What I'd like to know, Herr Sklansky...

Is why in the hell 112 is about the right age for the previously posted '50-50' proposition.

...it's funny you asked. For reasons unknown I thought about this today between hands (I have the attention span of a two year old and a wandering mind) and realized that the answer is the age where the population of people at that precise age is twice that of the population of people that age plus one year. Looking at it this way it is clear that the correct answer must be a very large number, although 112 would have been a wild guess for me, Sklansky, or anybody without an almanac.


Regards,


Rick

"the answer is the age where the population of people at that precise age is twice that of the population of people that age plus one year."


You're suggesting there are twice as many 112 year olds as there are 113 year olds.


Don't you mean to suggest there are twice as many 112 year olds as there are people who are at least 113 years old? Small differences can be important.

Both of us may be off but the concept makes it clear the number must be very high. Maybe a better way to say it is to have a life expectancy of 1 year the average age of all remaining people must be 113. Of course at this age the population is small and you probably need to measure life span in days and divide by 365.


BTW I just Googled something from the government you may want to look at - see link below). They lump everyone over 100 together and assign a life expectancy of 2.6 years. Perhaps deeper in the data is information about 112 year olds (I wish Mark Glover was here to find it).


Regards,


Rick

Rick,


very good reference chart. They do stop at 100+ but if you axtrapolate the data out to age 111 using the approximate 1.5 mths shorter age span per year from the ages of 90 to 100 I come up with age 111 where you have a 50/50 chance of living 1 more year.


Jimbo

Don't you mean to suggest there are twice as many 112 year olds as there are people who are at least 113 years old?


No, the way Rick said it was correct. Since 112 is the answer, we know that about half of 112 years will live for another year. Therefore, at any given time the number of 113 year olds should be half the number of 112 year olds.

The ratio of living 113 year olds has no relationship to the number of living 112 year olds. What is more relevant to that ratio is how many Americans were born in each year versus the other. Not that that ratio has any real value either unless the numbers are extremely out of whack due to some national emergency or natural disaster in one year or the other.


To elaborate just because there are 50 living 113 year old Americans does not mean there should be 100 living 112 year old Americans. It only means that the previous year there were 100 living 112 year old Americans and that half of them died in the following year.


Jimbo

Rick-


Goddamn your smart. I never would have thought of that. Even if you're not right, at least you've given me an answer that will get this off my mind.. /images/smile.gif


But, unlike you, I am very glad Mark Glover is not hear to help out.


I empathize with your attention span problems.

No, your logic is wrong. Probabilities are founded on past data. For example, the way we know that a coin has a 50% chance of comming up heads is because somebody tried it zillions of times. Likewise, from experience we know that each face on a regular solid has an equal chance of landing on top when the solid is thrown. Since last year half of the 112 year olds died, then barring any additional data the probability of any given 112 year old dying within one year is about 50%.


-MD

If Mark were here he would tell you that here is spelled here and not hear as you spelled it.

Technically, of course you could make the argument that just because 50% of 112 year olds last year made it to 113 doesn't mean that the same will be true this year and that we don't know how likely a 112 year old today is to live to 113. But saying that past data isn't an accurate predictor of what will happen invalidates the entire original question Sklansky asked, so it seems pretty silly.

MD,


You do not really believe that line about the coin flip do you? The reason we know the probability is 50/50 is NOT because someone flipped a coin zillions of times. It is because there are only 2 possible (reasonably possible excluding landing on it's edge) outcomes for a coin flip.


Jimbo

Who's to say that each side is equally likely to come up?


-MD

every normal person knows its even steven to come up heads/tails.


what a million times has shown is that this belief is not incorrect.


(stuff about nafta/gatt omitted)


take free clean needle exchange progams for iv drug users. everyone knows this will result in more drug use and more disease. the fact that science and (lets try it!) shows the opposite just goes to show that its hard to get people to change their mind.


im sure that for the first 100,000 coin tosses or so , somebody was making money because (...).


brad

Well MD I believe you said so, I quote from your post above: "For example, the way we know that a coin has a 50% chance of comming up heads is because somebody tried it..."


I just cannot believe that you think probabliities are determined by trial and error lab tests. You must be pulling our proverbial legs right?


Jimbo

Brad,


just how does that make us both wrong? I miss your point in your post after your marvelous heading.


Jimbo

'what a million times has shown is that this belief is not incorrect.'


well, you say your belief (theory)(in probability) makes it so, MD says empirical data makes it so.


i think a case could be made that its the empirical evidence which *has not* *disproved* the theory that makes it so.


'im sure that for the first 100,000 coin tosses or so , somebody was making money because (...).'


because their false theories hadnt been disproved yet (say, for example that heads would come up much more often because that was the face of god, etc.)


note that the main point here is that the theory need not be correct for the empirical data not to disprove it.


brad


p.s. it = 50% heads, 50% tails

To this stupid question which he says is obvious, but which doesn't seem to me at all obvious.

sx

Well, at risk of continuing this admitadly lame thread, I wasn't really saying that probabilities can't be calculated. My point was that the result of your calculation implies some faith in certain "obvious" truths. For example, anybody can calculate the probability of 3 heads in a row at 1/8, but to believe it you have to believe that a head is just as likely as a tail on any given flip. As this relates to Sklansky's problem, to say that "Just because half of the 112 year olds died last year doesn't necessarily mean that there is a 50% chance of a certain one of them dying this year" is equivalent to saying that "Just because half of the coin tosses on Earth to date were heads doesn't necessarily mean that there is a 50% chance of heads on the next toss."


I hope this clears up my intended point.


-MD