the example. It does not identify the reason for that offset. The table contains no information about the instrument, the reference source, repeated observations or the conditions under which real readings might be collected.
02
Apply the candidate rule only as a worked calculation
Subtracting two from each supplied reading produces adjusted values 10, 20 and 30. Each adjusted value then equals its paired reference, so the residual difference for these three rows is zero. That result follows directly from the invented numbers. If a later reading is 27, the same rule would calculate 25, but without a paired reference its residual remains unknown. Keep the calculated value and the evidence of agreement in separate fields. Applying a formula to a new number does not independently establish that the formula is appropriate there.
03
Keep a record that can support later investigation
The worksheet fields are reference value, reading, defined difference, candidate rule, adjusted value and residual where a reference exists. A real investigation would add units, equipment identification, reference provenance, conditions and repeated observations. Preserve the original readings alongside any adjusted values so another reviewer can reconstruct the arithmetic. Label the rule as proposed until the relevant review establishes its use. This small exercise teaches how to describe one pattern precisely while leaving uncertainty, behaviour outside the supplied points and operational suitability unresolved.