The relative risk score is related to the inspection frequency, the sprinklers will make that frequency longer but will not necessarily mean that less appliances are sent in the event of fire. If 1 hour compartmentation means the largest area that can be affected by fire is 100m2 then that might alter the number of appliances.
There are 3 sections of data on the form, 1 is for FSEC, which is the compartmentation/life risk/community importance/value/firefighter risk/environmental risk data etc, that part is directly related to turnout. Another is for the compliance level which gives an 'enforcement expectation' (i.e. You don't comply with RRFSO hardly at all, enforcement is likely but also dependant on other factors) and the final section helps fine tune the relative risk level, which alongside the compliance give a reinspection frequency.
Alot of IO where I am agree that sprinklers seem to knock too much off, but I am guessing that is all backed up by statistics that prove the efficiency of sprinklers on containing/extinguishing a fire.
Another thing to think about is sprinklers as part of an engineered solution. Does that also mean the risk is less when its part of a system of various elements all relying on each other for the maths to work? The data is blind to the difference between life/property protection or when it is a compensatory feature.