Dirty Data

It is a fact that real-world time series data distributed by vendors contain at least some errors. Their source can be traced to database corruption or faulty reporting by the exchanges themselves. Mechanical trading systems relying on these data can be sensitive to the errors that can lead to invalid entry and exit signals. These false alarms are at least a source of frustration and can potentially seriously damage a trading account. HistoryMaker simulates those errors by applying a data corruption algorithm to the generated time series. By using such data for system testing, a system's immunity to dirty data can be tested.