Agent-based computational economics is becoming widely used in practice. This paper explores the consistency of some of its standard techniques. We focus in particular on prevailing wholesale electricity trading simulation methods. We include different supply and demand representations and propose the experience-Weighted Attractions method to include several behavioural algorithms. We compare the results across assumptions and to economic theory predictions. The match is good under best-response and reinforcement learning but not under fictitious play. The simulations perform well underflat and upward-slopping supply bidding, and also for plausible demand elasticity assumptions. Learning is influenced by the number of bids per plant and the initial conditions. The overall conclusion is that agent-based simulation assumptions are far from innocuous. We link their performance to underlying features, and identify those that are better suited to model wholesale electricity markets.
