Surname survival statistics program and how to increase arti

Moderator: MOD_nyhetsgrupper

Svar
raylopez99

Surname survival statistics program and how to increase arti

Legg inn av raylopez99 » 25 sep 2007 13:24:23

See the below thread if you're interested in the probability of a
surname surviving. For Western populations having lambda (Poisson
mean) of 1.15 boys, the natural survival rate for surnames is in the
20%-30% region (that is, 70-80% of the time the surname dies out), but
you can increase the odds by (artificially) having more boys in the
first generation, see below.

RL

http://tinyurl.com/2qlzo3

I finished the beta version 2.0 of my program, which allows user
input. I'll double check later, but it seems to be working.

The ver 2.0 allows a user to input the first generation. For example,
if you say the first generation has 2 boys, you input two. You can
input a sequence as well (2,2,3 which means two boys, who each have
two and three boys, respectively).

Results (survival rate) are as follows: the survival rate went, for
Lambda 1.15, from the low to mid 20% region to as follows (100
simulations--but you really need to run 1000 to get stable results,
see my other posts):

2 boys (first generation) --> 39% survival rate
3 " " --> 60% " "
4 " " --> 64% " "
5 " " --> 78% " "
10 " " --> 91%

and the sequence of 1,2 (i.e. the first generation being told, say by
their father in a will, "you must have two boys to inhereit my money")

1,2 (one boy in the first generation having two boys) -->43% to 49%
(100 to 500 simulations).

In my next version of the program I'll see what the "winning
sequences" are in all probability--that is, what sequence of boys will
give a 50% surname survival rate? Clear 1,2 (see above) is one such
sequence, but there are others. For example, 2,1,0,3 is two boys, one
of whom has a boy (the other does not), which then has three boys--
does this sequence survive long? (I just ran this sequence, and for
100 simulations you get a 56% survival rate, so it does--and logically
it's the same as "3" above, so it's consistent with the above, given
the chaotic nature of Poisson sequences).

Hugh Watkins

Re: Surname survival statistics program and how to increase

Legg inn av Hugh Watkins » 25 sep 2007 21:27:26

But the same surname arises independently in many places

patronymics, locations, occupations . .

Hugh W



raylopez99 wrote:
See the below thread if you're interested in the probability of a
surname surviving. For Western populations having lambda (Poisson
mean) of 1.15 boys, the natural survival rate for surnames is in the
20%-30% region (that is, 70-80% of the time the surname dies out), but
you can increase the odds by (artificially) having more boys in the
first generation, see below.

RL

http://tinyurl.com/2qlzo3

I finished the beta version 2.0 of my program, which allows user
input. I'll double check later, but it seems to be working.

The ver 2.0 allows a user to input the first generation. For example,
if you say the first generation has 2 boys, you input two. You can
input a sequence as well (2,2,3 which means two boys, who each have
two and three boys, respectively).

Results (survival rate) are as follows: the survival rate went, for
Lambda 1.15, from the low to mid 20% region to as follows (100
simulations--but you really need to run 1000 to get stable results,
see my other posts):

2 boys (first generation) --> 39% survival rate
3 " " --> 60% " "
4 " " --> 64% " "
5 " " --> 78% " "
10 " " --> 91%

and the sequence of 1,2 (i.e. the first generation being told, say by
their father in a will, "you must have two boys to inhereit my money")

1,2 (one boy in the first generation having two boys) -->43% to 49%
(100 to 500 simulations).

In my next version of the program I'll see what the "winning
sequences" are in all probability--that is, what sequence of boys will
give a 50% surname survival rate? Clear 1,2 (see above) is one such
sequence, but there are others. For example, 2,1,0,3 is two boys, one
of whom has a boy (the other does not), which then has three boys--
does this sequence survive long? (I just ran this sequence, and for
100 simulations you get a 56% survival rate, so it does--and logically
it's the same as "3" above, so it's consistent with the above, given
the chaotic nature of Poisson sequences).



