Although the "base" system I am using for this contest was set in stone (ie, no discretion), the "add on to winners" portion of the strategy, where I increase size as the profit in a position grows, had discretion in it:
Initial entry - 100% mechanical, no discretion
Exit - 100% mechanical, no discretion
Additional entries, for profitable positions - 100% discretion
Although I liked the idea an adding on to winning positions, and tests I have run with other systems show that the idea is sound, I did not like the discretionary aspect of it.
So, I decided to fix this.
I went back, and divided the historical trades roughly in 2 - an in-sample "test" period, and an out-of-sample "verification" period.
I ran a couple of simple ideas (for when to add on) through the "in period" and chose one, based on test results.
I then checked, and the scheme holds up on the "verification" out-of-sample period.
Normally, I'd like to run this in real time (without real money) for a few months. But, in this case I am going to put it to use immediately.
So, here is what this system looks like now:
Initial entry - 100% mechanical, no discretion
Exit - 100 mechanical, no discretion
Additional entries, for profitable positions - 100% mechanical, no discretion
Why am I eliminating discretion? I'll talk about my reasons in a future post.
We will see how this works out!
Contest Update:
My Performance:
Your equity curve looking good now. If there is continuation today, it would be fantastic!
ReplyDeleteBy the way, just a question on market selection. Do you have an algorithm to select the markets to trade? In practice, does this holds up?
Thanks for the question.
ReplyDeleteFor the system I am using for the contest, I had an idea in mind, but I knew it would not produce enough trades in 1 market per year. But, over 20+ markets, it produces an acceptable number of trades per year (around 75-100, if memory serves me right). So all testing and evaluation was done with all these markets.
When everything was complete, I think all but 2 or 3 markets were profitable. But, I trade them all, because eliminating the loser markets would basically be cherry picking, and I like to avoid that when I can.
I do not use this multiple market approach for all systems I develop. For example, I recently created 5 systems, all unique, and all different markets. Sort of the opposite approach to what I am doing here.
5 Systems? You referring to SFE?
ReplyDeleteYes, thanks for noticing!
ReplyDelete