What is Blitz3D's floating point accuracy? Can it handle calculations with 16 decimal places? Blitz3D will surely round the answer - but does it calculate with 16 decimal place accuracy in the first place?
Single-Precision, which equates to a reasonable degree of accuracy with 7 digits ( including those on *both* sides of the floating point as Ross demonstrates above. )
If I recall correctly, there were some "unique qualities" to B3D's floating point precision which made it about 1 digit less accurate than the same code in other languages with single-precision floating point operations. So on second thoughts, make that six, not seven, unless those issues were fixed.