Designing an Automated Trading Program
- Thread starter expiated
- Start date
The program takes all costs into account, and the lot size is a mere 0.01.
I'm not familiar with trading forex. If you say that trade commissions of "a few pennies here and there" is accurate than that is fine with me. I only wanted to ensure that you had taken realistic trading costs into account. Otherwise you might have drawn incorrect conclusions from your back test.
It's one of the reasons I now prefer trading foreign currency pairs over trading equities, which cost me an arm and a leg every time I wanted to trade a stock.
And the other version you need to follow up on was Version #9... just so you don't forget.
Maybe, but an hour or two earlier it WAS letting me test using five- and even one-minute charts (if I remember correctly).
What's your Sample period though? In your video you posted, it looks like you are starting back in 2016, most brokers do not hold enough high-resolution data for you to poll 3 years of 1 minute or 5-minute data but maybe you can get a year's worth.
I have 5 years of tick data on the E-mini, its well over 1 billion rows of spreadsheet data, I also run my strategy on a Co-located server and I just migrated to a New server, it took an entire weekend to transfer that tick data, now Imagine trying to deal with years of 1 minute data on every Forex Pair, it would require massive amounts of storage and brokers can't justify the costs.
Hey, do you mind posting the details page with the Win/loss ratio, avg loser/winner and stuff?
I am an idiot, I forgot that I had my MT4 programmer make me a Chop Indicator. If I remember correctly, this indicator is a slope filter, so You add this to your chart, you define the lookback period and the threshold.. It works like this, if the MA 5 periods ago (or whatever you choose) is Less Than +10 pips and greater than -10 pips then the MA turns Blue for Neutral. We are straddling the range like bollinger bands with a (+/-) variable to define this range, in this case i used a random number of 10.
The idea is that Chop is usually within a Range, so the MA will center on that range and move sideways, so the MA's range is going to be a much smaller range than price, and when the MA stays between the High/Low threshold you set, it turns blue and outputs a "0", when it goes above the threshold it outputs a "1" saying its ok to go long and turns green, when it is below the threshold, it turns red, outputs a "2" to let my system know its ok to go short. I have to dig around for it, but you can have it if you want to put it on your chart and observe it. If it works just have your programmer integrate it into your system.
I currently use Ninja Trader and have outgrown it as well, I am looking at having my programmer code me my own custom Algo on top of Matlab. Once in a while if I come up with a new Idea, I will pull out Meta trader because I can get programs built for dirt cheap on that platform. Also the other problem is that I cant remember if its in MT4 or MT5, I think you are using MT4.
Last edited:
My sample was only a month, and it was working for a while. I had to switch to the longer time frame when testing on four-hour and especially daily charts to have enough data to evaluate.