
Number Theory Number Theory Math Forum 
 LinkBack  Thread Tools  Display Modes 
May 17th, 2019, 03:04 PM  #1 
Newbie Joined: May 2019 From: Spokane Posts: 2 Thanks: 0  Commonality of Normal Numbers?
Disclaimer...I'm artist not a mathematician By the definition of Normal Numbers I'm left with the question, how large of digit strings are shared by all normal numbers? For example the string 14639 should appear in a normal number and in all normal numbers and should likely or inevitably show up more than once (infinitely?). So, by extension it seems that an infinitely long string should also obey this and then, if that is the case, how is one normal number different than another? Is only the starting place different? Could 314... Be found somewhere in all normal numbers? Sorry I don't know enough math to ask this concisely. 
May 17th, 2019, 08:40 PM  #2 
Senior Member Joined: Sep 2015 From: USA Posts: 2,427 Thanks: 1314 
what is a "normal number" ?

May 17th, 2019, 08:49 PM  #3 
Newbie Joined: May 2019 From: Spokane Posts: 2 Thanks: 0 
Ha ok, I guess I'm referring to an irrational number that is normal  that has an even distribution

May 17th, 2019, 10:15 PM  #4 
Senior Member Joined: Aug 2012 Posts: 2,306 Thanks: 706 
First, a disjunctive number is a real number whose decimal expansion contains every possible finite sequence of digits. Somewhere in its infinite decimal expansion we can find 42343243243, and the first 11 million digits of pi, the ASCII encoding of the complete works of Shakespeare, and an ASCII description of each minute of your entire life from the moment of your birth to the moment of your death. [ASCII is the standard way of representing letters of the alphabet in binary code]. Such a number is often  very often, in fact  mistakenly called a normal number. A normal number is disjunctive, to start with. But more than that, a normal number's decimal representation contains each possible finite sequence a statistically equal number of times. In other words each digit occurs on average 1/10 of the time, as you look at more and more digits and take the limit to infinity. Each 2digit combination occurs 1/100 of the time; each 3digit combination occurs 1/100 of the time, and so forth. It's clear that each finite sequence of digits must recur infinitely many times somewhere in the decimal representation. It's been proven that a randomly chosen real number is normal with probability 1. That means that there do exist nonnormal reals, but they form a set of measure 0. They are statistically very rare. Strangely, we can easily produce nonnormal numbers. 1/3 = .333333333... is obviously one such and you can easily make up similar examples. Even more strangely, nobody has ever proven that any specific real number is normal. We think but we don't know that e and pi are normal. To answer the OP's question, of course normal numbers can all be different from each other. Just because they have to statistically contain every finite sequence in equal proportions, they can still be very irregular for a long time out. For example there could be some normal number that begins .333333333333...3 where there are a million 3's; but still ends up still being normal, because by the time you take the statistical frequencies all the way to the end, every finite sequence ends up occurring with equal frequency. [What I described is normal to base 10. A number might be normal in one base but not another. The actual definition of a normal number is that it is normal to every base]. Last edited by Maschke; May 17th, 2019 at 10:44 PM. 

Tags 
commonality, irrational, normal, numbers 
Thread Tools  
Display Modes  

Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
normal numbers and probability  dannyh532  Number Theory  5  June 26th, 2017 10:05 PM 
Probability Of Consecutive Numbers When Numbers Are Paired  EvanJ  Advanced Statistics  1  January 26th, 2015 05:32 AM 
normal to the curve/normal to the circle  cheyb93  Calculus  7  October 29th, 2012 01:11 PM 
Expectation of picking all N numbers from N distinct numbers  abcbc  Advanced Statistics  1  October 1st, 2012 04:14 PM 
Product normal implies components normal  cos5000  Real Analysis  3  November 18th, 2009 08:52 PM 