This article studies differences in the results of using different risk estimation tools in the same hazardous situations involving dangerous machinery. We investigated how (a) types of risk estimation parameters and methods of constructing tools, (b) the number of levels of each parameter, and (c) the number of risk levels influence the results. Consequently, 31 risk estimation tools were compared by using them to estimate risk levels associated with 20 hazardous situations. Risk estimation appears to be tool-dependent, as different tools give different results with identical hazardous situations. The scope of the tool, its use, and construction could explain these differences. This article also proposes a series of rules for constructing tools to alleviate many problems associated with the variability of risk estimations.