Danger Danger!

On Fri, 29 Nov 2013 19:33:35 -0800, Dan
wrote:

John B. writes:

On Fri, 29 Nov 2013 12:22:47 -0800 (PST), Dan O
wrote:

On Friday, November 29, 2013 8:26:34 AM UTC-8, Frank Krygowski wrote:
On Thursday, November 28, 2013 10:51:30 PM UTC-5, tussock wrote:
Graham wrote:
"Frank Krygowski" wrote:
Graham wrote:

I just received this in a news letter from the UK CTC of which I am a
member:

"My colleagues and I are absolutely devastated with the shocking rate at
which cyclists are dying on Londons roads. Six cyclists have died over
the last two weeks, all of them in collisions with large vehicles, three
of which were lorries. During this period, three pedestrians were also
killed in collisions with lorries in London."

From what I can tell, a typical year in London has about 75 pedestrians
killed, but fewer than 20 cyclists. See

http://www.theguardian.com/commentis...cyclists-like-
pedestrians-must-get-angry

"Less well known is that, mile for mile, it's more dangerous to be a
pedestrian than it is to be a cyclist..." Yet how many know that? How
many are being scared away from riding by the "Danger! Danger!" cries?

All I did was give you the latest *facts* from a respected cycling
organisation. Things are clearly getting worse and the recent ratio was 2
to 1 against cyclists.

Nonsense. Statistically, you'll see six cyclists die in an arbitrary
two-week period at some point in a year quite often, even at just 20/annum,
assuming there's at least some small correlating factors to deaths across
time.

If you start examining random data across arbitrary time periods (two
today!, four in a week!, eight since this time last month!) it gets even
sillier. You can always find meaningless noise in small random samples of
risk.

At 20 a year you'll randomly see somewhere between 14 and 26 for a
start, this could be a randomly high year and /everyone's/ odds are still
the same. Most years you will see two in a day somewhere, or 25-30% piling
up in an arbitrary 30 day period.

Now, you're /more/ likely to see big numbers come up there if there's
more danger to cyclists, but to tell if that's true you count the yearly
totals and do trend analysis across many years of all injuries and deaths (a
larger sample, less relative noise) to find the /chance/ it's getting more
dangerous, you don't try to /guess/ from a peak in the sample space.

Not if you care about finding the truth. There's a /lot/ of possible
press releases out there saying there was zero cyclist deaths in the last
two weeks, after all. Huzzah, problem solved.

Make /any/ change right now, and deaths will almost certainly trend down
in the short term, because that was a peak and it'll trend down no matter
what you do.

I accept this is a snap shot of the data which could be an aberation but
from what I am seeing here in the UK as the level of cycling is going up
- we have over doubled our club membership in the past year - the death
rate is also rising. I cannot see why you want to keep trotting out the
above mantra rather than getting behind the people who are trying to do
something about it.

More cyclists is normally found to lower the risk per cyclist. So you
just got safer, and all the new cyclists are also safer, but you can still
see more deaths with more cyclists even though it's safer per cyclist.

snip


For more on this subject: In his book _Risk_, John Adams examined all sorts of traffic crash data. I recall one section (around p. 78 in my edition) dealt with improvements to "black spots," or intersections and areas that had experienced surges of crashes. In response to a surge, authorities often made changes to the road design, and son of a gun, the crash counts dropped! Problem solved?

Not really. He also says that at "black spots" where nothing is done, you also see drops in crash counts. In some cases, you may see drops in crash counts at the black spot itself, but increases in crash counts at nearby locations.

It's conceivable that the cause-effect link is real. It might be straightforward (slow down all the traffic and crashes should decrease) and it might be secondary (tell road users "It's dangerous here!" and they may be more alert, but relax once past that spot).

But it's also conceivable that the changes are simply an effect of the probability of rare events.


People do get all up in arms about death; and people do arrive at silly,
illogical conclusions (or at least arrive at conclusions that might be
logical anyway but haven't been arrived at logically).

