So a news story turned up today with the headline "Pi record smashed as team finds two-quadrillionth digit" which I was naturally interested in. This new digit of Pi was found by means of some clever method of computing digits separately, which is kind of interesting. But then I got to the meat of the issue, what was the two-quadrillionth digit of Pi. It turns out that it is a 0, in binary.
In binary? Really? If we're doing it in binary you've got a fifty fifty chance of getting it bloody right. But that's not my major complaint against doing it in binary. My gripe is with claiming this is the two-quadrillionth digit of Pi. When I count digits of Pi I count decimal digits after the decimal point (the three is too easy and doesn't count). Pi to five digits is 3.14159. In binary the same number is 11.001001000011111 which is just a tad longer, and could have been much longer (my first run to calculate the binary value of 3.14159 got to over 350000 digits before I killed it for taking to long. I then decided to limit it to 15 digits for this example). My point is that calculating a bunch of binary digits is really easy and doesn't match what we normally consider the number of digits in a number. So calling this the two-quadrillionth digit of Pi is a bit of an exaggeration. In fact, I'd call it a bit of bullshit.