# C1172 — Reject rank alone as an explanation score

## Question

Could a formally smaller numeric dimension conceal a larger description of the source family?

## Inputs

```json
{}
```

## Sources

```json
[
  "C1171",
  "Strategy §6"
]
```

## Opened utc

2026-09-28T05:11:45.208647+00:00

## Results

```json
{
  "rank_counterexample": {
    "path": "model/rank_only_compression_counterexample.json",
    "sha256": "b54c1d3fc222a1ffb0f88d10a1ba8d4eb9df7b7f3e9d7c9e5c5c2ca7b4857d1a",
    "bytes": 469
  }
}
```

## Finding

A one-dimensional fitted ray can encode every observed value only by placing all56 values in its coefficients. This shows why lower rank is not itself the Strategy’s desired compression; source rules, masks and exceptional choices must be counted.

## Reassessment

Establish the algebraic domain of the row moves before using them as a shared operation grammar.

## Checks

```json
{
  "ray_rank1": true,
  "all56_stored": true,
  "no_score_weights": true
}
```

## Closed utc

2026-09-28T05:11:45.209858+00:00

## Record provenance

Exact record hash: `d35e57433c43dcf73a70acf751376cb503fbacbef9062064d5e6ecc7a08ba5f8`

Source journal: `Research_Cycles/C1132_C1431_Recovered/journal.json`