And people put too much stock in statistics - at least people who aren't
really looking at things from an actuarial standpoint.

Don't get me wrong, statistics are great. They're just very limited in
offering understanding of how and why things are.

That's why you've got to observe, investigate, analyze, and *think* of
*individual* events. Sure, then you can aggregate them if that helps
your purpose. But by then each one has been put through a pretty
subjective wringer... which it _needs to be in order to be understood_
and suitably classified or whatever.

The problem with tallying and aggregating and scoring and all-else-being-
equal is that... well, while death is a downer, for every death (or what-
ever reported phenomenon you're tallying), there are countless occurrences
of something that made a bicyclist's experience less optimum than it could
or *should* have been - often quite egregiously less optimum. And the
*reasons* for these things, if addressed, target the conflicts _that often
probably contribute to the death toll, etc._

Yes, bicycling deaths are *statistically* low risk, but they're not a rare
phenomenon. And besides, they're not even as impactful to bicycling on the
whole as, say, horn honks.

Well, the bicycle advocates (bicyclinginfo.org) state that there are
some 57 million cyclists (23.7% of the population) in the U.S. and the
National Highway Traffic Safety Administration says that there were
677 deaths in 2011, or 1 in 84,194 cyclists.

What's rare :-)


Hard to find many examples of.

Maybe but the war cry, "Danger! Danger! is apparently deserved. I
recently came across a site(a) that states:

"A new study shows that riding on regular saddles harms sexual health
not just in men, but women too."

So I can only assume that all you younger bike riders are sexually
dis-functional (being older has some advantages ) and I further assume
that being dis-functional you will eventually disappear as without
descendants you will probably just fade away.

Or perhaps a law similar to the 3 foot law - "No bicycle until you
have three kids" with the same penalties.

(a)
http://www.smartplanet.com/blog/scie...sexual-health/
--
Cheers,

John B.

On Tuesday, November 19, 2013 11:47:50 AM UTC-6, Graham wrote:
I just received this in a news letter from the UK CTC of which I am a member:



"My colleagues and I are absolutely devastated with the shocking rate at which cyclists are dying on London’s roads. Six cyclists have died over the last two weeks, all of them in collisions with large vehicles, three of which were lorries. During this period, three pedestrians were also killed in collisions with lorries in London."



Graham.

I am sorry that cycling deaths are up a bit.

I am curious if the U.K. allows cell phone use while driving ?

Other factors that could be involved may include :

1. Do cyclists stay real close to the curb ?
2. Do they ride on the sidewalk during the go to/come home from work period ?
3. Are they devoting 100% of their concentration on riding and being aware ?

Andy

On Saturday, November 30, 2013 10:30:15 PM UTC-5, Phil W Lee wrote:

There's also the fact that any increase in the number of cyclists must
logically mean there are more inexperienced cyclists around, who may
be expected to have a slightly higher risk level (usually a slight
lag, as they start to become confident but still lack experience).
If the rate of increase is stable, that will even out over time, and
if the rate of increase stops, the risk level statistics may show an
improvement for a while.

To find out if this is occurring, you'd need to examine the stats
closely enough to find out what proportion of casualties are novice
riders - but then you are reducing the data set even further, making
meaningful analysis even less likely.

Here in the U.S., we it's very hard to get reliable data on any factor affecting bike fatalities.

For example, we know that a disproportionate number of bike fatalities involve night riding on rural roads. It's strongly suspected that lack of proper bike lighting is a major contributor (since observation shows a large percentage of night riders without lights). But the data is scarce.

Bike fatalities are rare enough that researchers aren't motivated to get into the details. It just doesn't seem like a large problem, and it's difficult to get enough data points for statistical significance.

