it?

I'll leave Jon to reply to this fully but I have used Jon's suggestion of
NORMSINV(RAND()) to generate values from a normal distribution and I have
also used the noticeably faster sum of uniform deviates method taken as
[RAND() +RAND() +..to 12 values ] -6.  I have compared the results with
FREQUENCY and, with a few thousand calculated values,  FREQUENCY gives quite
a good bell-shaped curve with either method. It sounds a bit cumbersome but
it only takes a few seconds. There are polynomial approximations but they
seem very cumbersome for a one off calculation and another approach might be
to search on the Web for a tabulation.

it?

I may be mistaken as to why you want the bell-shaped curve but, if it's only
a one-off application, perhaps the best thing would be to type in 20-30
values from a standard tabulation and run a smoothed scatter plot thro'
them. There is not a simple formula, AFAIK.

Jim.

Gary -

Tushar Mehta shows the bell curve on his web site, http://tushar-mehta.com.

- Jon
-------
Jon Peltier, Microsoft Excel MVP
http://www.geocities.com/jonpeltier/Excel/index.html
_______