Backtesting System Modes

The FMZ Quant Trading Platform divides backtesting modes into Real-tick Level backtesting and Simulated-tick Level backtesting. Real-tick Level backtesting is completely based on complete historical data for backtesting; Simulated-tick Level backtesting generates tick data based on real candlestick data for backtesting. Both are based on real historical data for backtesting, but Real-tick Level backtesting has more accurate data and more reliable results. It should be noted that backtesting only reflects the performance of strategies under historical data, and historical data cannot fully represent future market conditions, so backtesting results should be treated with a rational and objective attitude.

Simulated-tick Level backtesting generates simulated tick data based on the underlying candlestick period, generating up to 12 backtesting time points on each underlying candlestick period. Real-tick Level backtesting uses real collected second-by-second tick data, which has a large amount of data and slower backtesting speed, so it is not suitable for backtesting particularly long time ranges. FMZ Quant's backtesting mechanism allows strategies to trade multiple times on a single candlestick, avoiding the limitation of only being able to trade at the closing price, ensuring accuracy while balancing backtesting speed.

Backtesting System Mechanism Description

• Simulated-tick Level
  Simulated-tick Level backtesting generates tick data for backtesting based on the underlying candlestick data of the backtesting system, simulating tick data within the price framework composed of the high, low, open, and close prices of the given underlying candlestick bar according to specific algorithms, serving as real-time tick data on the backtesting time series, which is returned when the strategy program calls the interface. For details, please refer to: Backtesting System Simulated Level Mechanism Description.

• Real-tick Level
  Real-tick level backtesting uses real tick-level data in the bar time series. For strategies based on tick-level data, using real-tick level backtesting is closer to actual conditions. The ticks in real-tick level backtesting are real recorded data, not simulated. It supports depth data, market trade record data playback, supports custom depth, and supports trade-by-trade data. Real-tick level backtesting data supports a maximum of 50MB, with no limit on the backtesting time range within the data limit. To maximize the backtesting time range as much as possible, you can reduce the depth level value settings and not use trade-by-trade data to extend the backtesting time range. Call GetDepth and GetTrades functions to obtain playback market data. At a certain market data moment on the timeline, calling GetTicker, GetTrades, GetDepth, GetRecords will not move time forward multiple times on the backtesting timeline (will not trigger jumping to the next market data moment). Repeated calls to any of the above functions will push the backtesting time forward on the backtesting timeline (jump to the next market data moment). When using real-tick level backtesting, it is not advisable to select too early a time, as earlier time periods may not have real-tick level data.

Real-tick Level and Simulated-tick Level mode backtesting system order matching mechanism: Order matching is executed based on price-taking and full-fill execution. Therefore, partial fill scenarios cannot be tested in the backtesting system.