- Frank Krygowski

On 11/29/2013 7:03 PM, John B. wrote:
On Fri, 29 Nov 2013 12:22:47 -0800 (PST), Dan O
wrote:

On Friday, November 29, 2013 8:26:34 AM UTC-8, Frank Krygowski wrote:
On Thursday, November 28, 2013 10:51:30 PM UTC-5, tussock wrote:
Graham wrote:
"Frank Krygowski" wrote:
Graham wrote:

I just received this in a news letter from the UK CTC of which I am a
member:

"My colleagues and I are absolutely devastated with the shocking rate at
which cyclists are dying on Londons roads. Six cyclists have died over
the last two weeks, all of them in collisions with large vehicles, three
of which were lorries. During this period, three pedestrians were also
killed in collisions with lorries in London."

From what I can tell, a typical year in London has about 75 pedestrians
killed, but fewer than 20 cyclists. See

http://www.theguardian.com/commentis...cyclists-like-
pedestrians-must-get-angry

"Less well known is that, mile for mile, it's more dangerous to be a
pedestrian than it is to be a cyclist..." Yet how many know that? How
many are being scared away from riding by the "Danger! Danger!" cries?

All I did was give you the latest *facts* from a respected cycling
organisation. Things are clearly getting worse and the recent ratio was 2
to 1 against cyclists.

Nonsense. Statistically, you'll see six cyclists die in an arbitrary
two-week period at some point in a year quite often, even at just 20/annum,
assuming there's at least some small correlating factors to deaths across
time.

If you start examining random data across arbitrary time periods (two
today!, four in a week!, eight since this time last month!) it gets even
sillier. You can always find meaningless noise in small random samples of
risk.

At 20 a year you'll randomly see somewhere between 14 and 26 for a
start, this could be a randomly high year and /everyone's/ odds are still
the same. Most years you will see two in a day somewhere, or 25-30% piling
up in an arbitrary 30 day period.

Now, you're /more/ likely to see big numbers come up there if there's
more danger to cyclists, but to tell if that's true you count the yearly
totals and do trend analysis across many years of all injuries and deaths (a
larger sample, less relative noise) to find the /chance/ it's getting more
dangerous, you don't try to /guess/ from a peak in the sample space.

Not if you care about finding the truth. There's a /lot/ of possible
press releases out there saying there was zero cyclist deaths in the last
two weeks, after all. Huzzah, problem solved.

Make /any/ change right now, and deaths will almost certainly trend down
in the short term, because that was a peak and it'll trend down no matter
what you do.

I accept this is a snap shot of the data which could be an aberation but
from what I am seeing here in the UK as the level of cycling is going up
- we have over doubled our club membership in the past year - the death
rate is also rising. I cannot see why you want to keep trotting out the
above mantra rather than getting behind the people who are trying to do
something about it.

More cyclists is normally found to lower the risk per cyclist. So you
just got safer, and all the new cyclists are also safer, but you can still
see more deaths with more cyclists even though it's safer per cyclist.

snip


For more on this subject: In his book _Risk_, John Adams examined all sorts of traffic crash data. I recall one section (around p. 78 in my edition) dealt with improvements to "black spots," or intersections and areas that had experienced surges of crashes. In response to a surge, authorities often made changes to the road design, and son of a gun, the crash counts dropped! Problem solved?

Not really. He also says that at "black spots" where nothing is done, you also see drops in crash counts. In some cases, you may see drops in crash counts at the black spot itself, but increases in crash counts at nearby locations.

It's conceivable that the cause-effect link is real. It might be straightforward (slow down all the traffic and crashes should decrease) and it might be secondary (tell road users "It's dangerous here!" and they may be more alert, but relax once past that spot).

But it's also conceivable that the changes are simply an effect of the probability of rare events.


People do get all up in arms about death; and people do arrive at silly,
illogical conclusions (or at least arrive at conclusions that might be
logical anyway but haven't been arrived at logically).

And people put too much stock in statistics - at least people who aren't
really looking at things from an actuarial standpoint.

Don't get me wrong, statistics are great. They're just very limited in
offering understanding of how and why things are.

