These Are the 10 Hardest Math Problems

That Remain Unsolved

The Collatz Conjecture
In September 2019, news broke regarding progress on this 82-year-old
question, thanks to prolific mathematician Terence Tao. And while
the story of Tao's breakthrough is promising, the problem isn't fully
solved yet.

A refresher on the Collatz Conjecture: It's all about that function f(n),

shown above, which takes even numbers and cuts them in half, while odd
numbers get tripled and then added to 1. Take any natural number, apply
f, then apply f again and again. You eventually land on 1, for every
number we've ever checked. The Conjecture is that this is true for all
natural numbers (positive integers from 1 through infinity).

Tao's recent work is a near-solution to the Collatz Conjecture in some

subtle ways. But he most likely can't adapt his methods to yield a
complete solution to the problem, as Tao subsequently explained. So, we
might be working on it for decades longer.

The Conjecture lives in the math discipline known as Dynamical

Systems, or the study of situations that change over time in semi-
predictable ways. It looks like a simple, innocuous question, but that's
what makes it special. Why is such a basic question so hard to answer? It
serves as a benchmark for our understanding; once we solve it, then we
can proceed onto much more complicated matters.

The study of dynamical systems could become more robust than anyone
today could imagine. But we'll need to solve the Collatz Conjecture for
the subject to flourish.

Goldbach’s Conjecture

CREATIVE COMMONS
One of the greatest unsolved mysteries in math is also very easy to
write. Goldbach's Conjecture is, "Every even number (greater than two)
is the sum of two primes." You check this in your head for small
numbers: 18 is 13+5, and 42 is 23+19. Computers have checked the
Conjecture for numbers up to some magnitude. But we need proof
for all natural numbers.

Goldbach's Conjecture precipitated from letters in 1742 between German

mathematician Christian Goldbach and legendary Swiss
mathematician Leonhard Euler, considered one of the greatest in math
history. As Euler put it, "I regard [it] as a completely certain theorem,
although I cannot prove it."

Euler may have sensed what makes this problem counterintuitively hard
to solve. When you look at larger numbers, they have more ways of being
written as sums of primes, not less. Like how 3+5 is the only way to
break 8 into two primes, but 42 can broken into 5+37, 11+31, 13+29, and
19+23. So it feels like Goldbach's Conjecture is an understatement for
very large numbers.

Still, a proof of the conjecture for all numbers eludes mathematicians to

this day. It stands as one of the oldest open questions in all of math.

ADVERTISEMENT - CONTINUE READING BELOW

3The Twin Prime Conjecture

WOLFRAM ALPHA

Together with Goldbach's, the Twin Prime Conjecture is the most famous
in Number Theory—or the study of natural numbers and their
properties, frequently involving prime numbers. Since you've known
these numbers since grade school, stating the conjectures is easy.

When two primes have a difference of 2, they're called twin primes. So 11

and 13 are twin primes, as are 599 and 601. Now, it's a Day 1 Number
Theory fact that there are infinitely many prime numbers. So, are there
infinitely many twin primes? The Twin Prime Conjecture says yes.
Let's go a bit deeper. The first in a pair of twin primes is, with one
exception, always 1 less than a multiple of 6. And so the second twin
prime is always 1 more than a multiple of 6. You can understand why, if
you're ready to follow a bit of heady Number Theory.

All primes after 2 are odd. Even numbers are always 0, 2, or 4 more than
a multiple of 6, while odd numbers are always 1, 3, or 5 more than a
multiple of 6. Well, one of those three possibilities for odd numbers
causes an issue. If a number is 3 more than a multiple of 6, then it has
a factor of 3. Having a factor of 3 means a number isn't prime (with the
sole exception of 3 itself). And that's why every third odd number can't
be prime.

How's your head after that paragraph? Now imagine the headaches of
everyone who has tried to solve this problem in the last 170 years.

The good news is that we've made some promising progress in the last
decade. Mathematicians have managed to tackle closer and closer
versions of the Twin Prime Conjecture. This was their idea: Trouble
proving there are infinitely many primes with a difference of 2? How
about proving there are infinitely many primes with a difference of
70,000,000? That was cleverly proven in 2013 by Yitang Zhang at the
University of New Hampshire.

For the last six years, mathematicians have been improving that number
in Zhang's proof, from millions down to hundreds. Taking it down all the
way to 2 will be the solution to the Twin Prime Conjecture. The closest
we've come—given some subtle technical assumptions—is 6. Time will
tell if the last step from 6 to 2 is right around the corner, or if that last
part will challenge mathematicians for decades longer.

ADVERTISEMENT - CONTINUE READING BELOW

4The Riemann Hypothesis

DAVE LINKLETTER

Today's mathematicians would probably agree that the Riemann

Hypothesis is the most significant open problem in all of math. It's one of
the seven Millennium Prize Problems, with $1 million reward for its
solution. It has implications deep into various branches of math, but it's
also simple enough that we can explain the basic idea right here.

