Target alignment, dilution and forecast selection when cross-sectional forecasts share a common target
Abstract
Forecasters often score the same units per date against one standardized realized outcome. We show that every standardized forecast splits exactly into a component aligned with this common target and a component uncorrelated with it. Three consequences follow: forecast-error correlation largely mirrors forecast correlation and is therefore a poor measure of diversity; an equally weighted combination beats a no-information forecast only when average alignment is large relative to the combination's dispersion; and the gain from adding a forecaster separates into genuine improvement and mere dilution, which equal-weight admission can mistakenly reward. We develop a cautious selection rule, study it in simulations, and apply it to language-model forecasts of US equity rankings and mechanical signals ranking exchange-traded funds. Selection removes most dilution losses, but no combination beats the no-information forecast.
Related
- Flowz: When Does Learning Forecast Errors Add Value? Conditional Flow Matching for Financial Price Forecasts
- VGAS: Variance-Reduced Guidance and Adaptive Selection for Training-Free Reward Alignment in Discrete Diffusion
- Exact Fragmentation and Terminal-Cluster Selection in One-Dimensional Finite-Range Normalized Alignment
- Bayesian Group Testing Regression with a Shape-Free Dilution Curve
- Subgroup Inconsistency and the Dilution Effect in Levene's Test for Homoscedasticity
BibTeX
@misc{soleimani2026target,
title = {{Target alignment, dilution and forecast selection when cross-sectional forecasts share a common target}},
author = {Masoud Soleimani},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.26303},
eprint = {2609.26303},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.26303},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - UNPB
TI - Target alignment, dilution and forecast selection when cross-sectional forecasts share a common target
AU - Masoud Soleimani
PY - 2026
DA - 2026/09/22
PB - arXiv
DO - 10.48550/arxiv.2609.26303
UR - https://arxiv.org/abs/2609.26303
AB - Forecasters often score the same units per date against one standardized realized outcome. We show that every standardized forecast splits exactly into a component aligned with this common target and a component uncorrelated with it. Three consequences follow: forecast-error correlation largely mirrors forecast correlation and is therefore a poor measure of diversity; an equally weighted combination beats a no-information forecast only when average alignment is large relative to the combination's dispersion; and the gain from adding a forecaster separates into genuine improvement and mere dilution, which equal-weight admission can mistakenly reward. We develop a cautious selection rule, study it in simulations, and apply it to language-model forecasts of US equity rankings and mechanical signals ranking exchange-traded funds. Selection removes most dilution losses, but no combination beats the no-information forecast.
ER -