-
Notifications
You must be signed in to change notification settings - Fork 223
Backtesting
In financial analysis, backtesting seeks to estimate the performance of a strategy if it had been employed during a past period.
About backtesting:
Backtesting is the main use case of ta4j.
Once you constructed [your time series](Time series and ticks) and [your trading strategy](Trading strategies), you can backtest the strategy by just calling:
TimeSeries series = ...
TimeSeriesManager seriesManager = new TimeSeriesManager(series);
Strategy myStrategy = ...
TradingRecord tradingRecord = seriesManager.run(myStrategy);
That's it! You get a TradingRecord
object which is the record of the resulting trading session (basically a list of trades/orders).
By providing different strategies to the TimeSeriesManager#run(Strategy)
methods, you get different TradingRecord
objects and you can compare them according to analysis criteria.
Let's assume you backtested strategy1
and strategy2
over a series
. You get two TradingRecord
objects: record1
and record2
.
In order to get the profitability ratio of each strategy you have to give those records to an analysis criterion:
AnalysisCriterion criterion = new TotalProfitCriterion();
criterion.calculate(series, record1); // Returns the result for strategy1
criterion.calculate(series, record2); // Returns the result for strategy2
If you just want to get the best strategy according to an analysis criterion you just have to call:
TimeSeriesManager seriesManager = new TimeSeriesManager(series);
Strategy bestStrategy = criterion.chooseBest(seriesManager, Arrays.asList(strategy1, strategy2));
Ta4j comes with several analysis criteria which are all listed in the Javadoc.
Ta4j allows you to perform a well-known Walk-forward optimization. An example can be found here.
Ta4j documentation (2014 - 2017)