Safespeed spoilsports



Status
Not open for further replies.
On Tue, 04 Mar 2003 21:31:45 +0000, "Just zis Guy, you know?" <[email protected]> wrote:

>>>There is a widely held belief that drivers can be exoected to drive safely within the speed
>>>limit, and some are so rash as to suggest that having been required to demonstrate this at the
>>>time of passing their driving test it is not unreasonable to suppose that they may continue to be
>>>required to do so without significant detriment to road safety as they become more experienced.
>>>Funny, that.

>>Some may. Some will. Some won't. The "won't" group is far too large to ignore.

>But you must accept that, on the evidence that they were able to manage it while they were still
>unqualified and inexperienced, it is very definitely "won't" rather than "can't." And for those who
>won't obey the law we have courts.

I don't regard any inexperienced driver as "safe".

There's no way to identify them, and the courts can't deal with (maybe) 20 million.

>>>Oops! Argument by assertion again. Where's the proof? Oh, yes, you have acknowledged you don't
>>>have any of course. Ah, well.

>>Lack of proof does not stop it killing 3 each day.

>Oops! Argument by assertion again. Lack of proof means precisely that: you can't prove that it has
>ever killed a single person

I'm quite sure it will be recognised and accepted within a few years. The international
comparison figures are becoming quite obvious, and it won't be much longer before lots of people
start to notice.

>>>And equally reasonable to suggest that this is not an argument for not enforcing speed limits.

>>It's an important component of one of the arguments, and goes towards reducing the benefit side of
>>the changes that cameras bring.

>So what? So cameras only bring X amount of benefit instead of 2X or
>3X. Big fat hairy deal. Speeding, by contrast, brings absolutely no benefit to safety.

An effective road safety strategy, based on sound priorities brings benefits amounting to perhaps 6%
annual reduction in fatality rates. As we can see when we look around the world.

>>>>speed cameras cost lives
>>>But only when combined with dangerous drivers who are prepared to put speed above safety.
>>>Obviously.
>>But that's most of us for some of the time.

>By "us" do you include yourself? Are you prepared to put speed above safety?

I certainly hope not. But at that instant when a camera or a ("modern") cop appears, perhaps I do. I
can't be certain.

>I came across three people today who put /their/ speed above /my/ safety - is that OK? How about
>the time last week when I was run off the road by someone whose speed was more important than my
>safety. Was that acceptable?

Safe Speed, every time.

You can be put at danger by another road user's speed at well below the speed limit, and 2/3rds of
the time (where excessive speed is a factor) that's what happens.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Tue, 4 Mar 2003 21:36:48 +0000, Michael MacClancy <[email protected]> wrote:

>>No speed is inherently dangerous.

>It's a pity you missed out the comma - "No, speed is inherently dangerous."

Why? What (numerical) speed do you think is inherently dangerous?
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Tue, 4 Mar 2003 22:10:39 -0000, "Tony W" <[email protected]> wrote:

>> >If you are right it is clearly an outrage. If we scrapped cameras
>tonight
>> >(easily done, a government minister stands up and says no-one will be charged based on camera
>> >evidence forthwith) we would cut death on the
>road
>> >by 40+%.

>> No chance. But we could immediately return to a 5% per annum improvement in fatality figures.

>> >I agree -- proof would be nice but those figures are too horrifying to wait for proof. Turn the
>> >cameras off and watch the accident figures tumble.

>> It doesn't work quite like that. The system has a great deal of inertia.

>OK -- this is interesting. Can you clarify a couple of points, please?

>I take it you are estimating the 3 deaths per day by extrapolating the trend line from 1992 on to
>2003. i.e. if we had continued the with the same rate of improvement pre 1992 to today we would be
>killing 7 per day on our roads rather than 10 (my apologies for getting the numbers wrong before, I
>was working from memory rather than checking).

>So these deaths cannot be directly related to cameras, per se -- but to the 'lost opportunity' to
>continue the pre 1992 trend lines that cameras brought about?

