Do the Nazca lines point at the Sun more often than chance?

The question

Since the 1940s the Nazca lines of southern Peru have been called a giant sun calendar, with lines aimed at sunrise and sunset on key days. In 2021 a German and Czech team (Richter, Teichert and Pavelka, Applied Sciences 11(4):1637, from HTW Dresden) checked all 2,308 straight lines in their survey database against the Sun's rising and setting directions at the solstices, the equinoxes and the zenith passage, when the Sun is directly overhead. They found 99 matches and concluded that the calendar theory was not supported. Earlier, smaller studies by Hawkins and Aveni had reached a similar verdict.

A count of 99 means little without a comparison. As far as this study could find, the paper does not report how many of 2,308 lines would point those ways by accident. So: is 99 more than chance, about the same, or fewer?

The test

Paul asked an AI model (Claude) to work this out on public data. The results are dated 30 September 2026. The individual line directions are not published, so the test used the paper's own chart of directions (Figure 13, open access under CC BY 4.0), whose 18 bars in 10 degree slices, read off the figure, add up to exactly 2,308, and its counts of 62 solstice, 24 equinox and 13 zenith hits.

For a flat horizon at latitude 14.7 degrees south, the code computed the target directions: solstice 65.45 and 114.55 degrees (matching the edges of the paper's own windows, 65.6 and 114.4), equinox 90, and zenith passage 105.2 and 74.8. It then generated 20,000 random sets of 2,308 lines, each with the same direction pattern as the real ones, and counted hits.

Unlike the other two Book 6 studies, no plan or pass mark was written beforehand.

What came out

The answer turns on one detail the study could not settle: whether the survey's stated tolerance of 2 degrees means 2 degrees either side of a target (a 4 degree window) or a 2 degree window in total (1 degree either side).

Bracketed figures read the chart's bar labels as the lower edge of each slice, not the upper.

If the rule was 2 degrees either side, chance gives about 234 hits and the real lines have 99. No random set was that low; the lowest had 177. Read literally, the lines point at these directions less than half as often as chance. If the rule was 1 degree either side, chance gives 117 to 121, and 99 is a little below it, but not strongly so: 4.5% of random sets (1.9% on the other reading) were as low as 99.

Either way, the calendar theory gets no support, which agrees with the survey authors. The two cases differ in what else they say. The first would be a strange finding, and a difference in how hits were counted is a likelier explanation than builders steering away from the Sun. That is a judgement, not a result.

What this does not show

It does not show that the builders avoided the Sun. That reading should wait for the authors' counting rule.

The open points:

• The chart has 10 degree slices, so directions inside each slice were assumed evenly spread.
• The survey adjusted each target for the height of the real horizon. The test used a flat one, which shifts the windows slightly but should not change the expected rate much.
• It is unknown whether "zenith passage" counted one direction or two. Counting one still gives about 42 to 51 expected, against 13 found.
• Whether hits were counted per line or per line end is also unsettled.

The survey authors were emailed on 30 September 2026 and asked for their exact counting rule or, better, the line directions. No reply had arrived when the book draft was compiled on 4 October 2026. If they confirm 1 degree, the "less than half" wording drops out. If they send the directions, the test will be re-run on those.

Where it goes in the books

This is Book 6, The Sky Remembers. Chapter 4, "The Desert Calendar That Wasn't, and the One That Was", sets Nazca beside Chankillo, a row of thirteen towers that really does track the Sun. The Interlude, "The Machine Checks the Sky", gives the full account, marked provisional.

Check the arithmetic yourself. The bars of Figure 13, labelled 10 to 180 degrees, read: 126, 91, 103, 100, 153, 127, 177, 116, 129, 76, 104, 86, 124, 165, 140, 172, 177, 142. Reading each label as its slice's upper edge, the equinox window of 88 to 92 degrees takes 2 degrees from the slice labelled 90 (129 lines) and 2 from the slice labelled 100 (76): 2 × 12.9 + 2 × 7.6 = 41. Read the labels the other way and you get 49.