The Collatz Conjecture: Why 3n+1 Always Seems to End at 1

There is a math problem you can explain to a ten-year-old in one sentence, and yet no professional mathematician on Earth can prove it. Pick any positive whole number. If it is even, cut it in half. If it is odd, multiply it by three and add one. Repeat. The claim is that you will always, eventually, land on 1. That is the whole thing. It has been checked for every number up past 2^68, it has never failed once, and after almost ninety years nobody has shown it must hold forever.

This is the Collatz conjecture, also called the 3n+1 problem. I want to walk through how it actually behaves, why the numbers bounce around so wildly, and why such a tidy little rule has resisted every attack.

The Rule, Stated Once

The entire engine is two cases:

If n is even, replace it with n / 2.
If n is odd, replace it with 3n + 1.

That is it. An even step shrinks the number. An odd step roughly triples it (and the result is always even, so the very next step halves it back down). You alternate between these moves depending on whether the current value is even or odd, and you keep going until you hit 1. Once you reach 1, the rule sends you 1 → 4 → 2 → 1 forever, so 1 is treated as the finish line.

Lothar Collatz wrote this down in 1937. Since then it has picked up a small museum of names — the 3n+1 problem, the hailstone problem, the Ulam conjecture, Kakutani's problem, the Syracuse problem — but every one of them describes this same pair of moves and the same open question.

A Worked Example: Following 6 to the End

Numbers are more convincing than rules, so let me trace one by hand. Start at 6.

6 is even → halve it → 3
3 is odd → 3×3 + 1 = 10 → 10
10 is even → halve it → 5
5 is odd → 3×5 + 1 = 16 → 16
16 is even → 8
8 is even → 4
4 is even → 2
2 is even → 1

The full path is 6, 3, 10, 5, 16, 8, 4, 2, 1. Count the arrows: that is 8 steps to reach 1. The largest value the path ever touched is 16, which we call the peak. Notice the shape — the number climbs (6 down to 3, then up to 10, down to 5, up to 16) before the long slide of halvings carries it home. That up-and-down wobble is exactly why these are nicknamed hailstone sequences: the values rise and fall like a hailstone tossed around inside a storm cloud before it finally drops to the ground.

You can reproduce this run in seconds with the Collatz Conjecture calculator. Type 6, and it lays out the same nine entries, reports the stopping time of 8, and flags the peak of 16. Then try changing the seed and watch the picture change completely.

Step Counts and Peaks Have No Pattern

Here is where the conjecture stops being cute and starts being strange. You might guess that small numbers settle quickly and big numbers take longer. They do not. The behaviour is jumpy in a way that feels almost random.

Take two neighbours. The number 26 reaches 1 in just 10 steps. Its next-door neighbour 27 takes 111 steps and, along the way, climbs all the way up to 9232 before crashing back down. Same neighbourhood, wildly different orbits. A seed under 100 with a peak in the thousands is not an exception; it is just how the problem works.

If you collect a few record-holders under 100 — 27, 31, 47, 54, 73, 97 — and compare their stopping times, you get a ragged little table with no clean trend. Each odd step inflates the value, each even step deflates it, and the order in which those steps arrive depends on bit patterns that shift unpredictably as the number grows. There is no formula that reads off the stopping time from the starting number. You have to run it.

One practical note while you experiment: peaks get large fast, and ordinary floating-point arithmetic starts losing exactness above roughly nine quadrillion (2^53). A good calculator uses exact big-integer math so a peak of 9232 reads as 9232, not as some rounded approximation. If you want to feed those sequences into other math work, you can copy them out clean and drop them next to a tool like the scientific calculator for whatever comes next.

Why It Refuses to Be Proven

So why is something this simple unsolved? The trouble is the mixture. Multiplication and division pull in opposite directions, and the rule keeps switching between them based on parity. An odd step multiplies by 3 (then the forced even step halves), so a rough "odd round trip" multiplies the number by about 3/2 — it grows. An even step multiplies by 1/2 — it shrinks. Whether a given starting number trends up or down depends on how many odd steps versus even steps it encounters, and that ratio is exactly the thing nobody can control in general.

Two things would each prove the conjecture false, and we have never found either: a starting number whose sequence shoots off to infinity, or a starting number that falls into a loop other than 4 → 2 → 1. Computers have ruled both out for an enormous range of inputs, but "no counterexample up to 2^68" is verification, not proof. Verification checks cases; a proof must cover all infinitely many numbers at once.

Paul Erdős, one of the most prolific mathematicians of the twentieth century, looked at it and said mathematics is not yet ready for such problems, then offered money for a solution anyway. That tells you something about its standing: it is not that nobody cares, it is that the standard toolkit of number theory does not have a handle on it.

What I Noticed Running Hundreds of Seeds

I sat down one evening and ran every starting number from 1 to 200 just to get a feel for it, and the thing that stuck with me was how local the chaos is. I expected smooth bands — short orbits for small seeds, gradually longer ones higher up. Instead I got spikes that came out of nowhere. I would watch a run of seeds finish in 15 or 20 steps, completely calm, and then 27 would detonate into 111 steps with a peak of 9232 sitting in the middle of otherwise quiet company. After enough of those, I stopped trying to predict anything. You genuinely cannot tell by looking at a number what its hailstone flight will do, and that is the whole charm. It also gave me real respect for how much computing has been thrown at this: every one of my little runs is one more verified case, and people have done that billions upon billions of times.

Where to Take It Next

The Collatz rule is a gateway into how integer sequences behave, and it pairs naturally with other generators. If you enjoy watching numbers march along a rule, the number sequence generator lets you build arithmetic, geometric, and custom progressions and compare how predictable they are next to the hailstone chaos.

The honest summary is this: the 3n+1 problem is a reminder that simplicity and difficulty are not opposites. Halve the evens, triple-plus-one the odds, and every number we have ever checked tumbles down to 1. Whether that is a theorem or merely a very lucky streak across every number ever tested is still, after all these decades, an open question. Run a few seeds yourself, watch the peaks jump, and you will understand why it keeps people up at night.

Made by Toolora · Updated 2026-06-13