That's why you've got to observe, investigate, analyze, and *think* of
*individual* events. Sure, then you can aggregate them if that helps
your purpose. But by then each one has been put through a pretty
subjective wringer... which it _needs to be in order to be understood_
and suitably classified or whatever.

The problem with tallying and aggregating and scoring and all-else-being-
equal is that... well, while death is a downer, for every death (or what-
ever reported phenomenon you're tallying), there are countless occurrences
of something that made a bicyclist's experience less optimum than it could
or *should* have been - often quite egregiously less optimum. And the
*reasons* for these things, if addressed, target the conflicts _that often
probably contribute to the death toll, etc._

Yes, bicycling deaths are *statistically* low risk, but they're not a rare
phenomenon. And besides, they're not even as impactful to bicycling on the
whole as, say, horn honks.

Well, the bicycle advocates (bicyclinginfo.org) state that there are
some 57 million cyclists (23.7% of the population) in the U.S. and the
National Highway Traffic Safety Administration says that there were
677 deaths in 2011, or 1 in 84,194 cyclists.

What's rare :-)


Hospital acquired infections fell more people than car
wrecks for about 34,000 per year, a much higher rate of a
larger population.

Then again everything has risk. I'd rather ride than not.

--
Andrew Muzi
www.yellowjersey.org/
Open every day since 1 April, 1971

AMuzi wrote:
On 11/29/2013 7:03 PM, John B. wrote:
On Fri, 29 Nov 2013 12:22:47 -0800 (PST), Dan O
wrote:

On Friday, November 29, 2013 8:26:34 AM UTC-8, Frank Krygowski wrote:
On Thursday, November 28, 2013 10:51:30 PM UTC-5, tussock wrote:
Graham wrote:
"Frank Krygowski" wrote:
Graham wrote:

I just received this in a news letter from the UK CTC of which I am a
member:

"My colleagues and I are absolutely devastated with the shocking rate at
which cyclists are dying on Londons roads. Six cyclists have died over
the last two weeks, all of them in collisions with large vehicles, three
of which were lorries. During this period, three pedestrians were also
killed in collisions with lorries in London."

From what I can tell, a typical year in London has about 75 pedestrians
killed, but fewer than 20 cyclists. See

http://www.theguardian.com/commentis...cyclists-like-
pedestrians-must-get-angry

"Less well known is that, mile for mile, it's more dangerous to be a
pedestrian than it is to be a cyclist..." Yet how many know that? How
many are being scared away from riding by the "Danger! Danger!" cries?

All I did was give you the latest *facts* from a respected cycling
organisation. Things are clearly getting worse and the recent ratio was 2
to 1 against cyclists.

Nonsense. Statistically, you'll see six cyclists die in an arbitrary
two-week period at some point in a year quite often, even at just 20/annum,
assuming there's at least some small correlating factors to deaths across
time.

If you start examining random data across arbitrary time periods (two
today!, four in a week!, eight since this time last month!) it gets even
sillier. You can always find meaningless noise in small random samples of
risk.

At 20 a year you'll randomly see somewhere between 14 and 26 for a
start, this could be a randomly high year and /everyone's/ odds are still
the same. Most years you will see two in a day somewhere, or 25-30% piling
up in an arbitrary 30 day period.

Now, you're /more/ likely to see big numbers come up there if there's
more danger to cyclists, but to tell if that's true you count the yearly
totals and do trend analysis across many years of all injuries and deaths (a
larger sample, less relative noise) to find the /chance/ it's getting more
dangerous, you don't try to /guess/ from a peak in the sample space.

Not if you care about finding the truth. There's a /lot/ of possible
press releases out there saying there was zero cyclist deaths in the last
two weeks, after all. Huzzah, problem solved.

Make /any/ change right now, and deaths will almost certainly trend down
in the short term, because that was a peak and it'll trend down no matter
what you do.

