Electricity market models are widely employed to study the role, impacts and economic viability of new technologies. Sources of arbitrage, such as storage and transmission, are increasingly seen as essential for integrating higher shares of variable renewables. Understanding their operation and business case requires models which accurately represent time-series of wholesale electricity prices. We show that the prevailing assumption of generators bidding short-run marginal cost, such as in the merit order stack, substantially underestimates the spread and volatility of hourly wholesale prices. To compound this, the lack of transparent outputs from previous electricity market modelling studies makes it impossible to scrutinise the prevailing methods or provide a detailed inter-comparison. We demonstrate a simple modification to the short-run marginal cost approach that delivers improved variability in modelled prices: allowing generators to make a spread of bids, below cost for their first megawatts of capacity, above for their last. Using this model we demonstrate the impact of price variability on the operation and profitability of storage, highlighting the urgent need for greater awareness of this aspect of market model performance.