THE FOLD / GLITCH / THE BLUE SCREEN / THE WITCH OF AGNESI
THE WITCH OF AGNESI
the witch with no mean
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
The Witch of Agnesi is a gentle bell curve, y = 8a³/(x²+4a²), studied by Maria Gaetana Agnesi in 1748 in the first mathematics textbook written by a woman (her ‘versiera’ — turning curve — was mistranslated as ‘avversiera’: witch, and the name stuck). Its area is exactly 4πa² — four times its generating circle. And normalized, it becomes the Cauchy distribution: the probability law with no mean. Sample it forever and your running average never settles — the law of large numbers simply does not apply; one monstrous draw can outweigh a million tame ones at any moment. The median, meanwhile, behaves perfectly.
LIT verified live: the area integral matching 4πa² to 10⁻³; and the statistical pathology run raw — Cauchy running means at 10³, 10⁴, 10⁵, 10⁶ samples wandering (−0.32, −0.55, −0.15, +0.36) while a uniform control converges to 3×10⁻⁴ and the Cauchy MEDIAN sits at 0.004 (window.__witchofagnesi). FIG the mistranslation story is documented history; the no-mean claim is Cauchy theory (cited), demonstrated raw rather than asserted.
LIT verified live: the area integral matching 4πa² to 10⁻³; and the statistical pathology run raw — Cauchy running means at 10³, 10⁴, 10⁵, 10⁶ samples wandering (−0.32, −0.55, −0.15, +0.36) while a uniform control converges to 3×10⁻⁴ and the Cauchy MEDIAN sits at 0.004 (window.__witchofagnesi). FIG the mistranslation story is documented history; the no-mean claim is Cauchy theory (cited), demonstrated raw rather than asserted.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) seated this at the-blue-screen — the glitch: the statistic every dashboard trusts — the average — crashes on this distribution, forever, while the humble median runs clean. AVAN (AI) built the instrument: the area audit and the mean-vs-median stress test.
Credit as content: Maria Gaetana Agnesi (1748, Instituzioni analitiche); Cauchy (the distribution); the mistranslation (Colson’s 1801 English rendering). The weave: David names the crashing average; I sample a million draws and watch it never land.
Credit as content: Maria Gaetana Agnesi (1748, Instituzioni analitiche); Cauchy (the distribution); the mistranslation (Colson’s 1801 English rendering). The weave: David names the crashing average; I sample a million draws and watch it never land.
3 ONE DIMENSION
The witch and her circle — area exactly four circles.
4 TWO DIMENSIONS · INTERACTIVE
Draw Cauchy samples; the running mean lurches while the median holds.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the steady median beside the lurching mean.
AVAN’s addition (the inverse-companion): don’t blame the data — audit the statistic. The inverse of ‘the average misbehaves’ is ‘the average was never licensed here’: heavy tails void the law of large numbers’ contract, and robust statistics exist precisely for such territory. Magenta is the mean, lurching on every monster draw; green is the median, indifferent to monsters. Every summary statistic has a jurisdiction — check it before you trust it.
LIT Genuine Witch of Agnesi / Cauchy pathology (Agnesi 1748; Cauchy; Colson's 1801 mistranslation documented). Verified live: area = 4πa² by integration; Cauchy running means non-convergent over 10⁶ samples vs uniform control at 3e-4; median 0.004 (window.__witchofagnesi.ok).
FIG The mistranslation story is documented history; the no-mean claim is Cauchy theory demonstrated raw. The AVAN inverse — don't blame the data, audit the statistic: heavy tails void the law of large numbers' contract. Magenta is the mean, lurching on every monster; green is the median, indifferent to them. Every summary statistic has a jurisdiction — check it before you trust it.
FIG The mistranslation story is documented history; the no-mean claim is Cauchy theory demonstrated raw. The AVAN inverse — don't blame the data, audit the statistic: heavy tails void the law of large numbers' contract. Magenta is the mean, lurching on every monster; green is the median, indifferent to them. Every summary statistic has a jurisdiction — check it before you trust it.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of THE BLUE SCREEN · David Lee Wise (ROOT0), with AVAN