The Collatz Conjecture

Can it be proven?

Phoenix Woodford
 The Collatz Conjecture

CONTENTS
Introduction...........................................................................................................................................2
Visualising the Collatz Function.............................................................................................................3
Noticeable patterns and smaller proofs................................................................................................4
Proof using more conjectures................................................................................................................4

Introduction
Here’s a fairly simple problem for you:

Choose any positive integer x.

If x is even, divide x by 2.
If x is odd, multiply x by 3 then add 1.
Keep repeating this process.
What do you get to?

As of 2020, all natural numbers up to 295,147,905,179,352,825,856 eventually reach

1 in their sequence of these steps. The function below describes this process , and
shall be referred to as the ‘Collatz function’ or ‘function C’ throughout this document. It is
defined as:

{
C ( x )= x /2if x ≡0 (mod 2)
3 x+ 1if x ≡1(mod 2)

We can make a sequence, beginning with any positive integer x and putting it into
function C to find the next term. We can put this output back into the input of the
function, each time generating the next term in the sequence. The Collatz Conjecture is:

“This process will eventually reach the number 1, regardless of which positive integer is
chosen initially.”

No such proof for this conjecture has been found as of Sunday 9 th June, 2024. This
document is my personal attempt to look into this conjecture and break it down into
smaller problems that may be simpler to comprehend. I would like to clarify now that
several of my points may not be logically true, and if this is the case and you find any
mistakes or questionable statements throughout, please attempt to contact me at
[email protected]. I would be more than happy to discuss any queries or
feedback, since I am very new to this style of work. (I do not expect to get it right first
try.)
With that said, let us dive into the Collatz Problem.
Visualising the Collatz Function
To be able to prove the conjecture, we must first establish how the function works, and
find what is noticeable among its sequences.
I have chosen to express how integers are joined through the Collatz function using a
number tree. This is one example of a Collatz number tree:

256 85 84 80 26 24

128 42 40 13 12

64 21 20 6

32 10 3

16 5

In this diagram, positive integers are connected by arrows pointing in certain directions.
An arrow pointing downward represents a division by 2 if x is even, while an arrow
pointing sideways represents a multiplication by 3 followed by an addition of 1 when x is
odd. The arrow going from 1 to 4 is also a multiplication of 3 and addition of 1, and is a
loop between these 3 numbers.
This can show us how single sequences join with others if they meet at a certain integer.
These integers which follow from an odd number will be called ‘connectors’ because they
connect two separate paths into one.
For example, the sequence following C(3) is:
{3, 10, 5, 16, 8, 4, 2, 1}
While the sequence following C(128) is:
{128, 64, 32, 16, 8, 4, 2, 1}
These sequences join at the connector 16, and continue on to 8, 4, 2, and 1. This can be
seen in the diagram by following the arrows from the starting integers and ending at 1.
(Note: unless specified in examples, the sequences are terminated at 1, or else they
would continue indefinitely.)

In the diagram, the tree ‘grows’ upward and outward to the right, creating columns of
integers. These columns shall be referred to as ‘branches’ from here onward. Notice that
at the bottom of each branch there is an odd number. This is because we move off the
branch by a sideways arrow when we multiply by 3 and add 1 (with a single exception -
more on that later). We can use this feature to describe a certain branch of the tree, by
using m × 2n, where m is the odd number at the bottom of the branch and n is the
amount of integers above m on the tree. For example, the branch with 3 at the bottom
shall be called 3 × 2n. If we were to find the second number above this, we do:
 m=3
n=2
m × 2n = 3 × 22
= 12
By this we can find that the numbers in the 3 × 2n branch are 3, 6, 12, 24, 48, 96, 192,
and so on.

Noticeable patterns and smaller proofs

 There is only one odd number in each branch
 There is a loop from 4, 2, 1
 There is only one route for C(x) to follow
 Connectors will always be even
 Connectors are in a branch every 2 numbers
 There are no connectors in a column where m is a multiple of 3

Proof using more conjectures

Now, so far by my own calculations I have not found a fully notated proof, but I believe I
have some thoughts on how it could be proven. I have split this conjecture down into
other conjectures that can either be proven or are perhaps simpler to understand and
find proof for.
I believe that the Collatz Conjecture can be proven if:

1. All natural numbers appear once and only once on the Collatz number tree
2. Every integer takes a finite amount of steps to reach 1
3. The number tree will never stop at some integer
4. Only the 1×2n branch can link back to itself
5. No branch is able to link back to itself through other branches

Using these 5 conjectures may be the key to solving this problem. It would rely heavily
on the number tree representation, but I shall continue to work hard and attempt to
notate this in a more understandable way. For now, this is my reasoning for why these
may form proof:

1. If all natural numbers appear somewhere on the number tree, it means they are
all able to be connected by C(x), and when paired with the other conjectures, will
mean every integer will eventually get to one.
2. We must prove that no number continues to move through branches infinitely,
and therefore reaches a time where it moves into the 1×2n branch.
3. A sequence of C(x) must not stop at any one integer (unless we decide to
terminate it at 1), for if this is the case, there could be a sequence that stops at
an integer without passing 1.
4. This must be the only loop that can link into its own branch, making the C(x)
sequence repeat. If there is another loop, it might never reach 1.
5. Similar to conjecture 4, but this moves through multiple branches. If there is a
loop anywhere in the Collatz tree that does not include 1, then the conjecture may
be proven false. If not, then it’s true!

Overall, if every number is on the tree, and the tree never stops, and if there are no
loops apart from 1 in the number tree, then any given integer must continue to divide by
2 until it reaches an odd number, where it moves into a new branch. This process must
repeat a finite number of times so that it may finally enter the 1×2n branch, and
therefore the Collatz conjecture must be true.
Now, off to proving these 5 conjectures..