Simply the overall effects of a road safety policy based on "speed kills". I suspect the
cameras enabled the policy, or maybe the policy enabled the cameras. Either way, it's the worst
we've ever had.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
"Paul Smith" <[email protected]> wrote in message
news:[email protected]...
>
> >I take it you are estimating the 3 deaths per day by extrapolating the
trend
> >line from 1992 on to 2003. i.e. if we had continued the with the same
rate
> >of improvement pre 1992 to today we would be killing 7 per day on our
roads
> >rather than 10 (my apologies for getting the numbers wrong before, I was working from memory
> >rather than checking).

Please confirm if the 3 extra deaths per day figure is or is not derived by comparing the pre 1992
rate of reduction in accidents with the reductions actually achieved. I think this is very important
and critical to the arguement

> Simply the overall effects of a road safety policy based on "speed kills". I suspect the cameras
> enabled the policy, or maybe the policy enabled the cameras. Either way, it's the worst we've
> ever had.

Sorry, this answer is a little too vague to identify exactly how you derive these figures. Precision
here is vital to your argument.

Tony
 
"Paul Smith" <[email protected]> wrote in
> "DR" <[email protected]> wrote:
>
> >> The point is that from the start to the finish of an average incident only about 2% of the KE
> >> ends up in the crash. Most of the rest is shed by driver response (braking etc).
>
> >No. You cannot assume that a crash is a single severe frontal impact. A crash can involve several
> >small impacts, metal grinding on road/other
metal,
> >destroying a hedge, cutting a furrow in a field etc, etc. All of these damage the car but may
> >keep severe acceleration off the occupants. It is low acceleration only that allows people to
> >survive crashes after travelling at speed; braking is not a pre-requisite and very naive of you
> >to consider it as the only possible reason.
>
> I'm not, and I have been considering those other factors. I'm reworking the 12mph page, and those
> sorts of factors are already included.
>
> I've been looking at various statistics, hoping to make deductions about the various places that
> energy might be dissipated, on average.

You don't appear to have read the bit I wrote above. It seems obvious to me where crash energy goes.

> At present there are too many assumptions, but the average KE remaining in the average injury
> crash may well turn out to be under .5% of the average KE in pre incident conditions on A roads.

Eh? I was once rear ended. I had a large enough gap to find the brake before hitting the car in
front. Lets say I ran into the leading car at 10mph and got 5mph off with the brakes. So I reached
15mph. The car behind was lighter than mine, so would have needed 35mph.

That's very conservative estimate of contact speed and assumes an elastic collision. This was nearly
30 years ago so structurally primitive cars by todays standards. There were 10 people in the 3 cars
and no one was hurt.

If 0.5% of original energy was released in the collision then the perpetrator would have needed to
be doing 500mph.

(Of course this actually was the simple crash scenario you envisage. I have not had a "through hedge
experience" but certainly don't imagine it doesn't happen)

David Roberts
 
On Wed, 5 Mar 2003 00:08:43 -0000, "Tony W" <[email protected]> wrote:

>> >I take it you are estimating the 3 deaths per day by extrapolating the
>trend
>> >line from 1992 on to 2003. i.e. if we had continued the with the same
>rate
>> >of improvement pre 1992 to today we would be killing 7 per day on our
>roads
>> >rather than 10 (my apologies for getting the numbers wrong before, I was working from memory
>> >rather than checking).
>
>Please confirm if the 3 extra deaths per day figure is or is not derived by comparing the pre 1992
>rate of reduction in accidents with the reductions actually achieved. I think this is very
>important and critical to the arguement
>
>> Simply the overall effects of a road safety policy based on "speed kills". I suspect the cameras
>> enabled the policy, or maybe the policy enabled the cameras. Either way, it's the worst we've
>> ever had.
>
>Sorry, this answer is a little too vague to identify exactly how you derive these figures.
>Precision here is vital to your argument.

http://www.safespeed.org.uk/fatality.html

You might want to consider what major factors might influence alterations in a national fatality
rate. I'd put these three in order of importance:

1) policy
2) investment in safety
3) technological change

But we can eliminate technological change because it's similarly present in all western countries,
and we don't all have the same results.

