Log in

View Full Version : The Collatz Conjecture


sqishy
October 22nd, 2015, 06:10 PM
I'm doing this thread to show a bit of mathematics that is not well known, yet relatively simple. In this context, a conjecture is a conclusion which has been made on a question, but with only incomplete information. A conjecture is not a proof; a proof is a definite answer, while a conjecture is an educated guess in a way. Lothar Collatz is the person who first put this conjecture out, hence it's called the Collatz Conjecture.

So what is the conjecture about? It is based off two very simple rules.

Take an natural number. Call this number X.

If X is even, divide it by 2. [IF X EVEN: X/2]
If, though, X is odd, then multiply it by 3, then add 1. [IF X EVEN: 3X + 1]

Repeat these steps forever.


It might be intuitive to think that these simple rules will give simple results. They don't, however. Though all natural numbers tested out so far have got to a stage where they reach 4,2,1,4,2,1... in a loop, many numbers take a long and windy journey to get to that, and the length of that journey is not easily relatable to the size of the number you start off with.

For example, if you start with the number 19, it goes as follows:
19, 58, 29, 88, 44, 22, 11, 34, 17, 52, 26, 13, 40, 20, 10, 5, 16, 8, 4, 2, 1.
The largest value reached was 88, and the sequence is 21 steps long.

The number 27, though, goes like this:
27, 82, 41, 124, 62, 31, 94, 47, 142, 71, 214, 107, 322, 161, 484, 242, 121, 364, 182, 91, 274, 137, 412, 206, 103, 310, 155, 466, 233, 700, 350, 175, 526, 263, 790, 395, 1186, 593, 1780, 890, 445, 1336, 668, 334, 167, 502, 251, 754, 377, 1132, 566, 283, 850, 425, 1276, 638, 319, 958, 479, 1438, 719, 2158, 1079, 3238, 1619, 4858, 2429, 7288, 3644, 1822, 911, 2734, 1367, 4102, 2051, 6154, 3077, 9232, 4616, 2308, 1154, 577, 1732, 866, 433, 1300, 650, 325, 976, 488, 244, 122, 61, 184, 92, 46, 23, 70, 35, 106, 53, 160, 80, 40, 20, 10, 5, 16, 8, 4, 2, 1.
The largest value reached was way up at 9232, and the sequence is 111 steps long.

255 is the next number which has a larger sequence than 27 when it comes to reaching the looping 4-2-1, and there are bigger and bigger sequences with certain bigger numbers.

The the question behind conjecture is this: Does every natural number, when subject to these rules, eventually reach 1, or does there exist a natural number that never reaches 1 and goes to infinity or some other pattern? The Collatz Conjecture says that every number will reach 1 eventually.

As I said at the start, no proof has been found which confirms or denies this conjecture. Every natural number up to 2 to the power of 60 (also written as 2^60, or just itself as 1,152,921,504,606,846,976) has been tested and shown to eventually reach 1, as of May of this year.
Whether the conjecture is proven or disproven, or if a proof can be found for or against it at all, is yet to be seen.

Just wanted to share this for whoever was interested, or bored, or whatever.

>>> https://en.wikipedia.org/wiki/Collatz_conjecture <<<

http://radar.oreilly.com/wp-content/blogs.dir/2/files/2011/10/1011-collatz-graph.png

Judean Zealot
October 22nd, 2015, 06:57 PM
As a matter of fact, Alan Turing's undecidability proof tells us that it is literally impossible to determine whether a given algorithm extends ad infinitum or reaches a peak with some rational number.

You have to remember, though, that Turing's undecidability proof only applies to machines following the standard linear model of computation (correct me if I'm wrong, I haven't had a formal education in mathematics, and as such I may have missed an angle of Turing-Church).

That might work fine with an absurdist like Vlerchan, but as a Mathematical Platonist (http://plato.stanford.edu/entries/platonism-mathematics/) (now that is an interesting discussion) I am fairly confident that there is a geometrical sort of ratio based proposition that can ultimately reach the essence of natural numbers, thus providing the tools necessary to avoid Turing's problem.

Emerald Dream
October 22nd, 2015, 07:01 PM
I would guess that every number would follow these rules....but it's just a guess.

