A problem that causes a program to produce invalid output or to crash (lock up). The problem is either insufficient logic or erroneous logic. For example, a program can crash if there are not enough validity checks performed on the input or on the calculations themselves, and the computer attempts to divide by zero. Bad instruction logic misdirects the computer to a place in the program where an instruction does not exist, and it crashes.

A program with bad logic may produce bad output without crashing, which is the reason extensive testing is required. For example, if the program is supposed to add an amount, but subtracts it instead, bad output results, although the computer keeps running. See abend, bug and buggy.