Synthesis of artificial bid sets for day-ahead power exchange models

We consider the scenario, when detailed bid data of a day-ahead power exchange is not accessible due to privacy issues, but the statistical extract of the data is available. More precisely, we assume that in the case of standard bids, the approximate joint distribution of bid quantity and price is available, while in the case of block bids we know the average and standard deviation of the bid parameters. Based on these statistical features we re-generate the bid set, and analyze, how much the market outcome differs from the original result, depending on the detail of the statistics used. In addition, we analyze how much the description details affect the performance of the regenerated bid set.