In today’s article, I’ll show you step-by-step how to reduce the efficiency of working with Internal Markets to 70% in just seven minutes (on a regular computer – using a laptop would take a little more time). So let’s get to that.
In this example, we will work with a very simple idea, which is to buy EMD.D (15-minute chart) when the first bar closes above yesterday’s close (i.e., the time to exchange at 8:45 or rather 15:45) local time for me in Spain).
This simple idea does not look bad at all; it seems to have considerable potential – although there is still a long way to go to achieve a complete and successful system.
There is no need to reduce the number of trades, filter out the “bad ones”, increase the average profit per trade and, most importantly, significantly reduce the discount. It’s not important to present full statistics either – a first look at equity already shows what we’re talking about. And how do you achieve that improvement? That’s exactly the job of the Internal Market!
First, I put the above code in the Market Internals smart code and prepared a special board / workspace for it.
This smart MI code has some of my MI requirements (it took me 6 months to put it all together) and now I let TradeStation run all those requirements and let it find the most suitable one. To avoid the risk of over-optimization from the outset, we need to apply this process to 70% of the sample data and save the remaining 30% as out-of-sample data.
Now we do the optimization, which will take about 2 minutes on average on a computer (about 6 minutes on a laptop).
As soon as the optimization is complete, I organize the sample data using the fitness function. In this case, it was the TS index.
Now is the time to choose only one of the TOP solutions. Most often there is more than one solution available. Among the best results we can find are 5-10 (this is not a classic system optimization, but also the search for the most suitable switches, i.e. Market Internal Conditions) there are many of these conditions in the smart code, so more than one can work very well). In this case, my favorite solution this time is the second issue. This is why I will choose this solution and look at the sample data.
The result looks great, so I’ll check the data outside of the sample.
Everything here looks good too, so I’ll quickly check the overall heritage.
Taking a look at overall equity told me that OOS is not very different from IS. That means everything is fine. Of course, I could do some strength testing, and so on, but that applies to each trader individually.
It took me less than 7 minutes for the entire procedure – and I was done.
Below are the results AFTER applying the MI condition.
Traction improved by 69%, NP / DD ratio improved by 120%, AVG Trade improved by 65%. What do you want after another 7 minutes of work?
This is another demonstration of the application and power of the Internal Market.