Re: I crash into religion - Page 7

Re: I crash into religion

Peter Clinch wrote:
>
> His points seem to me to be saying that when shooting something down in
> flames make sure every bullet is right on the mark. Which is a fair
> point. Furthermore, if one is going to hold up something as an example
> it helps if it's bulletproof. And that's true too.
>

As you will be well aware, science never provides absolute certainty and
this is widely exploited by the anti-brigade where the evidence is
very strong but can never be absolute. Examples include the whole MMR
fiasco, silicone implants, mobile phones and phone masts etc etc. You
can always make the claim that the moon may be made of green cheese but
just not the bits that people have visited or that Russell's Teapot
exists and they cannot be refuted. IMO p.k is just following that well
trodden path that ultimately leads to the Land of Wally who claimed that
none of it applied to him because no study had been done specifically on
the streets he rode in the town he rode in. YMMV.

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

Re: I crash into religion

Chris Gerhard wrote:
> Tony Raven wrote:
>>
>> So you should easily be able to see the step downwards in the number
>> of head injuries at the time the helmet law was introduced? Yes?
>>
>> OK, so now take the test that, every time I have posed it to helmet
>> supporters, they have all studiously ignored and avoided taking. On
>> http://www.cycling.raven-family.com/Helmet%20Graphs.jpg I have
>> reproduced the graphs of absolute head injury numbers in two countries
>> that have introduced helmet laws. So identify for me at which
>> horizontal tick mark the helmet law was introduced in each graph. It
>> should be very easy, just look for the downward step in head injuries.
>>
>> Go on, be the first helmet supporter to dare to take the test (without
>> cheating and looking up the answer).
>>
>
>
> Do you have a corresponding graphs of cycle use for the same time periods?
>
> --chris

Yes but to show it would give away the answer because there is a very
clear drop in cycle use at the time the MHL was introduced.

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

Re: I crash into religion

Tony Raven wrote:
> Peter Clinch wrote:
>>
>> His points seem to me to be saying that when shooting something down
>> in flames make sure every bullet is right on the mark. Which is a
>> fair point. Furthermore, if one is going to hold up something as an
>> example it helps if it's bulletproof. And that's true too.
>>
>
> As you will be well aware, science never provides absolute certainty
> and this is widely exploited by the anti-brigade where the evidence
> is very strong but can never be absolute. Examples include the whole MMR
> fiasco, silicone implants, mobile phones and phone masts etc etc. You
> can always make the claim that the moon may be made of green cheese
> but just not the bits that people have visited or that Russell's
> Teapot exists and they cannot be refuted. IMO p.k is just following
> that well trodden path that ultimately leads to the Land of Wally who
> claimed that none of it applied to him because no study had been done
> specifically on the streets he rode in the town he rode in. YMMV.

<ironic mode on>

As you will be well aware, science never provides absolute certainty
and this is widely exploited by the anti-brigade where the evidence is
very strong but can never be absolute. Thus the anti helmet brigade are able
to exploit lack of certainty in the statistics to undermine the case for
helmets.


pk

Re: I crash into religion

Peter Clinch wrote:
>
> Whether or not that is the case, we /do/ need to keep being careful that
> all our points stand up, whether they are trying to prop up one side or
> shoot down the other.
>

Agreed but that does not mean we have to accept Russell's Teapot attacks
on fundamentally sound research as evidence that our points don't stand
up and the case is flawed.

--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

Re: I crash into religion

Peter Clinch wrote:
> Tony Raven wrote:
>
>> Agreed but that does not mean we have to accept Russell's Teapot
>> attacks on fundamentally sound research as evidence that our points
>> don't stand up and the case is flawed.
>
> As I said, I think p.k. generally overstates his case.

Clearly, I can stop doing that now as the point has finally got through! ie
the case against compulsion is sound but is weakened by overstatement!

> But it's
> still a point to remember that one can have the right answer and
> still spout some questionable reasoning to justify it. If p.k. makes
> a point about there being some possible mechanisms for such and such
> that isn't necessarily an argument that such mechanisms are clearly
> in operation, rather that one should acknowledge they are possible
> and say why why one thinks they're not operating.

