A sequence of six 9s occurs in the decimal representation of ?, starting at the 762nd decimal place. These six 9s have come to be called the Feynman point.
N[Pi, 769] ends with …7211349999998
but
N[Pi, 768] ends with …721135000000
What gives?
A sequence of six 9s occurs in the decimal representation of ?, starting at the 762nd decimal place. These six 9s have come to be called the Feynman point.
N[Pi, 769] ends with …7211349999998
but
N[Pi, 768] ends with …721135000000
What gives?
If you allow just 768 digits,the one ending in 721135000000 is a better approximation of Pi than the one ending in 721134999999.
From the documentation:
“Pi can be evaluated to any numerical precision using N.”
Doesn’t seem like approximation is used.
If you have an exact number (like) Pi, then N can only give an approximation. And note that evaluate something with numerical precision x is the same as giving an approximation, that is the same.
As Gianluca says, this:
721134999999
is a worse approximation than:
721135000000
If you want to be sure to get non-rounded numbers, get a bunch more:
First[RealDigits[N[Pi, 800]]][[762 ;;]]