We have a set of problems that are relatively rare, they occurrence is 0 to 4 per reporting period*. This range is actually statistically stable when we use the u prime chart. (aka the Laney chart. see: Laney, David B., “Improved Control Charts for Attributes”, Quality Engineering, 14(4), 2002, pp. 531-537). so as John says we don't randomly celebrate a zero occurrence or beat people up when 4 events occur.
*our reporting period is not time based it is lot based as our lots are consumed by the user over a 9-12 month period.