Using Countermeasure Pack V3.02 my results are showing a small number of various countermeasures analyzed as having a negative severe crashes reduced per 100m . . . the value is not consistently negative for the countermeasure in general just occasionally. . . is this a problem with calibration or with the algorithms?
I never saw negative countermeasures before V3.02.
I think we need some more information in order to help. It looks like you are saying that some of the countermeasures are showing a reduction in fatal and serious injuries (FSIs) per 100m and that is what we would expect. The same countermeasures may have a greater reduction on FSIs in one location over another due to the different road attributes and traffic mix at each 100m segment.