Acts ix.17-18

pk

Re: I crash into religion

Ozark Bicycle wrote:
> frkrygow@gmail.com wrote:
> >
> > We know
> > you read _some_ information regarding bicycling and helmets. In
> > today's climate, that means you _have_ been told (in written form) to
> > "always wear a helmet."
> >
>
> You know what? What you know is crap. I haven't seen such a piece of
> propaganda. Or, if I have, it made no impression. Really.

Well, I certainly don't doubt that reading it made no impression on
you! However, we weren't discussing the capability of the reader. We
were discussing the intent of the writer.

(Yes, and I know you'll be furious that I trimmed some of your wonderul
prose. But really, we can't quote the entire thread with each post!
The internet would grind to a halt.)

> And more to the point: the subject of helmet use or non-use *never*
> comes up in everyday riding, with the sole exception of organized
> (i.e., club rides, charity rides, etc.) rides. Sorry to disappoint you,
> Mr. Roboto, but most cyclists simply don't make an issue of the matter.

Most cyclists have been sold a bill of goods. Most cyclists never get
to hear that there is any doubt about helmet effectiveness, let alone
about the need for helmets.

And FWIW, I have had many cyclists and non-cyclists bring the issue up,
the most recent being yesterday evening, by the head of a Metropolitan
Planning Organization. In fact, I recall a colleague of mine (PhD in
Engineering) telling me about his casual bicycling, and saying in an
abashed tone "I know I should wear a helmet, but I really don't want
to..." He was very interested when I responded that he _didn't_ have
to, and briefly told him about cycling's safety, about helmets' low
testing standards and apparent lack of effectiveness.


> Why is this helmet thing such a hobby horse for you?

First, why should it matter? Different people have different
interests. This is how society advances. If you don't happen to be
interested in it (enough to actually read and learn, that is) you're
free to stop posting. In fact, that would be the smart thing to do.

But the reasons I'm interested are several. First, I've long been
committed to promoting cycling, and the "Cycling is dangerous!!!!"
propaganda that goes with helmet promotion is anti-cycling. Second, I
am an engineering professor, and the mechanical issues (the patently
inadequate certification standards, the mechanics of brain injury) are
professionally interesting to me. Third, I have a low tolerance for
ignorance and bull****. Especially when it's detrimental to something
I love.


> BTW, your "Bicycling Life" site is really bull****, IMO. And I thought
> that *before* I noticed you were behind it, really. It just drips of
> geeky, self-inflated, self-imagined superiority.

:-) And so, when you run out of even lousy arguments, you flail around
attacking anything you can? Pitiful and shameful.

But I take comfort in the fact that less biased people have
emphatically disagreed with you. We've received many, many compliments
and requests for permission to reprint articles. And in fact, the
editor of Adventure Cycling magazine recommended our site this past
month. Read the letters to the editor column to see.

>
> Go visit the late Ken Kifer's site. Something to aspire to, Mr. Pocket
> Protector.

:-) You'll be interested to know that Ken and I were close online
friends. He had planned to visit my wife and I on a future bike tour,
especially once he learned a branch of his family was from nearby. Oh,
and he did, indeed, help us found the www.bicyclinglife.com website.

I wish he were still here. He'd have enjoyed "discussions" with you.
As an example, see http://www.kenkifer.com/bikepages/advocacy/mhls.htm
and http://www.kenkifer.com/bikepages/health/risks.htm

- Frank Krygowski

Re: I crash into religion

In article <zBOag.22322$43.218@nnrp.ca.mci.com!nnrp1.uunet.ca>,
jtaylor@deletethis.hfx.andara.com says...
>
> "Espressopithecus (Java Man)" <rickk@letterectomyTELUS.net> wrote in message
> news:MPG.1ed518b7b24d479598978f@shawnews.vc.shawcable.net...
> > >
> > > The population stats tell you you are more likely to be increasing your
> > > risk by wearing a helmet.
> >
> > > Now you may be a very cautious rider which
> > > means you will minimise the chance of you having an accident in the
> > > first place but 90% of the deaths or serious injuries are caused by
> > > being hit by a motor vehicle. Will your caution change the way the
> > > driver drives and the speed and position of the impact? So in the vast
> > > majority of accidents you will be no different from anyone else because
> > > events are not in your control - if they were you would have avoided the
> > > accident in the first place.
> > >
> > > If you had an illness and the doctor offered you a pill that he said had
> > > performed well in limited clinical trials but that recent studies had
> > > shown that there was a worrying number of deaths in people on the pill
> > > would you take the pill and say, "I'll be alright doc, I'm not other
> > > people"?
> > >
> > I would be less worried if I knew that 90% of those who died were, say,
> > diabetics and I wasn't. Of course, we can't let those pesky details get
> > in the way of a good generalization, now can we?
> >
>
> Now, what characteristic of yourself is it that makes you believe that you
> are so different from the average person (who will encounter death once
> every 450 years of cycling non-stop 24 hours a day) that wearing a helmet
> while cycing is worth doing?

