AM-00.06 · SPACE ACADEMY

AM-00.06 — Check whether a result is plausible: five tests before believing it

How can you suspect a result is wrong even when the calculator reports no error?

📄 Download A4 PDF

1 — Unit check

A speed result should end with speed units. If it ends in joules or kilograms, the chain of operations probably does not answer the question.

This check is independent of the numerical value.

Teaching diagram 1: 1 — Unit check
1 — Unit check

2 — Sign check

Distance travelled in a simple example should not become negative, but a velocity component can be negative if the direction convention allows it.

A sign is not good or bad without a defined reference frame.

Teaching diagram 2: 2 — Sign check
2 — Sign check

3 — Bound check

An efficiency defined as a simple fraction between 0 and 1 cannot become 1.4 unless the definition changes. A simple probability cannot be 230%.

Known bounds are powerful error detectors.

Teaching diagram 3: 3 — Bound check
3 — Bound check

4 — Limiting-case comparison

If a force goes to zero, should the predicted effect vanish? If distance doubles, should the quantity rise or fall?

Mentally testing extreme cases can reveal a transcribed equation error.

Teaching diagram 4: 4 — Limiting-case comparison
4 — Limiting-case comparison

Three complete examples: change one assumption to understand

Before each calculation, identify where every number comes from and whether it is measured, conventional, assumed, or calculated.

Three numerical examples in the course
Three compared cases

Example A — efficiency

98% = 0.98. A calculated 1.08 = 108% should trigger a question: are we still using the same efficiency definition?

A bound check does not prove 98% is correct; it reveals a possible inconsistency.

Example B — speed

120 km in 2 h gives 60 km/h. A result of 0.06 km/h would be three orders of magnitude too small.

Order of magnitude plus unit makes the problem visible.

Example C — pressure

If p = F/A, doubling F at constant area should double p. If your calculation predicts the opposite, check the relationship.

Sensitivity check before detailed calculation.

Inverse calculation

If p and A are known, recover F = p×A and substitute it back into p=F/A. Returning to p is a simple algebraic check.

Common trap and result check

Trap: “the result looks precise” is not a check. A number with six decimals can be precisely wrong.

Always check units, order of magnitude, and physical meaning before accepting a result.

Exercises and answers

Understand

Explain the relationship in your own words.

Answer: A good answer connects each operation to the physical or mathematical question being asked.

Change an assumption

Modify one input and predict the direction of change before calculating.

Answer: Make the qualitative prediction before using a calculator.

Verify

Name two checks after a calculation.

Answer: Units, order of magnitude, sign, physical consistency, or comparison with a known case.

Primary and educational sources