Most safety investment is by motor manufacturers and is equally applicable to many countries.

I've found no evidence that UK investment in road safety engineering has changed since the 80's, in
fact most figures appear to show a substantial increase.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Wed, 5 Mar 2003 00:40:31 -0000, "DR" <[email protected]> wrote:

>> >> The point is that from the start to the finish of an average incident only about 2% of the KE
>> >> ends up in the crash. Most of the rest is shed by driver response (braking etc).

>> >No. You cannot assume that a crash is a single severe frontal impact. A crash can involve
>> >several small impacts, metal grinding on road/other metal, destroying a hedge, cutting a furrow
>> >in a field etc, etc. All of these damage the car but may keep severe acceleration off the
>> >occupants. It is low acceleration only that allows people to survive crashes after travelling at
>> >speed; braking is not a pre-requisite and very naive of you to consider it as the only possible
>> >reason.

>> I'm not, and I have been considering those other factors. I'm reworking the 12mph page, and those
>> sorts of factors are already included.

>> I've been looking at various statistics, hoping to make deductions about the various places that
>> energy might be dissipated, on average.

>You don't appear to have read the bit I wrote above. It seems obvious to me where crash
>energy goes.

>> At present there are too many assumptions, but the average KE remaining in the average injury
>> crash may well turn out to be under .5% of the average KE in pre incident conditions on A roads.

>Eh? I was once rear ended. I had a large enough gap to find the brake before hitting the car in
>front. Lets say I ran into the leading car at 10mph and got 5mph off with the brakes. So I reached
>15mph. The car behind was lighter than mine, so would have needed 35mph.

>That's very conservative estimate of contact speed and assumes an elastic collision. This was
>nearly 30 years ago so structurally primitive cars by todays standards. There were 10 people in the
>3 cars and no one was hurt.

>If 0.5% of original energy was released in the collision then the perpetrator would have needed to
>be doing 500mph.

>(Of course this actually was the simple crash scenario you envisage. I have not had a "through
>hedge experience" but certainly don't imagine it doesn't happen)

You're asking me to put details to something which is work in progress, and I'm really not ready.
I'd rather gather all the info I can, and then discuss the results.

So many crashes result in little or no injury, and the fatal percentage is around 0.1%. Perhaps ten
times more even than these are complete misses due to drivers reacting and preventing the incident
from becoming an accident. So with one fatal per (ballpark) 10,000 incidents, what can we conclude
about the contribution of free travelling speed (and therefore KE) to the fatality?
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Wed, 05 Mar 2003 01:07:16 +0000, Paul Smith <[email protected]> wrote:

>So many crashes result in little or no injury, and the fatal percentage is around 0.1%. Perhaps ten
>times more even than these are complete misses due to drivers reacting and preventing the incident
>from becoming an accident. So with one fatal per (ballpark) 10,000 incidents, what can we conclude
>about the contribution of free travelling speed (and therefore KE) to the fatality?

That's "free travelling KE" of course. And I'm specifically concerned about free travelling speeds
because they are the ONLY speeds affected by enforcement.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
"Paul Smith" <[email protected]> wrote in message
news:[email protected]...
> >
> >Sorry, this answer is a little too vague to identify exactly how you
derive
> >these figures. Precision here is vital to your argument.
>
> http://www.safespeed.org.uk/fatality.html
>

Please answer the question. Imagine that I am the editor of the mass circulation Daily Blah. You
have got my attention (for the first time in months). I want to run the story that will bring your
campaign to the attention of everyone -- not least the Minister who can turn off the cameras this
afternoon.

