On Tue, 6 Mar 2007 15:57:31 -0600, Adam Thornton 
<[EMAIL PROTECTED]> wrote:

>I'll give you a dollar if you can show me a base in which 54 is more
>than two and a half times as big as 27.

Converting to mathematical terms:

j=54
k=27
n=2.5

You want j/k > n, where j and k are in base b.

Expressing j and k in terms of base b gives:

j=54 base b = 5*b^1 + 4*b^0 = 5b+4
k=27 base b = 2*b^1 + 7*b^0 = 2b+7

Substituting into j/k > n:
(5b+4) / (2b+7) > n, where n>2.5


Now choose an n > 2.5, and solve the following equation for b:
(5b+4) / (2b+7) = n

When n=3,
(5b+4) / (2b+7) = 3
5b+4 = 3(2b+7)
5b+4 = 6b+21
b=-17

To double check, we substitute b=-17 into:
(5b+4) / (2b+7) > 2.5, giving:
(-85+4) / (-34+7) > 2.5
(-81)/(-27) > 2.5
3 > 2.5, which is indeed true.

For larger values of n, b asymptotically approaches -3.5.
For values of n=2.5+x, where x>0 and as x->0, b->negative infinity

QED.


A long time ago I read a mathematical paper on some interesting propertie
s 
of negative base number systems.  Now it's earned me a dollar.

Brian Nielsen

Reply via email to