Offbyone

Off-by-one error, commonly called an off-by-one bug, is a frequent programming flaw where a loop boundary or index targets one element too many or too few. It typically arises from confusion between inclusive and exclusive bounds or between zero-based and one-based indexing. The result is that code accesses an element outside the intended range, or misses the first or last element.

Causes and patterns include using <= where < is intended in loops, or using length as the last valid index instead of length-1. It also occurs when iterating over ranges, slices, or arrays with incorrect start or end values, or when counting elements starting from 1 in a context that is zero-based. Language features such as range, slices, or iterators can either conceal or reveal the bug depending on whether bounds are inclusive or exclusive.

Examples are common across languages. In C, for (i = 0; i <= n; i++) accesses n+1 elements, whereas the correct form is i < n. In Python, range(n+1) yields 0 through n, which can be unintended if the goal is 0 through n-1. Strings and arrays are also susceptible, especially when calculating last indices as length-1 but looping with length.

Mitigation strategies include adopting half-open intervals (start inclusive, end exclusive), using language-provided iteration constructs, and writing explicit boundary tests. Boundary and stress tests, careful review of loop conditions, and static analysis can help detect off-by-one errors. Understanding the difference between inclusive and exclusive bounds and keeping indexing consistent are key to avoiding them.

See also: zero-based indexing; inclusive-exclusive bounds; boundary testing.