I need an answer that I can understand and convey to my readers in less than 100 words. Keep the 2"
thick folder for the men from the ministry.

Do you base your claim that cameras are costing 3 lives a day on an extrapolation of pre 1992 data
compared with current accident rates?

T
 
On Wed, 5 Mar 2003 09:05:29 -0000, "Tony W" <[email protected]> wrote:

>> >Sorry, this answer is a little too vague to identify exactly how you derive these figures.
>> >Precision here is vital to your argument.

>> http://www.safespeed.org.uk/fatality.html

>Please answer the question. Imagine that I am the editor of the mass circulation Daily Blah. You
>have got my attention (for the first time in months). I want to run the story that will bring your
>campaign to the attention of everyone -- not least the Minister who can turn off the cameras this
>afternoon.

>I need an answer that I can understand and convey to my readers in less than 100 words. Keep the 2"
>thick folder for the men from the ministry.

>Do you base your claim that cameras are costing 3 lives a day on an extrapolation of pre 1992 data
>compared with current accident rates?

I'm bored with your silly game. Why don't you just say your piece?
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Wed, 5 Mar 2003 12:44:12 +0000, Michael MacClancy <[email protected]> wrote:

>>>> http://www.safespeed.org.uk/fatality.html

>I've taken a look at this and the data that Paul Smith provides. I have to say that my and his
>understanding of trend lines differ.

>A key element of his argument is that something significant happened in 1993. He has developed two
>trend lines to demonstrate this. The first is a line trending the data 1950-1993. Cursory
>examination of the charts shows that this line is wrong. Most of the actual data points for this
>period lie above the supposed trend line. In fact, what he appears to have done is draw a straight
>line between the 1950 point and the 1993 point and call this the trend. A more accurate method of
>determination would have been to determine the slope and intercept using a least squares fit. Doing
>this generates a quite different picture. In fact, the least squares line is actually above the
>actual data line for all years after 1987, showing that actual fatalities were lower than the
>long-term trend would predict. It is true to say that the difference between actual and trend has
>been decreasing but there would appear to be little evidence to suggest that the long-term
>(1950-1993) trend has lost its validity.

There's no doubt that there were sudden and dramatic changes in 1993. Easily the fastest changes in
modern times. There's a reference on the page to other charts which all show this change.

Obviously there's random noise in the data, but an examination of log scale graphs clearly reveals
an approximate straight line. I find it helpful and illuminating to "look along the line" which help
to show a gently improving trend from 1950 to 1993 and a rapidly worsening trend thereafter.

I've looked at lots of different averages, rolling averages and curve fits (including least squares)
and none appear to match the long term trends overall as well as the exponential decay lines chosen.

>He also shows a trend for the period 1978-1993. Again, the line hasn't been fitted using least
>squares. It's not clear why he has chosen 1978 as the starting point for this trend line but it
>is interesting to note that this line has a substantially higher gradient than others that he
>could have selected (e.g. 1975-1993) and a steeper trend line evidently better suits his
>argument that the rate of reduction in fatalities has decreased in recent years. The graph does
>show that the rate of improvement was very high in the early 1990s but decreased in the second
>half of the decade.

By contrast, almost any later year than 1978 would have given a steeper (yellow) trend line.

>I'm not sure whether some of his data isn't suspect. For the period 1990-1993 he shows motor
>traffic billion vkm as 411, 412, 412, 412 which doesn't demonstrate the randomness one might expect
>and is completely at odds with the long-term upward trend.

Recession.

All data from official sources and sources listed.

>Summarising, although it is interesting that the rate of improvement has decreased in recent years
>it is currently no worse than the long-term trend would suggest and so it is clearly questionable
>to suggest that the introduction of speed cameras has resulted in more deaths than would be
>expected if the historic long-term trend continued.