Helmets protect against more than fatal brain injuries. But seeing your
statistic, I don't think I'll ever get off my bike again. After all,
the population averages suggest I'll live for 450 years! ;-)
>
> And why does this characteristic not apply while walking, as you've told us
> that you would not wear a helmet if walking in such a set of circumstance
> that the risk of head injury was greater than spending the same amount of
> time cycling?

I don't recall posting that. Could you kindly point out where I did?
The context may help in my answer.

Rick

Re: I crash into religion

p.k. wrote:
> frkrygow@gmail.com wrote:
> > What you're doing is postulating explanations that belie the actual
> > data, hoping to justify your point of view. But nothing justifies
> > your explanations! Really, the idea that you happen to like them
> > isn't sufficient!
> >
> > It's at least as likely that dedicated cyclists who choose not to stop
> > cycling after a MHL are the ones who are better at anticipating and
> > avoiding motorist mistakes, thus having fewer accidents that are the
> > motorists' fault; and who are more apt to obey traffic laws, thus
> > having fewer accidents that are their own fault.
>
>
>
> so you agree that there are a range of possible explanations for the
> population level data?

Of course. ANY phenomenon has a range of _possible_ explanations. The
likelihood of validity varies greatly, however. We can't blindly say
any possible explanation invalidates observed data.

It's also worth noting that in most instances, extra knowledge of
special cases merely fine-tunes the original rule, it doesn't
invalidate it. As an example, Einstein's Relativity doesn't really
prove Newton wrong, it just fine tunes Newton for extreme velocities.
Engineers and scientists still use Newton for all calculations at
ordinary velocities.

>
> I'd be interested to see data on the breakdown of cyclist head injury KSI's
> between different types of cyclist.

It would be interesting, but there's less and less value (and often,
more and more expense) in getting data for smaller and smaller
subgroups.

I have in my other office data from a survey by Moritz of members of
the League of American Bicyclists (an organization roughly comparable
to the CTC - although politically, much less effective). Moritz's
survey had experienced cyclists (something like 2500 miles per year)
experiencing a crash that caused minimum $50 damage or receiving some
medical attention once every 11 years. Median cost of the medical
attention was, IIRC, only $150 or so. I can check these figures in
detail later, but the crash frequency was certainly low, and the
typical crash was certainly rather mild, even though Moritz defined
them as "serious."

In _Effective Cycling_, John Forester gives much information without
citation. But he claims that crash rates drop dramatically with
experience.

> Is it the novice who falls off who makes up the bulk of the data or is it
> the 5,000 mile a year fast traffic commuter who is knocked off at speed?
> Are cycling accidents cyclists' fault or are the bulk the fault of
> motorists?
> Is it the inexperienced cyclist keeping to quiet roads who falls victim or
> is it the experienced roadie commuting in fast busy traffic?

FWIW, in America most "experienced roadies" do not commute by bike at
all! This is the land of the recreational cyclist. I belong to a 400
member bike club where 40 mile weekend rides are the norm, but AFAIK,
only about ten of us do a significant amount of utility riding.

- Frank Krygowski

Re: I crash into religion

David Damerell wrote:
> Quoting Espressopithecus (Java Man) <rickk@letterectomyTELUS.net>:
>> damerell@chiark.greenend.org.uk says...
>>> Quoting Espressopithecus (Java Man) <rickk@letterectomyTELUS.net>:
>>>> You are applying population statistics to an individual, which is
>>>> an error in logic. We do not know the underlying risk
>>>> distribution for individuals in the population, nor do we know the
>>>> underlying risk variance.
>>> Surely if we don't know that then we can use the population
>>> statistics.
>> Not to predict risk for specific individuals -- unless you think
>> individual risk propensity is all clustered around a very narrow,
>> peaky distribution curve with a very low sigma.
>
> Technically true but absolutely not pertinent.
>
> Lacking any data about how the risk breaks down across the
> population, we certainly _should_ use the population-level stats.
>


