Declared spelling-form score: a reproducible three-component formula

Methodology

Research question: Which pronunciation-covered spellings receive the highest values under the declared three-component formula, and how do its score bands distribute across ranking cohorts?

Why it is useful: Shows exactly how the declared formula ranks tier 1-2 spellings in this build; the score is an audit and exploration measure, not validated spelling difficulty or teaching guidance.

Included

• spelling-difficulty.covered-records: Include safe normalized ASCII spellings with at least one retained CMUdict pronunciation variant, scored against the primary variant.
• spelling-difficulty.tier-cohorts: Report the full covered cohort and the tier 1-2 cohort separately; exemplar rows come only from tiers 1-2.

Excluded

• spelling-difficulty.product-policy: Exclude spellings matched by the exact first-party content-safety policy before scoring.
• spelling-difficulty.no-pronunciation: Exclude records without a retained pronunciation variant; they receive no score, and nothing is inferred about their spelling difficulty.

Transformations

1. spelling-difficulty.surplus-v1: Compare normalized letter counts with the primary retained variant's phoneme count; each surplus letter contributes three declared points.
2. spelling-difficulty.grapheme-scan-v1: Scan each spelling left to right against the declared vowel-sequence list, longest sequence first, summing declared weights.
3. spelling-difficulty.doubling-v1: Count adjacent identical letter pairs; each pair contributes two declared points.
4. spelling-difficulty.banding-v1: Partition total formula scores into five declared numeric bands and rank the tier 1-2 exemplar by score with alphabetical tie-breaks, limited to 20 rows.

Calculations

research.count-v1

Count records in each declared population stage and score cohort. Formula: count(records).

research.share-v1

Calculate a cohort's exact share of its named denominator. Formula: 100 * subset_count / cohort_count.

research.ratio-v1

Divide letter count by primary-variant phoneme count, half-up to 2 places. Formula: letters / primary_phonemes.

research.maximum-v1

Select the highest heuristic score within a declared cohort. Formula: max(formula_score).

orthography.formula-component-v1

Compute the three declared formula components: letter-phoneme count surplus, vowel-sequence points from a longest-first scan, and doubled-letter pairs. Formula: surplus = max(0, letters - primary_phonemes); points = sum(declared grapheme weights, longest-first scan); pairs = count(adjacent identical letters).

orthography.formula-score-v1

Combine the declared components with fixed version-controlled weights. Formula: 3 * max(0, letters - primary_phonemes) + vowel_grapheme_points + 2 * doubled_letter_pairs.

Limitations

• Every weight is a fixed heuristic choice; the score is not a measured misspelling or reading-error rate.
• Only the first retained pronunciation variant is compared; other variants and dialects are ignored.
• CMUdict transcribes a primarily North American convention, so letter-phoneme surplus inherits its choices.
• The declared vowel-sequence list is finite; spellings outside it can still receive points from letter-phoneme count surplus or doubled letters.
• The vowel-sequence and doubled-letter components inspect spelling shape only; they do not compare those sequences with pronunciation.
• Tier cohorts are build-specific ranking bands, not universal usage claims.