A deterministic rule produces the same result from the same inputs under the same conditions. This simple idea helps explain why a calculation can be checked, a transformation revisited, or a surprising outcome investigated.
The hidden ingredients
What counts as an input is not always obvious. A result may depend on the current date, a device setting, a file read from elsewhere, or a rule that has since changed. Two attempts can look identical while relying on different information.
Repeatability therefore starts with describing the conditions. Saving the source material, the relevant settings, and the rule version can be more helpful than saving only the final result. A screenshot preserves an appearance; it may not preserve the steps that produced it.
Consistent does not mean correct
A rule can repeat a mistake perfectly. Repeatability makes an outcome easier to inspect, but does not prove that the rule reflects the intended meaning. A calculation still needs an appropriate formula and suitable source values.
There are also tasks where live information is the point. A result based on changing information should change when that information changes. Keeping a dated snapshot can make an earlier outcome understandable without freezing every future result.
When comparing two outputs, ask which ingredients stayed fixed and which changed. That question often turns an apparently unpredictable difference into an explainable one, and gives a useful place to begin checking.