Very nice.

- Pauli

Thanks!

It took me only five minutes or so to find this one on the 12C:

3.141592654

ENTER

1 e^x /

ENTER ENTER ENTER; now press * n times until something interesting shows on the display:

* => 1.335705708 not close enough to sqrt(sqrt(pi)) -> ignore

* => 1.543711618 don't recognize -> ignore

* => 1.784109737 not close enough to sqrt(pi) -> ignore

* and so on...

*

*

* => 3.183047554 From memory, close to 10/pi -> near idendity (pi/e)^8 ~ 10/pi, but useless for my purpose

*

*

*

* => 5.678906136 -> =(pi/e)^12. Since I have 12 and 56789, decide to try a pandigital approximation (only 3 and 4 missing).

Luckily e^-12 = e^(-3*4) does the job nicely! :-)

Gerson.

Five seconds on a 48G gave me this: 1146408/364913

The old regular 22/7 is a bit easier to remember.

Mark

But neither of them use the digits 1 through 9 in order :-)

- Pauli

Archimedes' upper bound for Pi is better because it was obtained by reasoning rather than sheer luck.

MathWord has plenty of Pi approximations to choose from:

http://mathworld.wolfram.com/PiApproximations.html

Regards,

Gerson.

-------

P. S.: The 33s gives instantly this even better and still easy to remember approximation:-)

355

Pi ~ ---

113

*Edited: 22 July 2009, 7:35 p.m. *

Quote:HP-35 users knew that one before (mentioned in its manual) ;^).

P. S.: The 33s gives instantly this even better and still easy to remember approximation:-)

355

Pi ~ ---

113