X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Sender: To: lml Date: Tue, 01 Aug 2006 14:31:48 -0400 Message-ID: X-Original-Return-Path: Received: from imo-d23.mx.aol.com ([205.188.139.137] verified) by logan.com (CommuniGate Pro SMTP 5.1c.2) with ESMTP id 1331776 for lml@lancaironline.net; Tue, 01 Aug 2006 11:04:47 -0400 Received-SPF: pass receiver=logan.com; client-ip=205.188.139.137; envelope-from=Sky2high@aol.com Received: from Sky2high@aol.com by imo-d23.mx.aol.com (mail_out_v38_r7.6.) id q.4a3.51380500 (63024) for ; Tue, 1 Aug 2006 11:03:49 -0400 (EDT) From: Sky2high@aol.com X-Original-Message-ID: <4a3.51380500.3200c755@aol.com> X-Original-Date: Tue, 1 Aug 2006 11:03:49 EDT Subject: Re: [LML] Safety statistics and other mythology X-Original-To: lml@lancaironline.net MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="-----------------------------1154444629" X-Mailer: 9.0 Security Edition for Windows sub 5318 X-Spam-Flag: NO -------------------------------1154444629 Content-Type: text/plain; charset="US-ASCII" Content-Transfer-Encoding: 7bit Hmmmmm........ A new soap opera: As the Thread Turns Brent's view of statistics (the plural of statistic) is from an engineer's perspective. Lorn didn't go far enough. Terrence used 20 year old information and threw in some strongly held personal views. Others have sought to explain certain sub-sets of the universe of air accidents by comparing dissimilar groupings. If you have gotten to this paragraph perhaps you will continue. Note that statistic is defined in an online dictionary as follows: 1. A numerical datum. 2. A numerical value, such as standard deviation or mean, that characterizes the sample or population from which it was derived. While it took me eleven years to get a simple BS (yes, that is the correct expression) in Math, the middle years (1960's) were devoted to computers, with some time spent writing programs on vacuum tube and transistor computers to turn vast quantities of #1 into a wide variety of #2 (yes, #2). Interestingly, statistics alone should not be our real concern. LML contributors so far have tried to perform the function of the insurance companies (Lorn, jump in any time). To wit, assigning a probability to an event which causes the death of an aircraft occupant when the reported statistic is grossly limited to gigantic subpopulations such as GA and Commercial Aviation. There are mathematical methods to determine the probability of heads on the next coin toss and even more complex event outcomes based on intervening conditions. I wonder if this is where Brent meant a lower branch of math? After all, a form of this math is applied to "theoretical" physics (Lorn, hit'm with the random nature of #1 above) to help explain various natural wonderments. I gave up on studying probability theory when the conditions got too complex. The insurance companies have not given up on conditional probabilities (this must be what they call actuarial science). Their motive is straight forward - What premium should be charged to some population so that when certain covered events occur, less is paid out than taken in. When a person obtains insurance coverage for an air event that leads to death and destruction (crash among others), the company tries to limit the population so that their motive is met. For example, a transitionally trained pilot with a minimum number of hours in similar type that is retrained each year and is young enough without too many DUI's can get such protection while someone falling outside that population is eliminated or must fit into a population where higher premiums are paid. If you are still with this piece, please note that the probability of your immediate demise is far more complex than using general aviation population statistics combined with the wing design of the airplane flown. There are so many ways to go. For example, in the greater population, the category "home accidents" is a killer, but since you built your plane and still have all your fingers and toes the chances are you won't perish that way (Hmmmm, an opinion, not a statistic). Now, to get personal I must describe the subpopulation I exist in - I drive a performance car without side airbags and use speed to get out of dicey situations, I frequently travel expressways in excess of 75 mph (65 Kts) on a motorcycle, I am overweight, I eat and drink almost anything in moderation, I operate dangerous home garden manicure equipment, I fly a hi performance airplane that I built in my garage and I don't challenge Mother Nature too much. I also avoid flying commercial airlines, especially those that keep turning up with drunk pilots or are based in certain exotic locations. Oh, I clean my own guns and I don't allow terrorists to board my aircraft either. So, is the probability less or greater that I will perish in my Lancair or of that because I am a member of the GA population? I don't care because I am a special case. Besides the advanced institutional care I receive and ever higher med doses, I respect the flight characteristics of my airplane. And that alone apparently puts me in a different population from those that think they are still flying a Cessna whilst directing their hi-perf experimental craft about the sky. Risk analysis is different for the individual or the insurance company because of very different interests. Statistics are useful to the latter and not particularly useful to the former. Scott Krueger AKA Grayhawk Lancair N92EX IO320 SB 89/96 Aurora, IL (KARR) Abnegate Exigencies! -------------------------------1154444629 Content-Type: text/html; charset="US-ASCII" Content-Transfer-Encoding: quoted-printable
Hmmmmm........  A new soap opera: As the Thread Turns
 
Brent's view of statistics (the plural of statistic) is from an enginee= r's=20 perspective.  Lorn didn't go far enough.  Terrence used 20 year ol= d=20 information and threw in some strongly held personal views.  Others hav= e=20 sought to explain certain sub-sets of the universe of air accidents by=20 comparing dissimilar groupings.
 
