I'm fascinated by the 'likelihood' aspect of risk assessment, touched upon by this thread, and its potential impact upon fire engineering.
One consequence is if you pour enough resources into prevention - either preventing ignition or slowing growth (e.g. by tight control of fire load), you could throw away any guidance on travel distances - have them as long as you like; dead-ends cease to be a problem; dispense with alarms! In effect, we go from the 'traditional' approach of assuming that a fire of significant size occurs, and designing/managing to deal with it to an assumption that it is improbable that a fire of significant size will occur. If you assume the latter, then established risk assessment techniques could show that spending ANY significant amount of money on additional precautions (fire doors, extinguishers, alarms etc) isn't justified, in pure life safety terms. The stonemasons could be an excellent example!
If we take this further, and look at the national statistics, the likelihood of any of us dying or suffering serious injury from fire in our place of work is infinitesimal - yet each year UK employers must spend millions on fire protection kit. In true risk assessment terms, how can this be money well spent? There must be a case to be made for reducing the current levels of fire protection, in these premises, especially if legislative changes improve prevention. Protection of property and assets argues against this, of course, but I know of no policy that encourages employers to consider this - even if you factor in the likelihood that a serious fire would cripple your business, the risk is still small (better to pour resource into contingeny planning, in many cases).
Totally 'off thread', of course, but I'm convinced that, come RRFSO, large organisations at least will see an opportunity to reduce safety spending in this way.