# Vumarii 2019 / Dunhuang Map 13: 34-unit traversal audit

## Finding

The apparent `34 = 34` connection is real as a **count**, but it is not a decryption. Map 13's scholarly Table 5 lists 34 asterisms. The complete Vumarii image has 34 whitespace-separated groups only if its two visible signature groups (`TENGRI137` and `666AB731`) are counted with the seven-line body. The body itself has 32 groups and 154 glyph slots.

The narrow count-derived test also fails: treating every group length as an asterism's published on-map star count gives no unusual agreement in the canonical Table-5 order or its reverse. That rejects this convenient length traversal, not all possible uses of the historical chart.

## Inputs and rule

- Vumarii Layer-1 transcription: `archive/gaps/tr1-00-files/vumarii-2019-TRANSCRIPTION.txt` (SHA-256 `2c6eb4492a3975c034e9fb4ba75ae676acad0fc979b76e7a0463bd77170ad6b3`)
- Full groups / body groups / signature groups: `34` / `32` / `2`
- Body / signature slot count: `154` / `17`
- Scholarly reference: `archive/recon/reference-papers/bonnet-bidaud-praderie-whitfield-2009-dunhuang-chinese-sky.pdf` (SHA-256 `cd449d46586d1303e1225bf418554400037b71b765da27b3e1e9da4cd991c7e1`), Table 5, pages 25-27 of the archived PDF
- Table order: north-to-south, then east-to-west. The test grants the reverse order too; neither order is selected by the Tengri image.
- Statistic: exact equality of group length and Table-5 `Nb of * on map`, Pearson correlation, and total absolute difference. Controls randomly permute the 34 star counts, while retaining the same best-of-forward-or-reverse selection.

## Results

| Order | Exact length matches | Pearson r | Total absolute difference |
| --- | ---: | ---: | ---: |
| Canonical Table-5 order | 5 | -0.0627 | 99 |
| Reverse order | 5 | 0.2011 | 85 |

Across `100,000` fixed-seed controls, the best-of-two-direction exact-match result is matched or exceeded `53,145` times (empirical `p = 0.5315`). The absolute correlation is matched or exceeded `46,147` times (`p = 0.4615`); the minimum total difference is matched or improved `38,574` times (`p = 0.3857`).

## Why the count does not select a decoder

- The 34-group count is available only when the two visible signature groups (TENGRI137 and 666AB731) are counted with the seven-line body. The signed image does not state that signatures are asterism tokens.
- The historical table supplies a canonical asterism order, but does not say to convert star counts to text-token lengths. The Vumarii image supplies no stated instruction connecting the diagram to the body.
- The source's red/black star-school notation was removed in the cleaned Vumarii drawing. The red glyph labels identify parts of the Vumarii legend/signature; they do not supply a 34-position source mapping.
- Therefore this result should prevent `34 = 34` from being presented as a solved payload. A future claim still needs an author-selected traversal, operation, and output that accounts for the glyph body.

## Direct comparison table

| # | Vumarii slots | Signature? | Map-13 asterism | Stars on map | Canonical equal? | Reverse stars | Reverse equal? |
| ---: | ---: | --- | --- | ---: | --- | ---: | --- |
| 1 | 3 | no | Tianchu | 6 | no | 6 | no |
| 2 | 4 | no | Wudizuo | 5 | no | 4 | yes |
| 3 | 2 | no | Chuanshe | 7 | no | 1 | no |
| 4 | 7 | no | Tianzhu | 5 | no | 1 | no |
| 5 | 2 | no | Liujia | 5 | no | 3 | no |
| 6 | 5 | no | Huagai | 7 | no | 1 | no |
| 7 | 2 | no | Gouchen | 6 | no | 5 | no |
| 8 | 6 | no | NI 1 | 1 | no | 4 | no |
| 9 | 4 | no | Tianhuang | 4 | yes | 7 | no |
| 10 | 7 | no | Ziwei | 15 | no | 3 | no |
| 11 | 4 | no | Zhuxiashi | 1 | no | 3 | no |
| 12 | 6 | no | Nüshi | 1 | no | 1 | no |
| 13 | 2 | no | Tianpei | 6 | no | 1 | no |
| 14 | 6 | no | Shangshu | 5 | no | 1 | no |
| 15 | 4 | no | Beiji | 4 | yes | 1 | no |
| 16 | 5 | no | Sifu | 4 | no | 4 | no |
| 17 | 5 | no | Neijie | 6 | no | 8 | no |
| 18 | 6 | no | Bagu | 8 | no | 6 | yes |
| 19 | 4 | no | Tianchuang | 4 | yes | 4 | yes |
| 20 | 6 | no | Taiyi | 1 | no | 4 | no |
| 21 | 7 | no | Tianyi | 1 | no | 5 | no |
| 22 | 6 | no | Tai | 1 | no | 6 | yes |
| 23 | 6 | no | Tian | 1 | no | 1 | no |
| 24 | 3 | no | Sangong (2) | 3 | yes | 1 | no |
| 25 | 7 | no | Tianqiang | 3 | no | 15 | no |
| 26 | 5 | no | Beidou | 7 | no | 4 | no |
| 27 | 6 | no | Tianli | 4 | no | 1 | no |
| 28 | 5 | no | Wenchang | 5 | yes | 6 | no |
| 29 | 2 | no | Xuange | 1 | no | 7 | no |
| 30 | 4 | no | Sangong (1) | 3 | no | 5 | no |
| 31 | 6 | no | Xiang | 1 | no | 5 | no |
| 32 | 7 | no | Taiyangshou | 1 | no | 7 | yes |
| 33 | 9 | yes | Shi | 4 | no | 5 | no |
| 34 | 8 | yes | Tianlao | 6 | no | 6 | no |

## Reproduce

```powershell
& '<bundled-python>' scripts\audit_vumarii_2019_dunhuang_traversal.py
```