Safety of machinery: how yearly recorded accident data relate to the fascination ``vision zero''

The issues "risk assessment" and "tolerable risk" are causing conflicting reactions not only among health and safety experts. Experienced designers in the machinery sector are sometimes unsettled, too. The controversies are mainly about numerical representations of hazard "probabilities" and what they mean in the context of a machine design and its operation. Admittedly, theoretical risk assessment starts in the hypothetical "what if" domain, where theoretical risk can be estimated only logically in cause and effect (at the most), but not in an absolute scale. However, actual relative risk reduction effects between different yearly accident records can be calculated and compared quite exactly, because the real risk actually can be "measured" (relatively) precisely as by the records of Health and Safety representatives (in Germany BGHM and DGUV), as well as in individual records of machine tool builders. A typical question is being reiterated, when it comes to required upgrades in the design of certain safety functions in machine tools: "Is there any evidence in the operations field that indicates the need to upgrade this safety function, or is it only because of hypothetical fears?" And it also happened again and again that the alleged need for a required upgrade could not at all been proven by corresponding failures and related accidents with injuries. Therefore, this paper tries to support plausible risk considerations connecting theory and reality.

Keywords: Accident data, risk, machine tool.