Oh really? Find another period of 5 years with a worse trend than the last five years. (I'll give
you a clue: there isn't one)

But most worrying of all, and not yet shown on the page is the rate of change of slope. Around 1993
the approximate annual reduction in fatalities was 7% pa... Then this figure ramps up year by year
to 0% for 2001 and (according to experimental provisional figures I've seen) shows around +1% (i.e.
worse) for 2002.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Wed, 5 Mar 2003 13:39:55 +0000 (UTC), Tim Woodall <[email protected]> wrote:

>>>>> http://www.safespeed.org.uk/fatality.html

>> I've taken a look at this and the data that Paul Smith provides. I have to say that my and his
>> understanding of trend lines differ.

>I've struggled with this as well but I really can't get any sort of trend line that fits the data
>at all well and I don't know enough stats to be able to make predictions from a poor fit or be
>able to estimate error bars. (My practical experience of using stats generally involves hundreds
>to thousands of results which can be reproduced multiple times so you have an estimate for the
>noise on any point irrespective of any other point. I don't know how to estimate noise when you
>only have one point for each measurement and you are trying to separate noise from trend. If you
>know the trend then estimating the noise is ok but in this case we seem to be attempting to
>determine the trend)

I've spent hours and hours on the same problems. One way to get noise reduction is to run a rolling
average, and I've done that for 2,3,4,5 and 10 years. All it appears to do is confirm the
exponential decay curves I'm using. I've seen no better long term fits than exponential decay.

I'd welcome alternative views of the data or trends (as it says on the page). My objective in this
is to highlight a problem which I'm certain is real and the more evidence and the more rigourous
the better.

>But, I've just discovered http://www.bast.de/htdocs/fachthemen/irtad/english/englisch.html

>Looking at "Road fatalities by traffic participation and age, Germany in 1999/2000 had the same
>fatalities as Britain in 1980. So, as a first approximation, it isn't surprising that Germany is
>achieving the safety improvements now that we were achieving in 1980. (Germany has done very well
>with child (0-14) fatalities and has practically caught up with the UK but the other figures
>swamp this)

Quite so. But the ~5% pa general trend has been long established, and I can find no reason to doubt
that the trend could have continued.

Looking at *all* the international data I have, the tendency to "hit the buffers" (in rich
western countries) shows a *perfect* correlation with road safety policy based on "speed kills"
and no correlation with low casualty rate, except for Britain and Sweden, (Who both have "speed
kills" anyway.)

>I would expect something similar with thermal insulation of houses between the UK and Scandinavia.
>Our building regulations are only at the level Scandinavia was in the 70s. I expect that the "trend
>in thermal insulation" is better in the UK than in Scandinavia and I can imagine somebody in
>Scandinavia saying - We don't need triple glazing. Look, since it became compulsory our rate of
>inprovements in thermal insulation have decreased. In the UK they have only recently made double
>glazing compulsory in new buildings and they are getting a far better improvement year on year.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Wed, 5 Mar 2003 13:51:01 -0000, "Tony W" <[email protected]> wrote:

>> >Do you base your claim that cameras are costing 3 lives a day on an extrapolation of pre 1992
>> >data compared with current accident rates?

>> I'm bored with your silly game. Why don't you just say your piece?

>Silly games?

>It seems clear that this is the difference between your extrapolation from pre 1992 figures to
>today as compared with the current figures (rounded and approximated).

>For some strange reason you will not confirm this preferring to provide links to pages on your
>web site.

I think the web page with graphs, a downloadable spreadsheet, and full reference sources says far
more than I ever could in a newsgroup post.

http://www.safespeed.org.uk/fatality.html

>Now your original statement of saving 3 lives a day (average) did, indeed, grab my attention. It
>will come as no surprise to you that I have considered you, to put it mildly, a bit of a loon. But
>if you really have stumbled onto a way of saving 1000+ lives per year (and presumably many more
>injuries) maybe I had better listen.

Look at the page. Play around with the spreadsheet. Ask any questions based on what I've done, which
is complete, transparent and essentially simple.

