Being able to anticipate where price might evolve after a trade can be useful to plan potential exit points, set take profits/stop loss as well as knowing when user set ones might get hit…etc.

All backtesting toolkits include a forecasting tool able to provide forecasts from the most recent trade executed by the backtester.

Using Forecasts

Forecasts can be enabled from the “FORECAST” settings group by enabling the “Show Forecast” toggle.

Each forecast is influenced by the price evolution made during previous respective long and short trades, as such forecasts for long positions are different from the ones of short positions.

Users can select the maximum length of forecast using the “Forecasting Length” setting.

Because of the forecasting algorithm used, forecasts can have a lower length than the one selected by the user depending on the average bars in trade.

Very frequent trades will generally return short price forecasts due to the lack of available data.

Forecasting Area

By default, forecasts are displayed alongside an area, indicating where future prices might evolve. The area extremities can help users potentially determines exit points with there interaction with the price, but can also help reference the performance of a current trade relative to past trades.

Areas extremities are representative of past trades performances. For example:

  • For long positions: A lower extremity not significantly going below the entry price can indicate good past performance for long positions.
  • For short positions: An upper extremity not significantly going above the entry price can indicate good past performance for short positions.

If an area is no longer visible in a point in time of the forecast it means that there is not enough data for that point in time.

Forecasting Memory

Users can control the influence older trades have on the forecast using the “Maximum Forecast Memory” setting, with lower values using a shorter term memory, discarding older information more quickly.

Low values of this setting allow obtaining more diverse forecasts for new trades, while higher values will return forecasts less subject to change over time.