A rounded quit-tracking summary is not the exact underlying record

Identify which number is an observation and which is a display

An individually recorded count, an average, and a shortened display have different roles. Whole observed events may be counted exactly under your definition, while a mean can contain a fraction. A screen showing a whole number for that mean does not imply that the same whole number occurred each day.

Ofqual’s statistical publication policy describes rounding and notes that sums of rounded values may differ from a total calculated from the original values. That is a general reporting example, not the rule used by No Vape. Check the actual tool’s documentation before claiming that its display rounds in a particular way.

Ofqual — GOV.UK: Rounding policy for statistical publications

Use a fictional pair to see what disappears

Imagine two invented daily averages of 2.24 and 2.26 recorded events. Rounding each to one decimal place by the familiar nearest-value rule gives 2.2 and 2.3. Rounding each to a whole number gives 2 and 2. The underlying difference has not disappeared merely because one display shows equal labels.

These are arithmetic examples, not puff allowances, actual app outputs, or clinically meaningful differences. The question is what the chosen precision preserves. Do not add more decimal places to a remembered estimate just to make a comparison seem exact; display precision cannot improve the observation that produced the number.

Avoid repeated rounding during a calculation

If you are making a summary yourself, keep the available original values for the calculation and round only the result you intend to present. Label the unit, period, and rounding choice. Do not try to repair a mismatch by changing an observed entry until the displayed totals agree.

For a hypothetical illustration, three values of 1.4 sum to 4.2. Rounding each first to a whole number and adding gives 3; rounding the original total to a whole number gives 4. The two procedures answer different display questions. Neither establishes a problem in a named tracker, and neither recovers values that were not retained.

Describe the limit alongside the comparison

A useful note might say: “Both summaries display 2 at whole-number precision; the exact underlying means are not available.” That is more defensible than saying the periods were identical. If the original observations are available, describe them using the same rule and known coverage before making a practical inference.

Rounding is separate from missing data, changing definitions, and nicotine exposure. Even an unrounded arithmetic result cannot resolve those issues or demonstrate a health change. Use precision appropriate for an understandable description, then return to the situation you want to prepare for. A small numerical difference may not require a new routine merely because the display can show it.