but we do know something about how the risk breaks down.:
There is a lot of data from various studies showing about 30% of pedestrian
ksi have high blood aclohol level (around 80% of pedestrian ksi's between 10
pm and 2 am in one study),
That pedesrtian error is the cause of a higher proportion of pedestrian
ksi's than cyclist error is of cyclist ksi's
That pedesrtian ksi's peaks very highly in the 10-19 age band but that
cyclist ksi's are more evenly spread at later ages with a less pronounnced
teenage peak
That pedestrian ksi's peak again (but lower) in old age
etc

The population level data point to questions about how money should be spent
improving road safety, but say very little about the level of risk faced by
an individual in unknown circumstances.


pk

Re: I crash into religion

p.k. wrote:
> <ironic mode on>
>
> As you will be well aware, science never provides absolute certainty
> and this is widely exploited by the anti-brigade where the evidence is
> very strong but can never be absolute. Thus the anti helmet brigade are able
> to exploit lack of certainty in the statistics to undermine the case for
> helmets.
>
>

I think you are confusing the intrinsic lack of absolute certainty in
good science with the questionable results of poor science.



--
Tony

"The best way I know of to win an argument is to start by being in the
right."
- Lord Hailsham

Re: I crash into religion

In article <1147922718.391928.73950@j33g2000cwa.googlegroups.com>,
frkrygow@gmail.com says...
>
> Espressopithecus (Java Man) wrote:
> > In article <1147887422.839839.201590@u72g2000cwu.googlegroups.com>,
> > frkrygow@gmail.com says...
> > >
> > > Espressopithecus (Java Man) wrote:
> > > > The last time I went on a 25 km walk, there were no bicycle helmets, so
> > > > the question is moot.
> > >
> > > That sounds like a rather desparate evasion.
> >
> > Facts are not a desperate evasion.
>
> You were hairsplitting to avoid admitting you would never consider a
> helmet for a 25 km walk. In that context, the "fact" was merely a
> delaying tactic.

I am suspicious of the population statistics on risk while walking and
cycling, and don't use them to make decisions about my own exposure. I
am convinced that individual risk-taking behaviour varies greatly
between individuals, and that the underlying risk distribution in the
population has a very high sigma. Unfortunately I no longer have the
reference, but I reviewed ~ 25 years ago showing that construction
workers accepted workplace risks that were ~ 4 times the risks factory
workers would accept and 16 times what office workers would accept.
Similar studies I reviewed at the time showed similar high variability
in individual risk tolerance. With an underlying population risk
tolerance distribution affected by such high individual variances, I
don't think one can draw conclusions from population data that are
useful for individual decision making.
>
> > You are applying population statistics to an individual, which is an
> > error in logic.
>
> It's not an error in logic. If we accept your premise, we would have
> to give up 90% of modern medicine!

It might not be a bad idea if we could give up about half of that.

> Nearly every therapeutic treatment
> in medicine has been studied and justified using population data.

Many have not, and many squeaked in with such low efficacy that the drug
companies, for example, should not have been allowed to claim their
products are effective.

> While some can be fine-tuned by more specific study, you'll _never_
> hear "Since you're between 6'1" and 6'2" tall, weigh between 200 and
> 205 pounds, are blonde and wear a beard, this vaccine should work for
> you. Good thing you don't have red hair!"

It is well known that people do not respond to drugs and treatments
uniformly. The more we know about the odds, the better -- or at least
that is what my doctor tells me.
>
> And you seem to ignore the fact that helmet proponents rely on worse
> data - data taken from demonstrably self-selected groups - which they
> then apply to everyone. Helmet promotions never say "If you are white,
> male, have good insurance coverage, ride mostly on bike paths, fall
> mostly on soft surfaces in incidents that don't involve cars, you
> should always wear a helmet." That is, however, a _slightly_ more
> accurate interpretation of the study that generated the famed "85%"
> figure.

I have made none of the statements in your previous paragraph. Please
don't assume that because I debate with you that I take the same
position as everyone else who debates with you. That, too, is an error
in logic.
>
> Instead, what do we hear? "Always wear a helmet." As if it applies to
> _everybody_.