There is a function, called the Riemann zeta function, written in the

image above.

For each s, this function gives an infinite sum, which takes some basic
calculus to approach for even the simplest values of s. For example, if
s=2, then 𝜁(s) is the well-known series 1 + 1/4 + 1/9 + 1/16 + …, which
strangely adds up to exactly 𝜋²/6. When s is a complex number—one
that looks like a+b𝑖, using the imaginary number 𝑖—finding 𝜁(s) gets
tricky.

So tricky, in fact, that it's become the ultimate math question.

Specifically, the Riemann Hypothesis is about when 𝜁(s)=0; the official
statement is, "Every nontrivial zero of the Riemann zeta function has real
part 1/2." On the plane of complex numbers, this means the function has
a certain behavior along a special vertical line. The hypothesis is that the
behavior continues along that line infinitely.

The Hypothesis and the zeta function come from German mathematician
Bernhard Riemann, who described them in 1859. Riemann developed
them while studying prime numbers and their distribution. Our
understanding of prime numbers has flourished in the 160 years since,
and Riemann would never have imagined the power of supercomputers.
But lacking a solution to the Riemann Hypothesis is a major setback.

If the Riemann Hypothesis were solved tomorrow, it would unlock an

avalanche of further progress. It would be huge news throughout the
subjects of Number Theory and Analysis. Until then, the Riemann
Hypothesis remains one of the largest dams to the river of math
research.

5The Birch and Swinnerton-Dyer Conjecture

CREATIVE COMMONS

The Birch and Swinnerton-Dyer Conjecture is another of the six

unsolved Millennium Prize Problems, and it's the only other one we can
remotely describe in plain English. This Conjecture involves the math
topic known as Elliptic Curves.

When we recently wrote about the toughest math problems that have

been solved, we mentioned one of the greatest achievements in 20th-
century math: the solution to Fermat's Last Theorem. Sir Andrew Wiles
solved it using Elliptic Curves. So, you could call this a very powerful new
branch of math.

In a nutshell, an elliptic curve is a special kind of function. They take the

unthreatening-looking form y²=x³+ax+b. It turns out functions like this
have certain properties that cast insight into math topics like Algebra
and Number Theory.

British mathematicians Bryan Birch and Peter Swinnerton-Dyer

developed their conjecture in the 1960s. Its exact statement is very
technical, and has evolved over the years. One of the main stewards of
this evolution has been none other than Wiles. To see its current status
and complexity, check out this famous update by Wells in 2006.

ADVERTISEMENT - CONTINUE READING BELOW

6The Kissing Number Problem
JJ HARRISON/CREATIVE COMMONS

A broad category of problems in math are called the Sphere

Packing Problems. They range from pure math to practical applications,
generally putting math terminology to the idea of stacking many spheres
in a given space, like fruit at the grocery store. Some questions in this
study have full solutions, while some simple ones leave us stumped, like
the Kissing Number Problem.

When a bunch of spheres are packed in some region, each sphere has a
Kissing Number, which is the number of other spheres it’s touching; if
you’re touching 6 neighboring spheres, then your kissing number is 6.
Nothing tricky. A packed bunch of spheres will have an average kissing
number, which helps mathematically describe the situation. But a basic
question about the kissing number stands unanswered.

First, a note on dimensions. Dimensions have a specific meaning in

math: they’re independent coordinate axes. The x-axis and y-axis show
the two dimensions of a coordinate plane. When a character in a sci-fi
show says they’re going to a different dimension, that doesn’t make
mathematical sense. You can’t go to the x-axis.

A 1-dimensional thing is a line, and 2-dimensional thing is a plane. For

these low numbers, mathematicians have proven the maximum possible
kissing number for spheres of that many dimensions. It’s 2 when you’re
on a 1-D line—one sphere to your left and the other to your right. There’s
proof of an exact number for 3 dimensions, although that took until the
1950s.

Beyond 3 dimensions, the Kissing Problem is mostly unsolved.

Mathematicians have slowly whittled the possibilities to fairly narrow
ranges for up to 24 dimensions, with a few exactly known, as you can
see on this chart. For larger numbers, or a general form, the problem is
wide open. There are several hurdles to a full solution, including
computational limitations. So expect incremental progress on this
problem for years to come.

ADVERTISEMENT - CONTINUE READING BELOW

7The Unknotting Problem

CREATIVE COMMONS
The simplest version of the Unknotting Problem has been solved, so
there's already some success with this story. Solving the full version of
the problem will be an even bigger triumph.

You probably haven't heard of the math subject Knot Theory. It's taught
in virtually no high schools, and few colleges. The idea is to try and apply
formal math ideas, like proofs, to knots, like … well, what you tie your
shoes with.

For example, you might know how to tie a "square knot" and a "granny
knot." They have the same steps except that one twist is reversed from
the square knot to the granny knot. But can you prove that those knots
are different? Well, knot theorists can.