I'm sure that decisions and approaches are open to criticism. But I can't find any other conclusions
which fit the facts as I understand them.

If there's ANY error or ambiguity, I'll be pleased to correct it.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
"Paul Smith" <[email protected]> wrote in message
news:[email protected]...

> >Do you base your claim that cameras are costing 3 lives a day on an extrapolation of pre 1992
> >data compared with current accident rates?
>
> I'm bored with your silly game. Why don't you just say your piece?

Silly games?

On 4/3/03 you made the following statement in a reply to Guy Chapman:

"Lack of proof does not stop it killing 3 each day".

I have been trying to get you to confirm what, exactly, "it" is and how the figure 3 deaths daily
has been derived.

In your first reply you provided 2 subtly different responses as to what "it" is.

Firstly, in reply to my question "I take it that the statement above suggests you believe that 3
people per day are losing their lives as a direct result of the government's policy of using speed
cameras to detect speeding drivers".

You replied "And all that has come with it. Yes. (including loss of traffic police, "speed kills"
as a primary road safety message, loss of focus and investment in other, better road safety
strategies, etc)"

But in reply to "So you could be claiming that over 40% of all deaths on the road are due to
speed cameras."

You replied (correcting an arithmetic error of mine) "A bit under 30%. Yes".

Now it is very tempting to pursue the second version as it represents a very clear cause and effect.
If it were true we could switch the cameras off this afternoon and watch our roads become safer
almost immediately.

However, it is fairly clear that you do not, in fact, mean cameras alone but cameras and a range of
related, consequential and supporting measures. This is particularly clarified by a further line in
your same reply of 4/3/03 in response to my suggestion that turning off the cameras would have an
immediate effect. :-

"No chance. But we could immediately return to a 5% per annum improvement in fatality figures".

Turning to the second point that I wanted clarified -- how the 3 deaths per day was calculated.

It seems clear that this is the difference between your extrapolation from pre 1992 figures to today
as compared with the current figures (rounded and approximated).

For some strange reason you will not confirm this preferring to provide links to pages on
your web site.

Now your original statement of saving 3 lives a day (average) did, indeed, grab my attention. It
will come as no surprise to you that I have considered you, to put it mildly, a bit of a loon. But
if you really have stumbled onto a way of saving 1000+ lives per year (and presumably many more
injuries) maybe I had better listen.

T
 
Paul Smith <[email protected]> wrote:
>
> Obviously there's random noise in the data, but an examination of log scale graphs clearly reveals
> an approximate straight line.

This reminds me of comment from my PhD supervisor many years ago when trend lines had to be manually
calculated:

"Any data looks like it fits a straight line trend if you plot it on a log-log graph"

Tony

--
http://www.raven-family.com

"I don't want any yes-men around me. I want everybody to tell me the truth even if it costs them
their job."

Samuel Goldwyn
 
On Wed, 5 Mar 2003 13:52:23 -0000, "Tony Raven" <[email protected]> wrote:

>> Obviously there's random noise in the data, but an examination of log scale graphs clearly
>> reveals an approximate straight line.

>This reminds me of comment from my PhD supervisor many years ago when trend lines had to be
>manually calculated:

>"Any data looks like it fits a straight line trend if you plot it on a log-log graph"

Which of course is why I've also provided exactly the same graphs with conventional scales.

