Java-based custom local back-testing platform for foreign exchange trading.
This app was created to have a custom backtesting system for evaluating fx strategies.
It is based on JForex platform from Dukascopy.
java source code is in project folder.
html customized report web code is in JFExtendedReports folder.
Classes in fradiani.env establish the communication with tick data loaded from csv files (this should be improved to work with a db and reduce space size) that for simplicity of manipulation emulate the tick data that can be publicly downloaded from dukascopy servers.
TickServer is in charge of loading the tick data and serve it to the TestingEnv, it uses the following logic:
- Chunks of string lines are read from the csv files an kept in cache
- The lines of the pairs are ordered serving those that are closer in time.
- When a chunk is completed, another one is loaded until the data of a particular day is done.
TestingEnv is the platform execution environment, run this class to test a particular strategy.
The testing environment class loads a strategy in order to execute its specific trading rules, the global parameters for a backtest are defined in
fradiani.strategies.ALX_COMMON i.e: periodStart, periodEnd, equity, strategyName, where strategyName is the full class name of the strategy
to be executed.
In order to be able to run tests, the following things need to be set:
-
In
TickServer, edit the path to load CSV tick datapublic static final String TICK_PATH = "/path/to/csv_data/";
the structure for storing the files is PAIR/YEAR/PAIR_YEAR_MONTH_DAY.csv, e.g: EURUSD/2013/EURUSD_2013_03_01.csv
the csv data can be obtained directly from Dukascopy or using tools like Tick Data Downloader.
-
in
TestingEnv, the path for the reporting must point to the JFExtendedReports folder, this is the html/js code that shows the performance report of strategies tested.public static final String REPORTS_PATH = "/path/to/AlxFx/JFExtendedReports/public_html/";
The SampleTrades class could be used as template and as guide to understand the life-cycle of a strategy, which is very similar to the way strategies are defined in the JForex platform. A couple of simple technical indicators are defined in the fradiani.indicators package, but they are at a very primitive stage and should be improved.