I accept this is a snap shot of the data which could be an aberation but
from what I am seeing here in the UK as the level of cycling is going up
- we have over doubled our club membership in the past year - the death
rate is also rising. I cannot see why you want to keep trotting out the
above mantra rather than getting behind the people who are trying to do
something about it.

More cyclists is normally found to lower the risk per cyclist. So you
just got safer, and all the new cyclists are also safer, but you can still
see more deaths with more cyclists even though it's safer per cyclist.

snip


For more on this subject: In his book _Risk_, John Adams examined all
sorts of traffic crash data. I recall one section (around p. 78 in my
edition) dealt with improvements to "black spots," or intersections
and areas that had experienced surges of crashes. In response to a
surge, authorities often made changes to the road design, and son of a
gun, the crash counts dropped! Problem solved?

Not really. He also says that at "black spots" where nothing is done,
you also see drops in crash counts. In some cases, you may see drops
in crash counts at the black spot itself, but increases in crash
counts at nearby locations.

It's conceivable that the cause-effect link is real. It might be
straightforward (slow down all the traffic and crashes should
decrease) and it might be secondary (tell road users "It's dangerous
here!" and they may be more alert, but relax once past that spot).

But it's also conceivable that the changes are simply an effect of the
probability of rare events.


People do get all up in arms about death; and people do arrive at silly,
illogical conclusions (or at least arrive at conclusions that might be
logical anyway but haven't been arrived at logically).

And people put too much stock in statistics - at least people who aren't
really looking at things from an actuarial standpoint.

Don't get me wrong, statistics are great. They're just very limited in
offering understanding of how and why things are.

That's why you've got to observe, investigate, analyze, and *think* of
*individual* events. Sure, then you can aggregate them if that helps
your purpose. But by then each one has been put through a pretty
subjective wringer... which it _needs to be in order to be understood_
and suitably classified or whatever.

The problem with tallying and aggregating and scoring and all-else-being-
equal is that... well, while death is a downer, for every death (or what-
ever reported phenomenon you're tallying), there are countless occurrences
of something that made a bicyclist's experience less optimum than it could
or *should* have been - often quite egregiously less optimum. And the
*reasons* for these things, if addressed, target the conflicts _that often
probably contribute to the death toll, etc._

Yes, bicycling deaths are *statistically* low risk, but they're not a rare
phenomenon. And besides, they're not even as impactful to bicycling on the
whole as, say, horn honks.

Well, the bicycle advocates (bicyclinginfo.org) state that there are
some 57 million cyclists (23.7% of the population) in the U.S. and the
National Highway Traffic Safety Administration says that there were
677 deaths in 2011, or 1 in 84,194 cyclists.

What's rare :-)


Hospital acquired infections fell more people than car wrecks for about
34,000 per year, a much higher rate of a larger population.

Then again everything has risk. I'd rather ride than not.

Exactly.


--
duane

On Sunday, December 1, 2013 1:41:53 PM UTC-5, AMuzi wrote:

Then again everything has risk. I'd rather ride than not.

I certainly agree.

I've thought quite a lot about the differences between my father and me at my present age. Some of the differences have been due to luck, I'm sure, and some due to different life choices.

That's not to imply that choices different from mine were wrong - well, except for the cigarettes, perhaps. And practically speaking, there was never any way he'd have been able to be the avid cyclist that I was; that was not on the menu of choices for a WW2 vet returning with a purple heart.

But I (and my family) have gotten so much benefit and so much joy from cycling. It's helped to keep me so much healthier than he was at this age, and helped me to see so much more of the world, even if we count his Army-sponsored trip to the Pacific.

I'm so glad that nobody managed to scare me away from cycling's many benefits.