Not from me.
>
> > > Let's rephrase the question: Should helmets be recommended for people
> > > embarking on 25 km walks?
> >
> > That wasn't, and hasn't been, the point I was arguing.
>
> No, it's more like the point you were avoiding - which is that cycling
> is not significantly riskier than walking or motoring. Yet you support
> helmet use only for cycling.

No, I WEAR A HELMET WHEN CYCLING. It is a personal decision which I
base on my risk in my circumstances. Feel free to guide all your
personal decisions by gross population metrics if you wish. I'll use my
own discretion.
>
> > >
> > > At present, they're obviously not. Browse any backpacking literature
> > > to see. (And those deal with - horrors! - isolated areas with insecure
> > > footing!)
> > >
> > > Should helmets be recommended for people embarking on 25 km runs?
> >
> > I don't know. What is the risk for people who are fit enough to run 25
> > km? Is it the same as the underlying risk for the entire population?
>
> Hmm. You're retreating into a defense from ignorance. "The available
> data shows one thing. But there's a slight possibility this subset is
> different, so we _must_ assume it _is_ different."

You're using population statistics slavishly, disregarding their utility
for individual decision making. Do you govern your entire life based on
population stats? For instance, do you drink the 1-2 glasses of wine
per day that the stats tell you are correlated with reductions in risk
of heart disease? If so, would you do it even if you didn't like wine?
I don't -- the data are not yet convincing. How about taking aspirin to
reduce the risk of stroke? Population studies show it does. Do you do
it? I don't -- there are individual risk factors that mitigate against
it. There are many more examples of questionable conclusions drawn from
population statistics. A key question in using population data to draw
conclusions about what specific individuals ought to do is the
underlying variance among individuals.
>
> > Surely there's enough data available about head injuries during
> > competitions like the Boston Marathon that you don't have to rely on the
> > population statistics to make your point?
>
> I doubt it. But we've produced data. Do you have some to prove your
> point? If so, let's see it.

Humour me. Do you think that head injury risk for participants in the
Boston Marathon is likely to be significantly different from population
data for head injuries to pedestrians? Would you use the gross
population stats to make a decision about whether or not participants in
organized marathons should wear helmets?
>
You've produced data that I doubt is useful for individual decision
making, just like a lot of other population data. When you assure me
that you use this data to make decisions about every facet of your life
-- not just bicycle helmet use -- I'll take your faith in the data more
seriously.

Rick

Re: I crash into religion

In article <1147922783.356051.131490@j55g2000cwa.googlegroups.com>,
frkrygow@gmail.com says...
>
> Espressopithecus (Java Man) wrote:
> > In article <4d1c87F187v9tU1@individual.net>, junk@raven-family.com
> > says...
> > > Espressopithecus (Java Man) wrote:
> > > >
> > > > Ignoring, of course, that helmets may be more protective in some types
> > > > of crashes and less in others.
> > >
> > > Which and what types of crashes are those and where is the evidence?
> >
> > The evidence on this is not firm because not enough studies have been
> > done. But my reading suggests that in lower speed crashes helmets may
> > be more effective than none, while in higher speed crashes, they may be
> > less effective than none.
>
> I'd like to see a citation.

So would I. Is there something misleading about "the evidence on this
is not firm"?

Rick

Re: I crash into religion

Peter Clinch wrote:
> Tony Raven wrote:
>
>> Agreed but that does not mean we have to accept Russell's Teapot
>> attacks on fundamentally sound research as evidence that our points
>> don't stand up and the case is flawed.
>
> As I said, I think p.k. generally overstates his case. But it's still a
> point to remember that one can have the right answer and still spout
> some questionable reasoning to justify it. If p.k. makes a point about
> there being some possible mechanisms for such and such that isn't
> necessarily an argument that such mechanisms are clearly in operation,
> rather that one should acknowledge they are possible and say why why one
> thinks they're not operating.
>

I think Simon addressed his proposed mechanisms up there somewhere and
showed they would support, not negate, the conclusions on helmets not
working so p.k shifted to another variant which still doesn't achieve
what he hopes for. He's also not willing to do one of the basic tests
of science; make a prediction and test it. I took him through the
prediction of his theory and provided him the data so he could do the
test. He ducked it. If he had an open scientific mind he would have
done the test and told us the results. But he didn't.