Knot theorists" holy grail problem was an algorithm to identify if some

tangled mess is truly knotted, or if it can be disentangled to nothing. The
cool news is that this has been accomplished! Several computer
algorithms for this have been written in the last 20 years, and some of
them even animate the process.

But the Unknotting Problem remains computational. In technical terms,

it's known that the Unknotting Problem is in NP, while we don't know if
it's in P. That roughly means that we know our algorithms are capable of
unknotting knots of any complexity, but that as they get more
complicated, it starts to take an impossibly long time. For now.

If someone comes up with an algorithm that can unknot any knot in

what's called polynomial time, that will put the Unknotting Problem fully
to rest. On the flip side, someone could prove that isn't possible, and that
the Unknotting Problem's computational intensity is unavoidably
profound. Eventually, we'll find out.

ADVERTISEMENT - CONTINUE READING BELOW

8The Large Cardinal Project
CREATIVE COMMONS

If you've never heard of Large Cardinals, get ready to learn. In the late
19th century, a German mathematician named Georg Cantor figured out
that infinity comes in different sizes. Some infinite sets truly have more
elements than others in a deep mathematical way, and Cantor proved it.

There is the first infinite size, the smallest infinity, which gets denoted
ℵ₀. That's a Hebrew letter aleph; it reads as "aleph-zero." It's the size of
the set of natural numbers, so that gets written |ℕ|=ℵ₀.
Next, some common sets are larger than size ℵ₀. The major example
Cantor proved is that the set of real numbers is bigger, written |ℝ|>ℵ₀.
But the reals aren't that big; we're just getting started on the infinite
sizes.

For the really big stuff, mathematicians keep discovering larger and
larger sizes, or what we call Large Cardinals. It's a process of pure math
that goes like this: Someone says, "I thought of a definition for a
cardinal, and I can prove this cardinal is bigger than all the known
cardinals." Then, if their proof is good, that's the new largest known
cardinal. Until someone else comes up with a larger one.

Throughout the 20th century, the frontier of known large cardinals was
steadily pushed forward. There's now even a beautiful wiki of known
large cardinals, named in honor of Cantor. So, will this ever end? The
answer is broadly yes, although it gets very complicated.

In some senses, the top of the large cardinal hierarchy is in sight. Some
theorems have been proven, which impose a sort of ceiling on the
possibilities for large cardinals. But many open questions remain, and
new cardinals have been nailed down as recently as 2019. It's very
possible we will be discovering more for decades to come. Hopefully we'll
eventually have a comprehensive list of all large cardinals.

ADVERTISEMENT - CONTINUE READING BELOW

9What’s the Deal with 𝜋+e?
ANDREW DANIELS

Given everything we know about two of math's most famous

constants, 𝜋 and e, it's a bit surprising how lost we are when they're
added together.

This mystery is all about algebraic real numbers. The definition: A real

number is algebraic if it's the root of some polynomial with integer
coefficients. For example, x²-6 is a polynomial with integer coefficients,
since 1 and -6 are integers. The roots of x²-6=0 are x=√6 and x=-√6, so
that means √6 and -√6 are algebraic numbers.

All rational numbers, and roots of rational numbers, are algebraic. So it

might feel like "most" real numbers are algebraic. Turns out, it's actually
the opposite. The antonym to algebraic is transcendental, and it turns
out almost all real numbers are transcendental—for certain
mathematical meanings of "almost all." So who's algebraic, and who's
transcendental?
The real number 𝜋 goes back to ancient math, while the number e has
been around since the 17th century. You've probably heard of both, and
you'd think we know the answer to every basic question to be asked
about them, right?

Well, we do know that both 𝜋 and e are transcendental. But somehow it's
unknown whether 𝜋+e is algebraic or transcendental. Similarly, we don't
know about 𝜋e, 𝜋/e, and other simple combinations of them. So there are
incredibly basic questions about numbers we've known for millennia that
still remain mysterious.

ADVERTISEMENT - CONTINUE READING BELOW

10Is 𝛾 Rational?

DAVE LINKLETTER

Here's another problem that's very easy to write, but hard to solve. All
you need to recall is the definition of rational numbers.

Rational numbers can be written in the form p/q, where p and q are
integers. So, 42 and -11/3 are rational, while 𝜋 and √2 are not. It's a very
basic property, so you'd think we can easily tell when a number is
rational or not, right?

Meet the Euler-Mascheroni constant 𝛾, which is a lowercase Greek

gamma. It's a real number, approximately 0.5772, with a closed form
that's not terribly ugly; it looks like the image above.

The sleek way of putting words to those symbols is "gamma is the limit of
the difference of the harmonic series and the natural log." So, it's a
combination of two very well-understood mathematical objects. It has
other neat closed forms, and appears in hundreds of formulas.
But somehow, we don't even know if 𝛾 is rational. We've calculated it to
half a trillion digits, yet nobody can prove if it's rational or not. The
popular prediction is that 𝛾 is irrational. Along with our previous
example 𝜋+e, we have another question of a simple property for a well-
known number, and we can't even answer it.