On Sun, 19 Jan 2003 12:35:44 +0000 (UTC), Tim Woodall <
[email protected]> wrote:
>>>Typical Smith. Reject the scientific approach for a bunch of ifs and maybes.
>> Oh, very good. Why don't you tell us how you would relate the Joksch curve to the ratio of near
>> misses to fatals. You might even like to infer an average impact speed.
>Please either learn to think or go away.
>The Joksch curve has _ABSOLUTELY_NOTHING_ to do with the risk of having an accident. The curve
>merely tells you what the probable outcome is once the accident occurs.
Quite. But since we know the risk of accident severity there's no reason we shouldn't use the curve
to infer average crash speed.
>A possible first approximation might be that the risk of accident it proportional to speed
>for a given road. (For small changes in speed e.g. 1mph the "psychological effects" won't
>have any effect)
Except that it's a robust general rule that faster roads are safer.
>Also, assuming a 30mph road where the mean traffic speed is 31.21mph we have a deltaV between
>collision and imminent collision of 10mph.
Ok, well that's slightly over half the accidents perhaps. What about those accidents on 60 and 70mph
roads where the risk of death actually turns out to be less?
>So Risk of fatality is
>P(s) = C s (s-10)^4
> 1.Assuming 1.5 million collisions, speed 31.21mph 1000 deaths:
>So C is approximately 1e-10
>now reduce the speed to a rigorously enforced 30mph
Even on the 70mph roads? Good plan.
>Correcting for rounding errors:
>at 31.21mph - 947 deaths at 30.00mph - 720 deaths.
>A 24% reduction in deaths.
Of course there's a big effect on mortality from impact speed. That's what Joksch's curve is
intended to show.
> 2.Assuming 15 million near misses
>C is approximately 1e-11
>Now reduce the speed to a rigorously enforced 30mph
Even on 70mph roads? Again?
>Correcting for rounding errors:
>at 31.21mph - 947 deaths at 30.00mph - 720 deaths.
>
>A 24% reduction in deaths.
>Now assume that the risk of accident is quadratic in speed.
>P(s) = C s^2 (s-10)^4
> 3.Assuming 1.5 million collisions, speed 31.21mph 1000 deaths:
>So C is approximately 1/3e11
>now reduce the speed to a rigorously enforced 30mph
>Correcting for rounding errors:
>at 31.21mph - 986 deaths at 30.00mph - 720 deaths.
>A 27% reduction in deaths.
> 4.Assuming 15 million near misses
>C is approximately 1/3e12
>Now reduce the speed to a rigorously enforced 30mph
>Correcting for rounding errors:
>at 31.21mph - 986 deaths at 30.00mph - 720 deaths.
>A 27% reduction in deaths.
>(Exactish figures follow)
> 1. 2. 3. 4.
>31.21 1000 1000 1000 1000
>30.00 760 760 730 730
>
>Reducion 24% 24% 27% 27%
>Finally, what makes me most happy about this is that counting crashes that don't happen doesn't
>affect the number of people dying or injured on the roads. I wouldn't expect it to any more than
>counting the number of bicycles in Mozambique is likely to affect deaths on UK roads.
There are roughly fixed ratios between:
Near misses: damage only accidents: injury accidents: serious injury accidents: fatal accidents.
If we can reduce near misses by a percentage we can reasonable expect to reduce fatals by the same
strategy and a similar percentage.
The fact that near misses outnumber fatals by something like 7500:1 tells us just how effective
drivers are at avoiding fatal accidents. When that ratio if fed into the Joksch equation we get a
strong implication that the average impact speed is 7.6mph.
Since we know that drivers are using speeds in excess of 70mph on UK roads every day, it's
completely clear that numerical speed is not playing much of a role in road fatalities. Be very
aware that most of those near misses could have been a fatality. At only 40mph the Joksch equation
suggests that the likelihood of a fatality is around 10%. Using the 7,500,000 near misses figure
that could have been three quarters of a million annual fatalities.
It isn't speed that kills. We can reduce the speed limits endlessly or enforce them perfectly
without ever hoping to get close to the thresholds where free travelling speed will play a larger
part in the outcome than driver based factors like skill, attention, attitude and training level. In
fact, small variations in these factors will have far more effect on accident rates and outcomes
than big variations in limited or enforced speed.
--
Paul Smith Scotland, UK
http://www.safespeed.org.uk please remove "XYZ" to reply by email Let's make
speed cameras as unacceptable as drink driving
