Skip to content
ASTROYANA

Astroyana research · marriage matching

How hard is the Guna Milan cut-off, really?

Across India, many families use the 36-point Ashtakoota or Guna Milan score when considering a match. Eighteen is the classical recommended minimum. We scored 50,000 random synthetic couples across all eight kootas. The study shows how demanding the threshold is and how often classical rules cancel Nadi and Bhakoot dosha.

79.1%

clear the 18-of-36 minimum

24

average score, out of 36

26.2%

blocked by Nadi dosha (after cancellation)

0.57%

score a perfect 36/36

vanishingly rare

The 18-point bar is high by design

20.9%, or about one in five, score below the classical 18-point minimum. Most pairs land in the “acceptable” to “very good” bands, while a perfect 36 is exceptionally rare.

Score tiers
  • Adhama — not recommended (under 18)20.9%
  • Madhyama — acceptable (18–23)25.5%
  • Very good (24–27)17.1%
  • Excellent (28–31)20.2%
  • Best possible (32–36)16.3%

50,000 random couples, classical Ashtakoota grading.

Score distribution (0–36)
0–82.5%
9–1718.3%
18–2325.5%
24–2717.1%
28–3120.2%
32–3616.3%

How often Nadi and Bhakoot actually block a match

Nadi (8 points) and Bhakoot (7 points) carry the most weight and are often treated as serious obstacles. Both have classical cancellation (parihara) rules. Nadi dosha appears in 33% of pairs and is cancelled in 20.7% of those cases. Bhakoot appears in 33.4% and is cancelled in 38.1%. The uncancelled blocking rates are therefore much lower than the initial flags suggest.

Present vs. actually blocking
Nadi dosha present33%
Nadi dosha blocking (uncancelled)26.2%
Bhakoot dosha present33.4%
Bhakoot dosha blocking (uncancelled)20.7%

See the classical cancellation logic in when Nadi dosha is cancelled.

Which gunas are hardest to score?

Each koota is shown as the average share of its points that a random pair earns. The smaller kootas tend to fill easily; the heavy ones — Bhakoot, Nadi, Gana — are where matches gain or lose the most ground.

Average points earned, per koota
Varna · 0.62/162%
Vashya · 0.98/249%
Tara · 1.67/356%
Yoni · 2/450%
Graha Maitri · 3.01/560%
Gana · 4.26/671%
Bhakoot · 5.55/779%
Nadi · 5.91/874%

What this means

A score below 18 is common and is not the whole analysis. Nadi and Bhakoot flags are often cancelled under the classical rules, while 36/36 is exceptionally rare. Guna Milan is therefore a first filter rather than a final verdict. A full classical match also considers Mangal dosha, the seventh house and its lord.

Methodology

Synthetic study — no real user data. Birth moments are sampled uniformly across the date range and clock; birthplaces are sampled across major Indian cities weighted by population. Percentages are the mathematical base rate per birth moment, not a population census. Reproducible from the committed script with the fixed seed. Each data point is one random synthetic couple scored on the 8-koota, 36-point Ashtakoota system; the classical minimum to recommend a match is 18.

Sample size
50,000 charts
Birth date range
1950-01-01 to 2009-12-31
Timezone
IST (UTC+5:30)
Birth hour
uniform
Locations
population-weighted across 25 major Indian cities
Ayanamsa
lahiri
Nodes
mean
House system
whole_sign
Seed
20260603

FAQ

What percentage of couples clear the 18-Guna minimum?

In a study of 50,000 random pairs, 79.1% reached the classical 18-of-36 minimum, with an average score of 24. A perfect 36 is extremely rare. Only 0.57% of random pairs scored 36/36.

How often does Nadi dosha actually block a match?

Nadi dosha is present in 33% of pairs, but classical rules cancel it in many of those — it ends up blocking (costing the 8 Nadi points uncancelled) 26.2% of the time.

Is a low Guna Milan score a dealbreaker?

Not on its own. Guna Milan is one layer; classical matching also weighs Mangal dosha, the strength of the 7th house and its lord, and dasha timing. The score is a starting point, not a verdict.

Free to cite with attribution to Astroyana. Figures are the mathematical base rate per birth moment from a synthetic sample, computed with our Swiss-Ephemeris (Lahiri) engine — not a survey of real people, and not a claim about anyone’s life.