Genre share of NYT Combined Print & E-Book Fiction list

Methodology

Starting probabilities are derived from a normal approximation model anchored to the 2025 full-year Romancing the Data genre share estimates, which track appearances (not unique titles) across all 52 weekly NYT Combined Print & E-Book Fiction lists.

Data sources

• Primary anchor: 2025 full-year shares — Romance & Romantasy 42%, Mystery & Thriller 34%, Literary Fiction 17%, SFF & Other 7%
• Monthly variance: April–May 2025 monthly samples used to estimate within-year standard deviation, scaled down by √12 to approximate annual aggregate smoothing
• Long-run trend: 15-year top-10 title shares (2011–2025) used as a shrinkage prior; R&R has risen from ~20% to 42% over this period while Literary Fiction has declined from ~38–40%
• Momentum signal: Fantasy unit sales fell 8.7% in 2025, suggesting R&R share growth is plateauing

Model

For each genre, a forecast mean and standard deviation are estimated as follows:

• Mean: weighted blend of 2025 anchor and long-run base rate (shrinkage weight 10–15% toward long-run), plus a momentum adjustment
• SD: monthly sample standard deviation ÷ √12, plus an irreducible uncertainty floor of 1.5pp
• Year-specific projection: 2027 and 2028 apply incremental mean drift and SD widening to reflect compounding trend uncertainty (+1.2pp SD per year for R&R, +0.8pp for others)

Probabilities are P(genre share ≥ threshold) evaluated from the normal survival function, clamped to [0.001, 0.999]. Thresholds outside the informative range are assigned the clamp boundary value (implying zero probability mass in that region), giving a unimodal implied distribution for each genre and year.

Resolution

Each calendar year resolves independently once the Romancing the Data annual summary is published (typically February of the following year). Romantasy is bundled with Romance, consistent with that source's taxonomy.