Miniature heart-monitoring devices, such as implanted defibrillators and cardioverters, detected onset of abnormal heart rhythms and automatically apply one or more shocks to the heart. The shocks restore normal heart function without human intervention. A critical part of these devices is the monitoring circuitry, which includes a microprocessor and stored instructions, or algorithms, that govern how the devices interpret and react to electrical signals indicative of abnormal heart rhythms. The inventor has devised a relatively simple and accurate algorithm for determining appropriate therapy options. One version of the algorithm computes three statistics a range statistic, a minimum interval statistic, and a dispersion index, from a set of depolarization intervals. This algorithm defines the range statistic as the difference between largest and smallest depolarization intervals, the minimum interval as the smallest of the intervals, and the dispersion index as the standard deviation of the intervals. A scalar interval dispersion assessment, based on the three statistics, is then compared to a threshold to identify a rhythm as a flutter or fibrillation. The three statistics can also define a point in a three-dimensional space, with rhythm identification based on relative position of the point and a surface in the space.