GUIDES · BETSLIP DOCTOR
How should I interpret an assist-to-turnover ratio when turnovers are zero?
If turnovers are observed as zero, report the assists and turnovers separately and mark the ratio undefined or N/A under the table’s convention. Zero is not missing data, and division by zero does not produce a finite assist-to-turnover ratio. If combining games, divide total assists by total turnovers when the combined denominator is positive; do not average individual game ratios.
For Bettors and cappers interpreting basketball statistics · Updated
Keep the raw counts when the denominator is zero
The NBA glossary describes Assist to Turnover Ratio as assists compared with turnovers. Read the two raw counts first. An observed line of 4 assists and 0 turnovers is different from a line where the turnover field is unavailable: the first has a recorded zero, while the second has no observed denominator.
As arithmetic, 4 divided by 0 has no defined finite ratio. Do not convert the result to zero or infinity, and do not imply the NBA mandates one display label; a report may use its own convention for an undefined value.
Combine counts before calculating a multi-game ratio
For a multi-game summary, add assists across the included games and add turnovers across the same games, then divide the totals if the combined turnover count is greater than zero. Averaging separate game ratios can give a different or undefined result and does not equal the ratio of the combined counts. State the included games and retain the raw totals.
Two fictional games and one combined line
Hypothetical Game 1: 4 assists and 0 turnovers. The zero is observed, but 4/0 has no defined finite ratio. Hypothetical Game 2: 6 assists and 2 turnovers, so its ratio is 6/2 = 3.0. Across both games the totals are 10 assists and 2 turnovers, giving a combined ratio of 10/2 = 5.0. Do not average an undefined Game 1 ratio with 3.0; these invented counts illustrate why the raw denominator and aggregation method matter.
Related questions
Should a zero-turnover game be treated as missing?
No. A recorded zero is an observed count. Missing data means the count was not available. Keep that distinction even though the ratio itself is undefined when turnovers equal zero.