- Frank Krygowski

On Sun, 01 Dec 2013 12:41:53 -0600, AMuzi wrote:

On 11/29/2013 7:03 PM, John B. wrote:
On Fri, 29 Nov 2013 12:22:47 -0800 (PST), Dan O
wrote:

On Friday, November 29, 2013 8:26:34 AM UTC-8, Frank Krygowski wrote:
On Thursday, November 28, 2013 10:51:30 PM UTC-5, tussock wrote:
Graham wrote:
"Frank Krygowski" wrote:
Graham wrote:

I just received this in a news letter from the UK CTC of which I am a
member:

"My colleagues and I are absolutely devastated with the shocking rate at
which cyclists are dying on Londons roads. Six cyclists have died over
the last two weeks, all of them in collisions with large vehicles, three
of which were lorries. During this period, three pedestrians were also
killed in collisions with lorries in London."

From what I can tell, a typical year in London has about 75 pedestrians
killed, but fewer than 20 cyclists. See

http://www.theguardian.com/commentis...cyclists-like-
pedestrians-must-get-angry

"Less well known is that, mile for mile, it's more dangerous to be a
pedestrian than it is to be a cyclist..." Yet how many know that? How
many are being scared away from riding by the "Danger! Danger!" cries?

All I did was give you the latest *facts* from a respected cycling
organisation. Things are clearly getting worse and the recent ratio was 2
to 1 against cyclists.

Nonsense. Statistically, you'll see six cyclists die in an arbitrary
two-week period at some point in a year quite often, even at just 20/annum,
assuming there's at least some small correlating factors to deaths across
time.

If you start examining random data across arbitrary time periods (two
today!, four in a week!, eight since this time last month!) it gets even
sillier. You can always find meaningless noise in small random samples of
risk.

At 20 a year you'll randomly see somewhere between 14 and 26 for a
start, this could be a randomly high year and /everyone's/ odds are still
the same. Most years you will see two in a day somewhere, or 25-30% piling
up in an arbitrary 30 day period.

Now, you're /more/ likely to see big numbers come up there if there's
more danger to cyclists, but to tell if that's true you count the yearly
totals and do trend analysis across many years of all injuries and deaths (a
larger sample, less relative noise) to find the /chance/ it's getting more
dangerous, you don't try to /guess/ from a peak in the sample space.

Not if you care about finding the truth. There's a /lot/ of possible
press releases out there saying there was zero cyclist deaths in the last
two weeks, after all. Huzzah, problem solved.

Make /any/ change right now, and deaths will almost certainly trend down
in the short term, because that was a peak and it'll trend down no matter
what you do.

I accept this is a snap shot of the data which could be an aberation but
from what I am seeing here in the UK as the level of cycling is going up
- we have over doubled our club membership in the past year - the death
rate is also rising. I cannot see why you want to keep trotting out the
above mantra rather than getting behind the people who are trying to do
something about it.

More cyclists is normally found to lower the risk per cyclist. So you
just got safer, and all the new cyclists are also safer, but you can still
see more deaths with more cyclists even though it's safer per cyclist.

snip


For more on this subject: In his book _Risk_, John Adams examined all sorts of traffic crash data. I recall one section (around p. 78 in my edition) dealt with improvements to "black spots," or intersections and areas that had experienced surges of crashes. In response to a surge, authorities often made changes to the road design, and son of a gun, the crash counts dropped! Problem solved?

Not really. He also says that at "black spots" where nothing is done, you also see drops in crash counts. In some cases, you may see drops in crash counts at the black spot itself, but increases in crash counts at nearby locations.

It's conceivable that the cause-effect link is real. It might be straightforward (slow down all the traffic and crashes should decrease) and it might be secondary (tell road users "It's dangerous here!" and they may be more alert, but relax once past that spot).

But it's also conceivable that the changes are simply an effect of the probability of rare events.


People do get all up in arms about death; and people do arrive at silly,
illogical conclusions (or at least arrive at conclusions that might be
logical anyway but haven't been arrived at logically).

And people put too much stock in statistics - at least people who aren't
really looking at things from an actuarial standpoint.

Don't get me wrong, statistics are great. They're just very limited in
offering understanding of how and why things are.

