Doctrine, Data, and the Death of DuPont

Doctrine, Data, and the Death of DuPont

  The full text of this Article may be found here.

36 Fordham Intell. Prop. Media & Ent. L.J. 678 (2026).

Article by Thomas A. Reichert*

ABSTRACT

[F]

or fifty years, trademark opinions have claimed to apply a comprehensive thirteen-factor test for trademark confusion. They are deeply mistaken. Using AI-powered analysis of over 4,000 TTAB inter partes decisions (2000–2025), this Article proves what practitioners have long suspected: in Section 2(d) adjudication, the test has collapsed to just two factors.

A simple categorical rule predicting confusion if and only if both mark similarity (Factor 1) and goods/services relatedness (Factor 2) (hereinafter “F1” and “F2” in figures, tables, and formulas) favor confusion achieves 99.55% accuracy across 4,651 comparisons. Cross-validated logistic regression confirms the pattern: a two-factor model achieves 99.46% accuracy, while adding the remaining eleven factors actually makes prediction worse (99.18%), a textbook
overfitting result. The thirteen-factor framework does not refine the
two-factor signal; it adds noise.

These findings reveal concrete doctrinal harms: practitioners brief eleven factors that do not matter, the Board analyzes them at length in every opinion, and the resulting complexity obscures what is actually a binary inquiry. The Article proposes reforms that center the two determinative factors and confine secondary considerations to narrow tiebreakers in genuinely ambiguous cases.

Finally, it advances a broader “multifactor collapse” hypothesis and outlines a research agenda for testing whether other legal balancing frameworks exhibit similar patterns where doctrinal complexity masks simpler underlying decision-making.