• Updated analysis of 1,328 industrial Norway spruce elements tested from 2018–2025. • Cross-validated assessment of strength prediction for 1,000 C30 boards and 328 T22 lamellae. • MOE and density show limited predictive power, with cross-validated (R2 < 0.43). • Several parametric families fit the data reasonably; no single distribution is uniformly best. • Lower-tail design estimates are sensitive to the adopted statistical estimation procedure. This study evaluates the reliability of modulus of elasticity (MOE) and density as predictors of bending strength (MOR) for graded structural timber. A dataset comprising nearly 1,000 full-size industrial Norway spruce boards was tested under standardized four-point bending, with MOE derived from force–deflection curves, density estimated via mass-to-volume ratios, and moisture content assessed after kiln drying. Regression models, both linear and machine-learning-based, were developed and cross-validated to ensure robustness. The best-performing model using MOE and density achieved a cross-validated R 2 of 0.43 for strength prediction in the C30 class and 0.22 for the T22 tension lamellae subset. Comparisons with key European datasets, including Gradewood and Steiger et al., show that the predictive power of MOE for strength classification in industrially sourced material remains moderate. Although the present regression relationships differ from some earlier published models, these differences should be interpreted only as observational cross-study comparisons rather than as evidence of specific temporal or silvicultural causes.
Aloisio et al. (Fri,) studied this question.