But in this case, the log graphs ARE genuinely clearer because we're dealing with an approximate
exponential decay which does show as a straight line on a graph with a log Y scale.
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
On Wed, 5 Mar 2003 12:44:12 +0000, Michael MacClancy <[email protected]> wrote:
> In message <[email protected]>, Paul Smith
> <[email protected]> writes
>>>> http://www.safespeed.org.uk/fatality.html
>
>
> I've taken a look at this and the data that Paul Smith provides. I have to say that my and his
> understanding of trend lines differ.
>
I've struggled with this as well but I really can't get any sort of trend line that fits the data at
all well and I don't know enough stats to be able to make predictions from a poor fit or be able to
estimate error bars. (My practical experience of using stats generally involves hundreds to
thousands of results which can be reproduced multiple times so you have an estimate for the noise on
any point irrespective of any other point. I don't know how to estimate noise when you only have one
point for each measurement and you are trying to separate noise from trend. If you know the trend
then estimating the noise is ok but in this case we seem to be attempting to determine the trend)

But, I've just discovered http://www.bast.de/htdocs/fachthemen/irtad/english/englisch.html

Looking at "Road fatalities by traffic participation and age, Germany in 1999/2000 had the same
fatalities as Britain in 1980. So, as a first approximation, it isn't surprising that Germany is
achieving the safety improvements now that we were achieving in 1980. (Germany has done very well
with child (0-14) fatalities and has practically caught up with the UK but the other figures
swamp this)

I would expect something similar with thermal insulation of houses between the UK and Scandinavia.
Our building regulations are only at the level Scandinavia was in the 70s. I expect that the "trend
in thermal insulation" is better in the UK than in Scandinavia and I can imagine somebody in
Scandinavia saying - We don't need triple glazing. Look, since it became compulsory our rate of
inprovements in thermal insulation have decreased. In the UK they have only recently made double
glazing compulsory in new buildings and they are getting a far better improvement year on year.

Regards,

Tim.

--
God said, "div D = rho, div B = 0, curl E = - @B/@t, curl H = J + @D/@t," and there was light.

http://tjw.hn.org/ http://www.locofungus.btinternet.co.uk/
 
On Wed, 5 Mar 2003 14:06:31 +0000, Michael MacClancy <[email protected]> wrote:

>>But most worrying of all, and not yet shown on the page is the rate of change of slope. Around
>>1993 the approximate annual reduction in fatalities was 7% pa... Then this figure ramps up year by
>>year to 0% for 2001 and (according to experimental provisional figures I've seen) shows around +1%
>>(i.e. worse) for 2002.

>I don't think there's any disagreement here although your efforts to make a link with speed cameras
>are nonsensical. You won't make that link by surmising. You need independent evidence that says
>that X% of fatalities (and rising) were caused by sudden braking/loss of control/undue care due to
>the presence actual or otherwise of cameras.

The link I've made is the opposite one. I knew very well from my expert knowledge of driving that
speed cameras (etc) would make the roads more dangerous in 1989 when they were first talked about.

I'm just watching the tragedy play out, and reporting it as best I can.

>I agree that the recent trend in statistics is a matter of concern. We differ in our willingness to
>infer more from the statistics than is contained in them.

And it's isn't the cameras as such... it's the subtle effect on drivers priorities which are
distorted by the cameras. See:

http://www.safespeed.org.uk/tiger.html
--
Paul Smith Scotland, UK http://www.safespeed.org.uk please remove "XYZ" to reply by email speed
cameras cost lives
 
"Paul Smith" <[email protected]> wrote in message
news:[email protected]...
>
> I'm sure that decisions and approaches are open to criticism. But I can't find any other
> conclusions which fit the facts as I understand them.
>
> If there's ANY error or ambiguity, I'll be pleased to correct it.

Well the biggest problem I see is that you need to identify specific changes to the different
components of your hate list.

Cameras didn't all pop up over night. Why have you identified 1993 as year zero? How many cameras
were there then? How have they spread? Is that speed consistent with reduced road safety?

Similar questions for changes in policing, various road safety campaigns etc.

Specifically, since you allocate much of the problem to driver distraction -- what evidence do you
have of accidents consistent with driver distraction? Have there been significant changes in single
car accidents near camera sites? etc.

Why did the government paint the cameras yellow? Did they know something we don't? Has this changed
the accident figures? (That was a very quick change. If there were significant numbers of accidents
at camera sites caused by late sighting of the camera and lose of control and rear ending accidents
as a consequence these should be obvious from the figures.