That's why you've got to observe, investigate, analyze, and *think* of
*individual* events. Sure, then you can aggregate them if that helps
your purpose. But by then each one has been put through a pretty
subjective wringer... which it _needs to be in order to be understood_
and suitably classified or whatever.

The problem with tallying and aggregating and scoring and all-else-being-
equal is that... well, while death is a downer, for every death (or what-
ever reported phenomenon you're tallying), there are countless occurrences
of something that made a bicyclist's experience less optimum than it could
or *should* have been - often quite egregiously less optimum. And the
*reasons* for these things, if addressed, target the conflicts _that often
probably contribute to the death toll, etc._

Yes, bicycling deaths are *statistically* low risk, but they're not a rare
phenomenon. And besides, they're not even as impactful to bicycling on the
whole as, say, horn honks.

Well, the bicycle advocates (bicyclinginfo.org) state that there are
some 57 million cyclists (23.7% of the population) in the U.S. and the
National Highway Traffic Safety Administration says that there were
677 deaths in 2011, or 1 in 84,194 cyclists.

What's rare :-)


Hospital acquired infections fell more people than car
wrecks for about 34,000 per year, a much higher rate of a
larger population.

Then again everything has risk. I'd rather ride than not.

As you say, everything has some risks associated with it. People die
from falls, from accidents occurring while walking, during sex, in
fact every individual that survives birth is doomed to die.

On the other hand it does seem a bit "odd" for one to engage in a
voluntary activity that one considers dangerous, when all one needs to
do to be "safe" is to cease.

I remember as a kid being told, "Don't pull the dog's ears, he'll bite
you". If you continued in the ear pulling activity and did get bitten
you didn't get much sympathy :-)
--
Cheers,

John B.

On Sunday, December 1, 2013 3:03:54 PM UTC-5, Phil W Lee wrote:

I'm surprised if you have a high proportion of unlit night riders on
rural roads - certainly here they are overwhelmingly an urban species.
One peculiarity I discovered here (but I can't find the reference, I'm
afraid) is that on a straight risk per mile basis, unlit night riders
have a lower accident rate than lit ones. I suspect (but can't prove)
that it is because unlit cyclists have a higher tendency to stay well
away from motor vehicles than lit ones do. Risk compensation in
action - they KNOW they are invisible, and act accordingly.

I've wondered about those points myself. I think there are several parts to the explanation.

One is, I think, a difference between my country and yours (although I haven't been to Britain for 12 years). I think our "rural" is different from your "rural." To explain: I almost never see nighttime cyclists in isolated rural situations. But we seem to have far more ex-urban sprawl than you do.

So the last night fatality that we had in our county (last year) was doubtlessly classed as being on a rural road, although he was less than 200 yards from the local fire station, and within perhaps half a mile of a "rural" neighborhood with at least 100 houses. We have a tendency to build these "developments" in cornfields, for people who want to live the country life - surrounded by bunches of other people who also moved out into the country. Of course, the gas stations, convenience stores, etc. follow them out, and are soon flanked with truck repair businesses, used car lots, fast food franchises, etc. What some claim to be the largest hardware store in Ohio is not very far from me, on about twenty acres of commercial development on an otherwise rural road.

About the "know they are invisible" bit: The guy killed near here was riding home from work, IIRC. And he had a headlight - something that surprised me. But he had no taillight (although Ohio law requires one) and no rear reflectors except pedal reflectors. But soon after the crash, other motorists reported that he was almost invisible from the rear. The Highway Patrol thought that his pedal reflectors were obscured by the extra-long pants he wore.

I've read studies (and probably have them on file) showing that pedestrians routinely over-estimate their own nighttime visibility. I wouldn't be surprised if some cyclists do the same. That's one of the reasons I advocate having a friend ride your bike for you, as you observe from different perspectives. It's best if that includes observing while driving by in a car.

- Frank Krygowski