Ugly Duckling Theorem Calculator

🦆

Watanabe: any two distinct objects are equally similar without predicates

Setup

Objects N (2–6)

Predicate Counts

Objects (N)

4

Items to classify

Possible subsets

2ᴺ = 16

Binary predicates over N objects

Binary relations

2^N² = 65,536

All pairwise comparisons

Unique pairs

6

Object pairs to compare

Unweighted Similarity — All Pairs Equal

Watanabe's theorem: Without feature weighting, every pair of objects shares exactly the same number of predicates (2^(N−2) subsets). All pairs are equally similar — the "ugly duckling" is as like a swan as two swans are like each other.

Obj 1 ↔ 2
0.2500
equal similarity

Obj 1 ↔ 3
0.2500
equal similarity

Obj 1 ↔ 4
0.2500
equal similarity

Obj 2 ↔ 3
0.2500
equal similarity

Obj 2 ↔ 4
0.2500
equal similarity

Obj 3 ↔ 4
0.2500
equal similarity

About the Ugly Duckling Theorem

Satosi Watanabe (1969) proved that without a prior preference for certain features, any two objects are equally similar to each other — making classification arbitrary without background knowledge.

Machine learning implication: Every classification algorithm implicitly encodes feature weighting. There is no "bias-free" learner — the No Free Lunch theorem for similarity is a direct corollary.

Practical takeaway: Feature engineering and domain knowledge are not optional shortcuts — they are the fundamental source of useful generalisation.

Add Ugly Duckling Theorem Calculator Free On Your Website

Paste this HTML on your site.

Embed code

<!-- Calculators.is embed -->
<iframe src="https://calculators.is/embed/ugly-duckling-theorem-calculator" title="Ugly Duckling Theorem Calculator" width="100%" height="620" style="max-width:720px;border:0;border-radius:12px;display:block" loading="lazy" referrerpolicy="strict-origin-when-cross-origin"></iframe>

What is Ugly Duckling Theorem Calculator?

Ugly Duckling Theorem Calculator demonstrates the Ugly Duckling Theorem from pattern recognition, showing why we need to assign importance to features when comparing objects.