for doubles, VBA/Excel uses the IEEE standard which is approximately 15
digits of precision, so the error is usually around the last couple of
digits. How that translates into your "epsilon" would depend on your
numbers.

Here is an illustrative example of accumulating the results of imprecision:

Sub abcd()
Dim v As Double
v = 0#
For i = 1 To 300
v = v + CDbl(1 / 3)
Next
Debug.Print Format(v, "0.0000000000000000000")
Debug.Print Format(CDbl(1 / 3) * 300, "0.0000000000000000000")

End Sub


--
Regards,
Tom Ogilvy

"Edward Ulle" wrote in message
...
Tom,

I suspected as much but in your opinion, if I am dealing with small
numbers what is a minimum tolerance, say 10 to the minus 10 or is that
too small?


*** Sent via Developersdex http://www.developersdex.com ***