If you have gotten to this paragraph perhaps you will=20 continue.  Note that statistic is defined in an online dictionary=20= as=20 follows:
 
  1. A numerical datum.=20
  2. A numerical value, such as standard deviation or mean, that characteri= zes=20 the sample or population from which it was derived.
While it took me eleven years to get a simple BS (yes, that is the corr= ect=20 expression) in Math, the middle years (1960's) were devoted to computers, wi= th=20 some time spent writing programs on vacuum tube and transistor computer= s to=20 turn vast quantities of #1 into a wide variety of #2 (yes, #2). =20 Interestingly, statistics alone should not be our real concern.
 
LML contributors so far have tried to perform the function of the insur= ance=20 companies (Lorn, jump in any time).  To wit, assigning a probability to= an=20 event which causes the death of an aircraft occupant when the reported stati= stic=20 is grossly limited to gigantic subpopulations such as GA and Commercial=20 Aviation.  There are mathematical methods to determine the probability=20= of=20 heads on the next coin toss and even more complex event outcomes based on=20 intervening conditions.  I wonder if this is where Brent meant a lower=20 branch of math?  After all, a form of this math is applied to "theoreti= cal"=20 physics (Lorn, hit'm with the random nature of #1 above) to help explain var= ious=20 natural wonderments.  I gave up on studying probability theory whe= n=20 the conditions got too complex.
 
The insurance companies have not given up on conditional probabilities=20 (this must be what they call actuarial science).  Their motive is strai= ght=20 forward - What premium should be charged to some population so that=20 when certain covered events occur, less is paid out than taken=20 in.  When a person obtains insurance coverage for an air event tha= t=20 leads to death and destruction (crash among others), the company tries to li= mit=20 the population so that their motive is met.  For example, a=20 transitionally trained pilot with a minimum number of hours in similar type=20= that=20 is retrained each year and is young enough without too many DUI's can g= et=20 such protection while someone falling outside that population is eliminated=20= or=20 must fit into a population where higher premiums are paid.
 
If you are still with this piece, please note that the probability= of=20 your immediate demise is far more complex than using general aviat= ion=20 population statistics combined with the wing design of the airplane flown.&n= bsp;=20 There are so many ways to go.  For example, in the greater population,=20= the=20 category "home accidents" is a killer, but since you built your plane and st= ill=20 have all your fingers and toes the chances are you won't perish that way (Hm= mmm,=20 an opinion, not a statistic).
 
Now, to get personal I must describe the subpopulation I= =20 exist in - I drive a performance car without side airbags and use speed to g= et=20 out of dicey situations, I frequently travel expressways in excess of 75 mph= (65=20 Kts) on a motorcycle, I am overweight, I eat and drink almost anything in=20 moderation, I operate dangerous home garden manicure equipment, I fly a hi=20 performance airplane that I built in my garage and I don't challenge Mother=20 Nature too much. I also avoid flying commercial airlines, especially th= ose=20 that keep turning up with drunk pilots or are based in certain exotic=20 locations.  Oh, I clean my own guns and I don't allow terrorists to boa= rd=20 my aircraft either.  
 
So, is the probability less or greater that I will perish in my=20 Lancair or of that because I am a member of the GA population?&nbs= p; I=20 don't care because I am a special case.  Besides the=20 advanced institutional care I receive and ever higher med doses, I resp= ect=20 the flight characteristics of my airplane.  And that alone apparen= tly=20 puts me in a different population from those that think they are still flyin= g a=20 Cessna whilst directing their hi-perf experimental craft about the sky.
 
Risk analysis is different for the individual or the insurance company=20 because of very different interests.  Statistics are useful to the= =20 latter and not particularly useful to the former.
 
Scott Krueger=20 AKA Grayhawk
Lancair N92EX IO320 SB 89/96
Aurora, IL=20 (KARR)

Abnegate Exigencies!
 
 
 -------------------------------1154444629--