--
For genealogy and help with family and local history in Bristol and
district http://groups.yahoo.com/group/Brycgstow/

http://snaps4.blogspot.com/ photographs and walks

GENEALOGE http://hughw36.blogspot.com/ MAIN BLOG

raylopez99

Re: Surname survival statistics program and how to increase

Legg inn av raylopez99 » 26 sep 2007 09:02:29

The program is modeled after this process: http://en.wikipedia.org/wiki/Galton-Watson_process

It does not account for any "virtual" surname survivial such as
independent derivation.

RL

On Sep 25, 1:27 pm, Hugh Watkins <hugh.watk...@gmail.com> wrote:

Hugh Watkins

Re: Surname survival statistics program and how to increase

Legg inn av Hugh Watkins » 26 sep 2007 12:04:12

raylopez99 wrote:
The program is modeled after this process: http://en.wikipedia.org/wiki/Galton-Watson_process

It does not account for any "virtual" surname survivial such as
independent derivation.

RL

On Sep 25, 1:27 pm, Hugh Watkins <hugh.watk...@gmail.com> wrote:


my point

"There was concern amongst the Victorians that aristocratic surnames
were becoming extinct."

does it really matter?

Hugh W

--
For genealogy and help with family and local history in Bristol and
district http://groups.yahoo.com/group/Brycgstow/

http://snaps4.blogspot.com/ photographs and walks

GENEALOGE http://hughw36.blogspot.com/ MAIN BLOG

David Harper

Re: Surname survival statistics program and how to increase

Legg inn av David Harper » 26 sep 2007 12:33:48

Hugh Watkins wrote:
raylopez99 wrote:
The program is modeled after this process:
http://en.wikipedia.org/wiki/Galton-Watson_process

It does not account for any "virtual" surname survivial such as
independent derivation.

RL

On Sep 25, 1:27 pm, Hugh Watkins <hugh.watk...@gmail.com> wrote:


my point

"There was concern amongst the Victorians that aristocratic surnames
were becoming extinct."

does it really matter?

It remains a matter of grave concern to the noble families of England,
where the situation is now so desperate that many poor souls are now
forced to call themselves by such ludicrous names as Sir Ranulph
Twistleton-Wykeham-Fiennes. And to add to their suffering, their
parents play cruel practical jokes on them during their formative years,
such as convincing them that the correct pronunciation of "Ranulph" is
"Rayf".

The twisted, one might even say tortured, history of poor Sir Ranulph's
bizarre nomenclature is recounted here:

http://uncyclopedia.org/wiki/Sir_Ranulp ... am-Fiennes

David Harper
Cambridge, England

Nick

Re: Surname survival statistics program and how to increase

Legg inn av Nick » 01 okt 2007 23:21:19

"raylopez99" <raylopez99@yahoo.com> wrote in message
news:1190723063.412308.246700@w3g2000hsg.googlegroups.com...
See the below thread if you're interested in the probability of a
surname surviving. For Western populations having lambda (Poisson
mean) of 1.15 boys, the natural survival rate for surnames is in the
20%-30% region (that is, 70-80% of the time the surname dies out), but
you can increase the odds by (artificially) having more boys in the
first generation, see below.

RL

http://tinyurl.com/2qlzo3

I finished the beta version 2.0 of my program, which allows user
input. I'll double check later, but it seems to be working.

The ver 2.0 allows a user to input the first generation. For example,
if you say the first generation has 2 boys, you input two. You can
input a sequence as well (2,2,3 which means two boys, who each have
two and three boys, respectively).

Results (survival rate) are as follows: the survival rate went, for
Lambda 1.15, from the low to mid 20% region to as follows (100
simulations--but you really need to run 1000 to get stable results,
see my other posts):

2 boys (first generation) --> 39% survival rate
3 " " --> 60% " "
4 " " --> 64% " "
5 " " --> 78% " "
10 " " --> 91%

and the sequence of 1,2 (i.e. the first generation being told, say by
their father in a will, "you must have two boys to inhereit my money")

1,2 (one boy in the first generation having two boys) -->43% to 49%
(100 to 500 simulations).