--
Tony

"If the facts don't fit the theory, change the facts."
- Albert Einstein

Re: I crash into religion

In article <OS0bg.22437$43.19706@nnrp.ca.mci.com!nnrp1.uunet.ca>,
jtaylor@deletethis.hfx.andara.com says...
>
> "Espressopithecus (Java Man)" <rickk@letterectomyTELUS.net> wrote in message
> news:MPG.1ed635d64405275e989798@shawnews.vc.shawcable.net...
>
> > >
> > > Now, what characteristic of yourself is it that makes you believe that
> you
> > > are so different from the average person (who will encounter death once
> > > every 450 years of cycling non-stop 24 hours a day) that wearing a
> helmet
> > > while cycing is worth doing?
> >
> > Helmets protect against more than fatal brain injuries. But seeing your
> > statistic, I don't think I'll ever get off my bike again. After all,
> > the population averages suggest I'll live for 450 years! ;-)
> > >
> > > And why does this characteristic not apply while walking, as you've told
> us
> > > that you would not wear a helmet if walking in such a set of
> circumstance
> > > that the risk of head injury was greater than spending the same amount
> of
> > > time cycling?
> >
> > I don't recall posting that. Could you kindly point out where I did?
> > The context may help in my answer.
> >
>
> Here you go...it was only yesterday afternoon that you wrote:
>
> <quote follows>
>
> > > > So when you do go for 25km walks, you do wear a helmet?
>
>
> > > The last time I went on a 25 km walk, there were no bicycle helmets, so
> > > the question is moot.
>
>
> > Suppose you were to go on a 25km walk tomorrow. Would you wear a helmet?
>
>
>
> No. I don't believe the population statistics are representative of my
> personal risk profile. Do you?
>
> <end quote>

So where dis I say I would "not wear a helmet if walking in such a set
of circumstance that the risk of head injury was greater than spending
the same amount of time cycling"?

You're putting words in my mouth.
>
> Now, about this "personal risk profile"; what is it that makes cycling so
> "dangerous" for you that does not apply while doing an activity (walking),
> which for the general population is as, or more, "dangerous"?
>
I don't think cycling is "dangerous" or I wouldn't do it. I do think
that walking where and how I do, I am at less risk of head injuries than
while cycling where and how I do. Based on my own experience -- which I
believe to be more appropriate to assess my personal risk than gross
population statistics -- I've suffered 4 head injuries while cycling,
and none while walking. That's one of the reasons I wear a helmet while
cycling but not while walking. What would you do under those
circumstances -- assume your own experience was a statistical anomally,
and that your REAL risk is the same as the general population? You're
welcome to do so -- it is your life after all -- but I certainly
wouldn't make those assumptions.

Rick

Re: I crash into religion

In article <zrh*5vYgr@news.chiark.greenend.org.uk>,
damerell@chiark.greenend.org.uk says...
> Quoting Espressopithecus (Java Man) <rickk@letterectomyTELUS.net>:
> >damerell@chiark.greenend.org.uk says...
> >>Quoting Espressopithecus (Java Man) <rickk@letterectomyTELUS.net>:
> >>>You are applying population statistics to an individual, which is an
> >>>error in logic. We do not know the underlying risk distribution for
> >>>individuals in the population, nor do we know the underlying risk
> >>>variance.
> >>Surely if we don't know that then we can use the population statistics.
> >Not to predict risk for specific individuals -- unless you think
> >individual risk propensity is all clustered around a very narrow, peaky
> >distribution curve with a very low sigma.
>
> Technically true but absolutely not pertinent.
>
> Lacking any data about how the risk breaks down across the population, we
> certainly _should_ use the population-level stats.

Do you guide all your decisions in life based on population stats? When
you can assure me that you do, I'll have more confidence that your
argument above is serious.
>
> Imagine we have a drug that provides some trivial health benefit but kills
> 25% of the recipients. What we don't know yet is how that risk maps across
> the population; secretly, it actually kills 50% of men and no women at
> all. Now the population statistic does not predict any individual's risk -
> which is either nothing or 50%, obviously - but if we don't know that
> secret it is still quite right for someone considering taking the drug to
> base their decision on a 25% chance of being killed.
>
Imagine that we have statistics showing a correlation between moderate
daily wine drinking and reductions in premature death from heart
disease. Does that mean we should all drink wine daily? Do you?

Rick