Primes Have a Favorite Number. Mathematicians Can't Explain Why.

Take every odd prime and sort it by its remainder when divided by four. Every odd prime is either 1 mod 4 or 3 mod 4 — two camps. Dirichlet proved in 1837 that asymptotically the two camps grow equally. The Prime Number Theorem for Arithmetic Progressions confirms it: π(x; 4, 1) ~ π(x; 4, 3) as x → ∞. So if you count primes in both classes as you go, the running counts should bob up and down, the lead alternating roughly evenly forever. That's the prediction. The reality is not that. In 1853, Pafnuty Chebyshev wrote a letter to Hermann Fuss noting something odd: the 3-mod-4 primes persistently outnumber the 1-mod-4 primes. Every checkpoint, every count, every modulus — 3-mod-4 leads. Chebyshev conjectured the lead is absolute: π(x; 4, 3) ≥ π(x; 4, 1) for all x. For 104 years, the conjecture held. Then in 1957, John Leech finally pushed the search past 26,000 and found it: at x = 26,861, the 1-mod-4 count edges ahead by exactly one prime. After a century, Chebyshev was wrong — but only barely. The 3-mod-4 lead returned at the very next prime. In 1962, Knapowski and Turán proved infinitely many such crossovers exist in both directions. In 1994, Michael Rubinstein and Peter Sarnak (Princeton, in Experimental Mathematics) computed the precise logarithmic density: assuming the Generalized Riemann Hypothesis plus linear independence of the zeros, primes ≡ 3 mod 4 lead primes ≡ 1 mod 4 exactly 99.59% of the time. Out of every ten thousand checks, the 3-mod-4 side is ahead on 9,959 of them. The same phenomenon repeats for mod 3, mod 5, mod 10, mod every modulus. There's always a winning side. The percentages differ, but the bias is universal. The video unpacks the entire story — and saves the explanation until the final 4 minutes. The reveal: it has nothing to do with the primes themselves. It has to do with prime POWERS. Every odd prime, when squared, lands in residue 1 mod 4. The residue class 1 mod 4 "collects" all the prime squares; the residue class 3 mod 4 doesn't. To keep the natural weighted count balanced, the actual primes ≡ 1 mod 4 have to come up short by ~√x / log x. That's the bias, forced by the simplest algebraic fact: the square of any number is always a quadratic residue. Chapters 00:00 The race 01:45 Something's off 03:30 The bias holds 05:53 Chebyshev notices (1853) 07:38 Does one-mod-four ever win? 09:22 1957 — the crack 11:16 Quantified: 99.59% 13:07 Why? Prime squares. References Chebyshev (1853), "Lettre de M. le Professeur Tchébychev à M. Fuss," Bull. Classe Phys. Acad. Imp. Sci. St. Petersburg Littlewood (1914), "Sur la distribution des nombres premiers" Leech (1957), "Note on the distribution of prime numbers" Knapowski & Turán (1962), "Comparative prime-number theory" Rubinstein & Sarnak (1994), "Chebyshev's bias," Experimental Mathematics Granville & Martin (2006), "Prime number races," American Mathematical Monthly #math #primes #numbertheory #chebyshev #primenumberrace #riemannhypothesis #manim