In my next version of the program I'll see what the "winning
sequences" are in all probability--that is, what sequence of boys will
give a 50% surname survival rate? Clear 1,2 (see above) is one such
sequence, but there are others. For example, 2,1,0,3 is two boys, one
of whom has a boy (the other does not), which then has three boys--
does this sequence survive long? (I just ran this sequence, and for
100 simulations you get a 56% survival rate, so it does--and logically
it's the same as "3" above, so it's consistent with the above, given
the chaotic nature of Poisson sequences).


In my own family, my greatgrandfather who was from Russia changed his
surname on leaving Russia.

Therefore one can say that he founded the family with that surname - it can
generally be said that descendants are not related to anyone else of that
surname.

He married twice - his first wife dying.

By his second wife he had 13 children.

I am descended from the second wife.

Just dealing with the second wife - there were 7 daughters and 6 sons. 4 of
the daughters never married.

Of the 29 grandchildren, 7 were sons through the male line.

Those grandsons have 3 sons themselves (including myself and a first
cousin). None of us have any children as far as I know - the third cousin I
am not that familiar with.

In other words there will be no descendants in the male line past the 4th
generation - if you say that the 1st generation was my greatgrandfather.

Nick

raylopez99

Re: Surname survival statistics program and how to increase

Legg inn av raylopez99 » 01 okt 2007 23:36:14

On Oct 1, 3:21 pm, "Nick" <tulse04-ne...@yahoo.co.uk> wrote:

Just dealing with the second wife - there were 7 daughters and 6 sons. 4 of
the daughters never married.

Of the 29 grandchildren, 7 were sons through the male line.

Those grandsons have 3 sons themselves (including myself and a first
cousin). None of us have any children as far as I know - the third cousin I
am not that familiar with.

In other words there will be no descendants in the male line past the 4th
generation - if you say that the 1st generation was my greatgrandfather.


Having no male heirs to survive a surname, even with six or seven boys
initially, is not that unusual (happens 18-22% of the time), see this
thread, which updates the present thread:

http://groups.google.com/group/soc.gene ... 3e24?hl=en

RL

Nick

Re: Surname survival statistics program and how to increase

Legg inn av Nick » 02 okt 2007 11:23:00

"raylopez99" <raylopez99@yahoo.com> wrote in message
news:1191278174.150291.244000@d55g2000hsg.googlegroups.com...
On Oct 1, 3:21 pm, "Nick" <tulse04-ne...@yahoo.co.uk> wrote:

Just dealing with the second wife - there were 7 daughters and 6 sons. 4
of
the daughters never married.

Of the 29 grandchildren, 7 were sons through the male line.

Those grandsons have 3 sons themselves (including myself and a first
cousin). None of us have any children as far as I know - the third cousin
I
am not that familiar with.

In other words there will be no descendants in the male line past the 4th
generation - if you say that the 1st generation was my greatgrandfather.


Having no male heirs to survive a surname, even with six or seven boys
initially, is not that unusual (happens 18-22% of the time), see this
thread, which updates the present thread:

http://groups.google.com/group/soc.gene ... 3e24?hl=en

Does this figure apply to after 4 generations?

Interesting.

Nick

raylopez99

Re: Surname survival statistics program and how to increase

Legg inn av raylopez99 » 02 okt 2007 14:44:38

On Oct 2, 3:23 am, "Nick" <tulse04-ne...@yahoo.co.uk> wrote:

Having no male heirs to survive a surname, even with six or seven boys
initially, is not that unusual (happens 18-22% of the time), see this
thread, which updates the present thread:

http://groups.google.com/group/soc.gene ... _thread/...

Does this figure apply to after 4 generations?


Yes, the number of generations have no bearing on whether or not
extinction occurs, I just thought it was interesting (see the other
thread) that even with 9 or 10 boys, you can have the family surname
go extinct in as little as four generations (I could not find one in
three generations in my simulations, but probably that's occurred at
least once in practice).

