Forecasting human pricing decisions in competitive dynamic pricing

Cesaret, Bahriye
Journal of Revenue and Pricing Management · Springer Nature · 2026-09-12 · DOI: 10.1057/s41272-026-00624-2

Abstract

Equilibrium models provide normative guidance for competitive dynamic pricing but may not accurately predict human decisions. Using secondary laboratory data from capacity-constrained human–automated and human–human duopolies, we develop two behavioral forecasting heuristics. Competitive Anchoring combines persistence in a seller’s preceding price with adjustment to the competitive price gap; Anticipatory Undercutting extrapolates recent competitor-price movements and prices one cent below the forecast. We evaluate the methods on nonoverlapping holdout participants and cohorts against normative, estimated, and parameter-free alternatives. Competitive Anchoring performs particularly well for low-capacity sellers, whereas Last Own Price performs best for high-capacity sellers. Adding the Anticipatory Undercutting trend signal to Competitive Anchoring produces little change in holdout accuracy for either capacity role, while adding time and capacity variables does not improve holdout accuracy for either role. Counterfactual simulations show that recursive interactions between pricing rules can limit transaction-price erosion and generate higher average revenues than observed human–human outcomes within the experimental environment, although results depend on capacity role, the competing rule, and initial prices. Predictive accuracy and economic performance therefore require separate evaluation.

Read at publisher →

Similar articles