This article examines how electricity market liquidity, renewable production and cross-border activity together in combination explain price spikes in the Hungarian Power Exchange day-ahead auctions. In the applied logit model, the dependent variable representing the price spike is binary, and the key explanatory variable is a modified bid-ask spread depicting liquidity. Weather-dependent renewable production and the difference between exports and imports appear as control variables in the model. The empirical analysis was based on data from 2017 and 2018. The results show that the control variables have no effect on the bid-ask spread and that the model explains 96 per cent of the spikes well, with an AUC-ROC of 0.75 and a Gini coefficient of 0.5. Based on the results, it may be worthwhile for traders to incorporate their data from sales and purchase curves into their forecasts, as this will improve their chances of successfully predicting extreme prices.