The Collatz Conjecture: A Tiny Rule With No Known Ending

01A rule a child can execute

Take a positive integer n. If it is even, divide by two. If it is odd, replace it with 3n + 1. Repeat. The Collatz conjecture claims that every positive starting value eventually reaches 1, after which the cycle 1 → 4 → 2 → 1 repeats.

Lothar Collatz introduced the problem in 1937. Its nickname, the hailstone problem, is apt: trajectories rise and fall unpredictably, sometimes soaring far above their starting point before coming back down.

02Why computation is persuasive—and insufficient

Computers have checked enormous finite ranges, and every tested starting value has eventually reached the 1–4–2 cycle. That is strong evidence about the landscape, not a proof about an infinite domain. A counterexample could be larger than every checked bound, and a proof must rule out both divergent orbits and non-trivial cycles everywhere.

The charted stopping times make the trap visible. The starting value 7 takes 16 steps; 837,799 takes 524. The rule has only two branches, but parity choices compound into a branching tree whose long tails are difficult to control.

03The 27 sequence and the illusion of monotonicity

The orbit beginning at 27 is a classic warning. It takes 111 steps to reach 1 and peaks at 9,232. A naive “it usually shrinks” intuition fails locally: odd steps multiply by three and add one, while the subsequent divisions may or may not compensate quickly.

On average, a two-step odd/even pattern tends to contract in logarithmic terms, which explains why random-looking experiments descend. But average behavior does not exclude a rare path that escapes forever. The mathematical challenge is to turn a statistical tendency into a universal guarantee.

04The results that narrowed the gap

Riho Terras proved that almost every positive integer has a finite stopping time—in other words, almost every orbit eventually falls below its starting value. In 2019, Terence Tao pushed further using logarithmic density: almost all orbits descend below any function of the starting point that grows to infinity, however slowly.

These results explain why numerical experiments look so compelling. They describe the overwhelming statistical bulk of integers. They do not identify a single finite threshold after which every orbit must fall, and they leave open the possibility of an exceptional set too thin for density arguments to eliminate.

05The hidden computational difficulty

Collatz is a dynamical system on the integers, but it is also a problem about information. Each odd step injects multiplication; each even step removes factors of two. The parity sequence acts like a compressed itinerary, and proving that every possible itinerary eventually contracts would require controlling correlations that simple probability models wash away.

Conway proved that a broad class of generalized Collatz-like systems is algorithmically undecidable. That result does not prove the original conjecture undecidable, but it is a warning: tiny piecewise-linear rules can encode surprisingly deep computation.

06Cycles, divergence, and what a proof must close

A complete proof needs to show that the only positive cycle is 1 → 4 → 2 → 1 and that no orbit diverges to infinity. Proving that every number eventually dips below its starting value would be enough, because a minimal counterexample could not exist under that descent property.

Conversely, a disproof would need a concrete divergent trajectory or a different cycle. The absence of such an example after extensive computation is meaningful evidence, but it is not the same thing as a structural theorem about all integers.

07Why the simplest problems endure

The Collatz conjecture survives because it sits at a productive boundary: elementary to state, computationally accessible, statistically suggestive, and resistant to the standard tools of number theory. It teaches a general lesson in unsolved problems: “easy to verify many cases” and “easy to prove universally” are different species of knowledge.

The open question is not whether the rule looks chaotic. It is whether that apparent chaos has a hidden invariant strong enough to force every hailstone back to the ground.