RL

Nick

Re: Surname survival statistics program and how to increase

Legg inn av Nick » 02 okt 2007 15:21:38

"raylopez99" <raylopez99@yahoo.com> wrote in message
news:1191332678.450323.261870@n39g2000hsh.googlegroups.com...
On Oct 2, 3:23 am, "Nick" <tulse04-ne...@yahoo.co.uk> wrote:

Having no male heirs to survive a surname, even with six or seven boys
initially, is not that unusual (happens 18-22% of the time), see this
thread, which updates the present thread:

http://groups.google.com/group/soc.gene ... _thread/...

Does this figure apply to after 4 generations?


Yes, _the number of generations have no bearing on whether or not
extinction occurs_, I just thought it was interesting (see the other
thread) that even with 9 or 10 boys, you can have the family surname
go extinct in as little as four generations (I could not find one in
three generations in my simulations, but probably that's occurred at
least once in practice).

Are you sure about the comment that I have highlighted?

I might add that I am a statistician.

Clearly once the name has become extinct it can't become extinct - but there
must be a probability that the name becomes extinct after 1, 2, 3
generations etc.

That would also give a marginal probability ie the Probability that it
became extinct after 2 generations, given that it didn't become extinct
after 1 generation and so on.

I would therefore assume that there is a probability of becoming extinct at
each succeeding generation - I don't believe that the number of generations
would have no bearing on the case - although I can see that there might be a
probability of extinction over all generations.

Nick

raylopez99

Re: Surname survival statistics program and how to increase

Legg inn av raylopez99 » 03 okt 2007 16:09:37

On Oct 2, 7:21 am, "Nick" <tulse04-ne...@yahoo.co.uk> wrote:
"raylopez99" <raylope...@yahoo.com> wrote in message

Having no male heirs to survive a surname, even with six or seven boys
initially, is not that unusual (happens 18-22% of the time), see this
thread, which updates the present thread:

http://groups.google.com/group/soc.gene ... _thread/...

Does this figure apply to after 4 generations?

Yes, _the number of generations have no bearing on whether or not
extinction occurs_, I just thought it was interesting (see the other
thread) that even with 9 or 10 boys, you can have the family surname
go extinct in as little as four generations (I could not find one in
three generations in my simulations, but probably that's occurred at
least once in practice).

Are you sure about the comment that I have highlighted?

I might add that I am a statistician.

Clearly once the name has become extinct it can't become extinct - but there
must be a probability that the name becomes extinct after 1, 2, 3
generations etc.

That would also give a marginal probability ie the Probability that it
became extinct after 2 generations, given that it didn't become extinct
after 1 generation and so on.

I would therefore assume that there is a probability of becoming extinct at
each succeeding generation - I don't believe that the number of generations
would have no bearing on the case - although I can see that there might be a
probability of extinction over all generations.

Nick

Yes, I am talking about the trivial case of either you are extinct, or
not extinct; I'm not referring to the marginal probability.

Actually I could use your help for a future iteration of the program,
where I might calculate marginal probability, but for the moment I
wrote this program as a C# training exercise. THe way I presently
calculate probability is as follows:

I first run N simulations (N=1000 to 6000) of "Monte Carlo" type
scenarios, where based on a Poisson array, either the family does or
does not go extinct; I define non-extinction as having over 10000 male
offspring from the initial parent node--that is, 10000 boys, dead or
alive--figuring if you go past this arbitrary 10k limit likely the
tree will grow forever (an approximation, but I had to stop the N-ary
tree somewhere or it would have crashed my PC).

Then, for each such simulation i of N, I record the value: extinct or
not?

Finally, I simply take a ratio of the sum of non-extinctions over all
simulations.

I find that the results are as reported in the other thread. For six
or seven initial boys (first node) extinction happens 18-22% of the
time.

BTW, I read that in Thailand, some kings took multiple wives (as in 40
wives!) to guard against extinction of the male line.

Please do tell me how to calculate marginal probability for Poisson
distributions--I take it a Poisson is not path dependent for Markov
chain purposes, it's like flipping a coin, but the process described
herein clearly is path dependent (what's happened before influences
the present). As I posted elsewhere, the above process is very
variable, but does converge to the mean if you run enough simulations.

RL

Svar

Gå tilbake til «alt.genealogy»