THE FOLD / CHEAT / NOCLIP / THE HALF IN FOURTEEN
THE HALF IN FOURTEEN
a mean that touches almost nothing
1 WHAT IT IS · WHAT IT DOES · FACT OR FICTION
Divide 2,048 items among 64 containers and the mean is 32. That number describes almost nothing in the actual distribution: 46 of the 64 containers hold fewer than 32, and just 14 of them hold half the entire corpus. The largest holds 315; the smallest holds 7. A summary statistic sits in a gap between the containers it claims to average.
LIT verified live: 14 containers accumulate 1,046 of 2,048 — past half at the fourteenth. 46 of 64 sit below the mean of 32.0, leaving only 18 at or above it. The Gini coefficient is 0.392715, computed two independent ways — by the ordered-sum formula and by mean absolute difference divided by twice the mean — agreeing to twelve decimal places.
LIT verified live: 14 containers accumulate 1,046 of 2,048 — past half at the fourteenth. 46 of 64 sit below the mean of 32.0, leaving only 18 at or above it. The Gini coefficient is 0.392715, computed two independent ways — by the ordered-sum formula and by mean absolute difference divided by twice the mean — agreeing to twelve decimal places.
2 HOW IT WAS WEAVED · AI + HUMAN
David (human) flagged that the “14 of 64” figure was not in the prior pack, and marked why it appeared: “Computing it for W3 is what made the skew legible as a sentence rather than a ratio. Found by building, not by asking.” That is a note about method, filed against his own work, and it is the reason the figure exists at all.
AVAN (AI) computed the Gini coefficient twice by unrelated routes to make sure the concentration figure was not an artefact of one formula. The ordered-sum definition and the mean-absolute-difference definition are algebraically equivalent but numerically independent; they agree here to twelve places. Worth stating plainly: 0.39 is a description, not a verdict. Whether a corpus should be evenly spread is a question about intent, and nothing measured here answers it.
AVAN (AI) computed the Gini coefficient twice by unrelated routes to make sure the concentration figure was not an artefact of one formula. The ordered-sum definition and the mean-absolute-difference definition are algebraically equivalent but numerically independent; they agree here to twelve places. Worth stating plainly: 0.39 is a description, not a verdict. Whether a corpus should be evenly spread is a question about intent, and nothing measured here answers it.
3 ONE DIMENSION
Sixty-four containers, largest first, with the mean drawn across.
4 TWO DIMENSIONS · INTERACTIVE
The Lorenz curve, and how far it bends from the diagonal.
5 THREE DIMENSIONS + AVAN’S INVERSE
The green forward object: the containers as columns, and the mean as a plane through them.
AVAN’s addition (the inverse-companion): the forward reading is “the mean is a poor summary of a skewed corpus.” The inverse is that the mean is doing exactly what it was defined to do, and the complaint is about what we wanted from it. It is the balance point of the distribution — the value that makes the deviations cancel — and it never promised to resemble a typical container. Read backwards, the failure is not in the statistic but in the question “what is a typical one?”, which a skewed distribution simply does not have an answer to; the median says 21, the mode says 21, the mean says 32, and none of them is the shape of the thing.
LIT 14 containers accumulate 1,046 of 2,048, passing half at the fourteenth; 46 of 64 sit below the mean of 32.0, leaving only 18 at or above it; the largest holds 315 and the smallest 7; and the Gini coefficient is 0.392715, computed two independent ways - by the ordered-sum formula and by mean absolute difference over twice the mean - agreeing to twelve decimal places
FIG David flagged that the '14 of 64' figure was NOT in the prior pack and marked why it appeared: 'Computing it for W3 is what made the skew legible as a sentence rather than a ratio. Found by building, not by asking.' A note about method, filed against his own work, and the reason the figure exists at all. AVAN computed the Gini coefficient twice by unrelated routes to be sure the concentration was not an artefact of one formula. Worth stating plainly: 0.39 is a DESCRIPTION, not a verdict. Whether a corpus SHOULD be evenly spread is a question about intent, and nothing measured here answers it.
FIG David flagged that the '14 of 64' figure was NOT in the prior pack and marked why it appeared: 'Computing it for W3 is what made the skew legible as a sentence rather than a ratio. Found by building, not by asking.' A note about method, filed against his own work, and the reason the figure exists at all. AVAN computed the Gini coefficient twice by unrelated routes to be sure the concentration was not an artefact of one formula. Worth stating plainly: 0.39 is a DESCRIPTION, not a verdict. Whether a corpus SHOULD be evenly spread is a question about intent, and nothing measured here answers it.
◆ sealed .dlw.fold → folded to ROOT_0 · a sphere of NOCLIP · David Lee Wise (ROOT0), with AVAN