The Nazca Construction Bench: Visualisation · Signals from the Periphery

← Intro 中文

View

THE PLOT

Subject

—

Number

—

State

—

THE NAZCA CONSTRUCTION BENCH
PAMPA DE NAZCA · 500 BC – AD 500
a centre-line and a knotted cord scale a sketch to any size, error-free-ish
a 1-arcminute eye holds a kilometre straight to metres
the whole pampa's clearing bill: ~257 person-days a year
why they drew it stays open, here included
drive the stakes · climb the tower · read the file

…

…

…

Evidence

fact ⇄ what kind of claim · live

Dark varnished stones over pale lime-crusted silt; ~4 mm of rain a year

lift the dark layer and the ground is the ink; the crust and the wind keep it

REPORTED

Figures 24–285 m, each one unbroken line; straight lines run kilometres

hummingbird 93 m, condor 134 m; a wooden stake C-14 dates one line AD 525 ± 80

REPORTED

Centre-line coordinates scale a sketch with no size penalty: RMS = σ

independent errors refuse to add; the chain method's σ√N gap is still only metres

EXACT

A 1-arcminute eye holds a kilometre straight to under 2 m

sighting arithmetic + Monte Carlo; post remains at mile intervals are the record's poles

EXACT

The labour bill: a figure is days; the whole pampa ~257 person-days a year

area ÷ the 1984 Earthwatch measured rate, 2.3 m² per person-hour

EXACT

1982: six people staked a full-size condor in nine hours, from the ground

Nickell, Skeptical Inquirer 1983; the first aerial view was the after-check

REPORTED

Hawkins tested 186 lines against the sky: 39 hit, chance predicts 37

the astronomical reading failed; Aveni's lines point at water instead

REPORTED

Why they drew it

the one question this bench refuses to close for you

READING

View

…

The plot

Survey error σ5 cm

One cord measurement's error, on a log dial from 1 to 50 cm. The named stops snap. Deviations draw magnified (factor on the canvas); the readout stays honest.

Stakes120

Points plotted. Nickell's crew used 165 on a full-size condor and reckoned the Nazca used fewer.

The ground

Stakes & cordthe construction vignette, along the outline

A maker1.6 m of scale, standing on the line

Labels

The labour

Crew size40

Reprices every job on the labour card, at the measured 1984 clearing rate.

The sky test

Tolerance±1°

How close counts as a hit.

Sky targets18

Solstice, equinox, lunar extremes, bright stars: the more you admit, the more chance pays for free.

Readout · live

Outline length

—

Stakes

—

Figure RMS (true)

—

Closure gap

—

Eye · grazing

—

Foreshortening

—

Hummingbird bill

—

Chance hits

—

…

What is exact vs modelled here

Exact: the error-propagation closed forms (RMS = σ for centre-line coordinates, σ√k for the chain, σ√N closure), the sighting arithmetic at a 1-arcminute eye, the foreshortening geometry sin(grazing angle), the labour division, and the chance-hit arithmetic, all checked in scripts/nazca-tune.mjs against Monte Carlo. Reported: the figure dimensions, the pavement's two tones, the 1984 clearing rate, Nickell's 1982 reproduction, Hawkins' 39-of-186, Aveni's ray centres, the balloon flight, the C-14 stake. Modelled: the hummingbird redraw, the 3D pampa and its foothill height, the stake spacing, the 0.5 m working width. There is no fitted constant anywhere on this bench.

The model, in full

Every stake carries one unit-Gaussian noise pair, drawn once from a seeded stream and multiplied by the σ dial live, so the picture is an honest Monte Carlo and the dial never re-rolls the crew. The coordinate method applies noise per point; the chain method applies the running sum of the same stream. The straightaway integrates the leapfrog recurrence stake by stake against the closed form σ_t·N^1.5/√3. The ground view reads eye height and grazing angle off the true camera and computes 1/sin α; the labour and sky cards are pure arithmetic on data.js. Anchors: scripts/nazca-tune.mjs.

Where this instrument takes liberties

The hummingbird outline is a stylised single-stroke redraw, not a survey tracing; the pampa's terrain, colours and foothill height are staging (the record gives no clean relief figure, so ~90 m is labelled as a variable); the working width of 0.5 m and the six-hour work day are staged bookkeeping choices; the real lines' "few yards every mile" straightness is a 1974 journalistic figure, carried with that label; and the monkey's length and the largest figure's size are disputed across sources, so the bench quotes ranges on the intro page and uses the best-attributed single values here.

…1 / 15

…

…

…