Method & Instrument
How the assessment works — scoring, classification, and sources
Instrument v1.0.0
The Prakriti assessment uses 30 lifetime-trait items across three weighted domains:
Domains & Weights
Domain A — Structure & Body
Weight: 1.510 itemsBody frame; lifetime weight pattern; untreated skin; complexion tendency; hair; eyes; teeth and gums; joints; veins and tendons; voice.
Domain B — Physiology
Weight: 110 itemsAppetite; digestion; bowel habit; thirst; sweating; weather most disliked; sleep; dreams; stamina; cravings.
Domain C — Mind & Conduct
Weight: 0.7510 itemsMental pace; memory; stress response; decisions; money; default emotional tone; conversation style; routine; friendship and change; work rhythm.
Consistency Probes
Two hidden probe items (one structural, one digestive) are repeated in different wording. They are not shown as dosha-labeled and are used only to adjust the confidence band. Probe answers are compared to their original items; disagreement lowers confidence.
Vikriti Items (9 items, 0–3 frequency, 30-day window)
- Vata: Dry skin or lips; Bloating/gas/hard stool; Broken sleep, racing thoughts; Joint or nerve pain.
- Pitta: Acidity/reflux/burning; Loose stool or urgency; Rash/acne/mouth ulcers/temper.
- Kapha: Congestion or phlegm; Daytime lethargy and heaviness after meals; Recent weight gain and sugar craving.
Scoring & Classification
- raw[d] = sum of item weights for each chosen option mapping to dosha d ∈ {V, P, K}
- pct[d] = 100 × raw[d] / (raw.V + raw.P + raw.K)
- Sort pct descending → lead, second, third
- If (lead − second) ≥ 12 → Single type (Vata / Pitta / Kapha)
- Else if (lead − third) ≥ 12 → Dual type (Vata-Pitta / Pitta-Kapha / Vata-Kapha)
- Else → Sama (balanced / tridoshic)
Confidence Bands
Confidence combines three factors:
- answered = scored items answered / 39 (30 prakriti + 9 vikriti); below 85% → low, no definitive label
- spread = lead − second
- domainSD = standard deviation of lead-dosha percentage within each domain A, B, C
Bands:
- High: answered ≥ 95% AND spread ≥ 15 AND domainSD < 8
- Medium: answered ≥ 85% AND spread ≥ 8
- Low: otherwise — shows 'retake in 7 days'
When domainSD ≥ 8: 'Your structure reads [X] while your current digestive and mental pattern reads [Y] — that gap usually means a present imbalance sitting on top of your constitution.'
Season Resolution (Ritu)
Six ritus with Indian-calendar windows (mid-month boundaries). Resolution order:
- Date-based ritu for the user's hemisphere (southern flips by ~6 months)
- Override from user's 'weather now' answer (hot-dry / hot-humid / cool-dry / cool-wet / cold) mapped to matching ritu qualities
- Show resolved ritu + reason; allow manual override
Ritusandhi (transition): last 7 days of a ritu + first 7 of the next → blended chart, labelled as transition, tapering old regimen and phasing in new.
Chart Generation — Resolution Order
Every output line carries a source tag [Season] / [Vikriti] / [Prakriti] and a classical reference.
- 1. Season rules apply to everyone in that climate.
- 2. Vikriti overrides food QUALITIES (hot/cold, heavy/light, dry/oily) on conflict; season keeps timing, quantity, routine. Matching direction → mark priority.
- 3. Prakriti is the tie-breaker and framing: favour qualities OPPOSITE to the dominant dosha (Charaka Samhita Sutrasthana 7/39-41); for Sama type, habitual use of all six tastes.
- 4. Filter by food habit, allergies, and long-term staples (satmya) — never remove a lifelong staple without showing a reason.
- 5. Safety filter runs last and may only remove or soften, never add.
Classical Sources
- Ashtanga Hridaya — Sutrasthana Ch. 3 (Ritucharya), Ch. 7 (Dinacharya), Ch. 10 (Dravya)
- Charaka Samhita — Sutrasthana Ch. 5 (Matrashitiya), Ch. 6 (Tasyashitiya), Ch. 7 (Naveganniya)
- Sushruta Samhita — Uttaratantra Ch. 64 (Ritucharya)
Known Limitations
- Self-reported traits without clinical examination (no pulse, no tongue, no palpation).
- No disease diagnosis, no herbal formulas, no dosages, no Panchakarma instructions.
- Confidence bands are heuristic, not statistically validated against a gold standard.
- Seasonal rules are generalised; microclimate, altitude, and individual variability are not captured.
- Hindi translations are editorial where classical Sanskrit terms lack direct equivalents.
