Forex Performance
The main reason for this post is to jot down some notes that we will have for later reference as you will see why toward the end of these thoughts.
A continuation of the past stats to improve performance post these are some additional thoughts with a solution that esmarting believes in.
What I think, but cannot prove, is that in ranging markets, which is most of the time, these adjustments could protect/benefit us a bit more.
Currently the Conquer trading system like any solid trend following systems make say 80% of our profits on 20% of our trades (trends) and the thought presented is that trading ranging markets could be more profitable.
More specificly according to the past data, there are some scenarios which could improve overall performance by more than 10% from the conquer’s historic performance.
Knowing ranging markets can be quite choppy and directionless here is the jest of what is being presented with some modifications, not when or when or what to enter, but when to took profit, modifying orders’ order, and targets.
Wisdom: one never knows when a trend will begin, thus we need to follow a trend system’s signals thru the ranging period and plan for drawdown – my 2 cents
Best scenario would be, according to this: same stop @0,33 ATR, BE @0,5 locking 10 pips. There everything remains intact with the current system.
Modifications are:
- order 1 tp@ 0,66ATR locking 0,22atr
- order 2 tp@ 1,32 atr locking 0,64atr
- order 3 tp@ 2,64atr
- orders would be: 1: 0,4% ; 2: 0,8%; 3: 1,2%
The frequency of hitting more tp would be greater and of course hitting first take profits more times, would help us in ranging markets. These scnerios consider the frequency that conquer currently hits take profits. These proposals/ideas been not been backtested/forward tested. According to this scenario, all 6 scenarios perform better that current settings for conquer, and 2 of them perform better by more than 10%.
The question then becomes isn’t past data being taken and then fitting in a set of criteria to that past data so it works or curve fitting. All computer programmers do this in trying to develop a trading system and this always fails when forward tested. Past performance data is taken curve fitting criteria to past data so that it works better. This is called curve fitting. This is so common with all computer programmers trying to develop a trading system, and this always fails when then forward tested.
I is a possibiliy of course, , but what esmarting is seeing from past performance is that half of our profits come from stopouts and half from targets hit. There seems to be merit but then I caution myself of falling into the same trap as all who try to develop systems through curve fitting. Each and everyone of them thinks they have something and we are talking about some pretty smart people who excel in math. They are not traders though and from curve fitting to past data when forward tested over and over again these systems fail.
This would be easy enough to backtest once the EA is finished since most of the system is the same (only changing a couple of variables) but I am sure it would backtest well since it was created through looking at our past performance. There is only one way to truly tell if there is something here that can truly help us and that is through forward testing. Itneeds to be forward tested for at least four months to know if you should continue or not, maybe with the EA on a demo account would be perfect.
The purpose of this post was mainly to have these thoughts written out so when the EA is complete this can be tested as well since the parameters or the core of the system is the same it will be easy enough to do. By the way, the auto script is ready and I will be writing that up on a thread and announcing it shortly.
Comments on this entry are closed.