Picking a change in trend and just saying that you can't think of anything other than cameras that
caused it simply is not adequate.

Paul, you may be onto something. I am fairly convinced that you are not and that you are just a loon
-- but I concede that it is possible that you have seen something in the figures. If so you need
solid evidence not vague speculation. And I dismiss almost everything currently in your web site at
present as either vague speculation or pretty but meaningless pictures.

T
 
On Wed, 05 Mar 2003 14:00:55 +0000, Paul Smith <[email protected]> wrote:
> On Wed, 5 Mar 2003 13:39:55 +0000 (UTC), Tim Woodall <[email protected]> wrote:
>
>>>>>> http://www.safespeed.org.uk/fatality.html
>
>>> I've taken a look at this and the data that Paul Smith provides. I have to say that my and his
>>> understanding of trend lines differ.
>
>>I've struggled with this as well but I really can't get any sort of trend line that fits the data
>>at all well and I don't know enough stats to be able to make predictions from a poor fit or be
>>able to estimate error bars. (My practical experience of using stats generally involves hundreds
>>to thousands of results which can be reproduced multiple times so you have an estimate for the
>>noise on any point irrespective of any other point. I don't know how to estimate noise when you
>>only have one point for each measurement and you are trying to separate noise from trend. If you
>>know the trend then estimating the noise is ok but in this case we seem to be attempting to
>>determine the trend)
>
> I've spent hours and hours on the same problems. One way to get noise reduction is to run a
> rolling average, and I've done that for 2,3,4,5 and 10 years. All it appears to do is confirm the
> exponential decay curves I'm using. I've seen no better long term fits than exponential decay.
>

I've done a least squares regression fit on a 10 year rolling average using an exponential trend and
the data fits well up until the years including 1999 onwards. But we already know that the method of
data collection changed in 1999, all of the stats produced have a footnote "Data from 1999 onwards
cannot be directly compared to years before 1999".

And I don't know how you can possibly point to a particular year when you have a 10 year rolling
average. But at the very least, if there really was a step change in 1991, I would expect it to be
obvious in the rolling average by 1986-1996 but it isn't (at least to me).

Taking the 10 year improvements,

1960-1970 through to 1976-1986, accident rates fell by about 40% per decade.

1977-1987 through to 1981-1991 accident rates fell by about 45% per decade although this isn't as
consistent a reduction, even when averaged over 10 years.

1982-1992 thought to 1986-1996 accident rates fell by about 50% per decade

1987-1997 through to 1989-1999 accident rates fell by about 45% per decade.

Beyond 1999 the method of stats collection changed but the figures are sitting somewhere abound
35-45% for the two years I have. Yes the decade 1991 to 2001 is the worst in the period I have
considered but this is only one period, there were worse periods in the 1960-1970 through 1976-1986.

Finally, there were four exceptional years in succession, 1991 - 12% fall, 1992 8% fall, 1993, 10%
fall, 1994 7% fall. These four years occured after the magic 1991 "speedcamera" year. I would expect
speed cameras to be placed first at the worst accident blackspots and then only later get installed
at the less "accident prone" sites so although 1991 may only have had a few speed cameras, they were
probably sited at the places which could give the best improvement. I remember the Bham New Road
that regularly used to have serious accidents because of cars jumping red lights. You could almost
overnight see the difference once the traffic light cameras got installed. There were a few speed
cameras on this road but, in general, if people drove at speeds where they could stop for a red
light, then then generally wouldn't be exceeding the speed limit much anyway. (I rarely drive this
road now and am not local so I don't know whether the accident improvements have been maintained but
I hope so)

Tim.

--
God said, "div D = rho, div B = 0, curl E = - @B/@t, curl H = J + @D/@t," and there was light.

http://tjw.hn.org/ http://www.locofungus.btinternet.co.uk/
 
Status
Not open for further replies.

Similar threads