It's funny that you should bring this up, because I remember reading a book when I was much younger (at least 7 years ago, though not exactly sure) that listed the rules you stated above as a "calculator game." I have done this (killing time) many times sitting at my desk with a calculator, and trying to get to 1 over and over again....and I still do it occassionally. There are many, many paths to get there (probably more than you would first think).

Even - divided by 2
Odd - times 3, then plus one :)

(I'm showing my nerdiness admitting that I play this)

sqishy
October 22nd, 2015, 07:10 PM
As a matter of fact, Alan Turing's undecidability proof tells us that it is literally impossible to determine whether a given algorithm extends ad infinitum or reaches a peak with some rational number.

You have to remember, though, that Turing's undecidability proof only applies to machines following the standard linear model of computation (correct me if I'm wrong, I haven't had a formal education in mathematics, and as such I may have missed an angle of Turing-Church).

That might work fine with an absurdist like Vlerchan, but as a Mathematical Platonist (http://plato.stanford.edu/entries/platonism-mathematics/) (now that is an interesting discussion) I am fairly confident that there is a geometrical sort of ratio based proposition that can ultimately reach the essence of natural numbers, thus providing the tools necessary to avoid Turing's problem.

My knowledge of mathematics is random, and I haven't heard of Turing's undecidability proof clearly, but bringing it up does help. There is indeed an educated consensus on the Collatz conjecture being undecidable, in terms of using algorithms available at the moment. I don't know anything reliable further than that, except that there are some ideas that proving the undecidability of the conjecture is itself impossible.

http://math.stackexchange.com/questions/1156004/how-could-collatz-conjecture-possibly-be-undecidable

https://www.quora.com/Why-is-proving-the-Collatz-Conjecture-considered-impossible

The second link does refer to Turing himself.
I hope my reply is helpful, if you needed it.

Mathematical platonism would be an interesting topic for sure; perhaps sometime soon I'll bring up a thread on that or something leading to it.

I would guess that every number would follow these rules....but it's just a guess.

It's funny that you should bring this up, because I remember reading a book when I was much younger (at least 7 years ago, though not exactly sure) that listed the rules you stated above as a "calculator game." I have done this (killing time) many times sitting at my desk with a calculator, and trying to get to 1 over and over again....and I still do it occassionally. There are many, many paths to get there (probably more than you would first think).

Even - divided by 2
Odd - times 3, then plus one :)

(I'm showing my nerdiness admitting that I play this)

I myself found out about this from a book by a mathematican Ian Stewart; it was either Professor Stewart's Cabinet of Mathematical Curiosities or Professor Stewart's Hoard of Mathematical Treasures, I honestly can't remember which one :P . I have them back home so I'll check it out on the weekend again.

Sir Suomi
October 22nd, 2015, 07:15 PM
Dude, I'm barely making it through Trig right now. You're killing me with this hypothetical math shit. :P

However, I think it's pretty neat, in a weird math-way.

Judean Zealot
October 22nd, 2015, 07:21 PM
I don't know anything reliable further than that, except that there are some ideas that proving the undecidability of the conjecture is itself impossible.



Hmm. That seems to be a sophistry, as Turing never tried asserting undecidability as an inherent aspect of natural numbers. He only published proof as regards a finite input of linear computation.

sqishy
October 22nd, 2015, 07:31 PM
Hmm. That seems to be a sophistry, as Turing never tried asserting undecidability as an inherent aspect of natural numbers. He only published proof as regards a finite input of linear computation.

I don't doubt that (our metaphorical 'friends') sophistry and rhetoric exists with some mathematicians. It does make more sense, at least intuitively, that undecidability is to be found in algorithms dealing with properties and 'processing' of numbers, than to be found in the numbers themselves. Epistemological versus metaphysical uncertainty/etc, in a way. The latter is making a bigger claim than the former. Mathematical platonism could come into this for sure.

phuckphace
October 22nd, 2015, 09:38 PM
don't have a compiler for this machine code but just posting to say I'm impressed when ppl can compile it in their heads

sqishy
October 23rd, 2015, 03:42 PM
don't have a compiler for this machine code but just posting to say I'm impressed when ppl can compile it in their heads

It's about staying away from Windows Vista mostly.