For those who are interested: It is conjectured that pi is "normal" in the sense that the digits occur with equal liklihood. Making that statement precise is

For those who are interested: It is conjectured that pi is "normal" in the sense that the digits occur with equal liklihood. Making that statement precise is exceedingly hard, and not very enlightening. EDIT: philh points out that the full definition of "normal" says that every block occurs equally often (a concept that needs care to make formally correct) and that does imply that every finite sequence does occur. pi is known to be transcendental. It is not only not rational (cannot be expressed as the ratio or integers (although 355/113 is close and you can get as close as you like by using bigger numbers)) but it is also not the root of any polynomial with integer coefficients (rationals being a special case). The digits of pi cannot become periodic. None of the above is enough to show that every finite number occurs "eventually". As an example, consider the number: 0.123 112233 111222333 111122223333 111112222233333 ... Here the digits 1 through 3 occur equally often, it never becomes periodic, it is transcendental, and yet the sequence 321 will never occur. Now instead of using 1 through 3, generalise to using 1 through 9. It is generally believed that the digits are "effectively random" which means any finite sequence will occur with probability 1, but the observations made above are not enough to ensure that. (EDIT: although the full definition of "normal" does, rather thanthe limited version I originally put - thanks to philh for the correction). (edited a typo 133 -> 113 : thank you for the correction)