falcon_a's Introduction

FALCON_A

Designed to work best with Decision Support System developed inside Lazy Trading Project

https://vladdsm.github.io/myblog_attempt/topics/lazy%20trading/

Introduction

Model-based trading robot. Main features are listed below:

Does not require optimization of parameters
Learns from past trading experience
Able to read Market Type
Able to log Market Type status to the file
Works with Reinforcement Learning Policy to identify which market type is more suitable to trade in
IMPORTANT: Setup Environmental Variables to syncronize and use scripts inside the repository

Reference

This EA is deprecated.

It was previously explained in the Udemy course

Developing Self Learning Trading Robot

This repository is to keep existing EA working on 28 currencies and forcedly trained with kNN methods used code from here

Original code: Stat_Euclidean_Metric.mq4

Repository will not be maintained or even be deleted in the future.

How this will work?

Go to Tester and set up parameter Base = True.

Select large amount of time period and start trades simulation.

Then change parameter Base = False.

Run simulation again

Finally in order to use in trading mode move files from /tester folder to the /Files folder

Optimization Method (on Saturdays!)

Check which systems require optimization in the folder 'TEST' see file 're-train'
Write down magic numbers and corresponding currency pair
Select robot FALCON_A in Terminal 2 strategy tester
Select M15, Open Prices, Spread 3
set up date to be | Start: Today - 2month; End: Today
during 'optimization' set max 25 trades, lots = 0.01, Base = True
set up magic number to the one selected in the table (see point 2)
press start in the tester to generate trades
repeat steps 7 - 8 for all magic numbers in the table
execute script _CopyDATFilesToProd.bat to copy generated files to the T1, T3, T4...
Make sure that platforms/computer are restarted