Some testing results

First, here are some sweet pics of the sunset in colorado I took today. Also, accidents from yesterday in Illinois and Nebraska

Fuckin sweet I know ;P

Ok, this was a real pain in the ass to do, but I went back through each trade that was made and implemented the martingale risk profile. Here is one minor detail I came to find out. Over the two month test period from January to February of 2013 the largest string of losses was 13 in a row. There is just no way to double your position size that many times so I had to make a rule. I could only have 5 losses and then I would start my position sizing over again. At my last size I would be trading $16/pip even though with a 10k account I could technically trade for about $40/pip after a string of losses that would need that size. Here is how my position sizing went.

-$10
-$20
-$40
-$80
-$160
-$10
etc...

I did get lucky and have a few strings finally win on the last trade before the reset for a win of +$240. If that had not happened and one more loss had occured, the results would be quite different I think. I will have to save my charts and make copies of them using different data points. That will be quite intensive because that guy who I emailed about updating the robot has basically ignored me. I don't have a clue what's going on with him.

Definitely make sure you download this chart so you can zoom in and see it. this is a comparison of the 289 trades during Jan and Feb of 2013. I wasn't super accurate with my placement of the box, but my goal was to place it at the start of each new daily bar and leave it the entire length of the day.

The top chart is with original static position sizes using $1/pip and the bottom one uses the martingale position sizing as mentioned above.

The area with the largest drawdowns are obviously the string of losses. Click HERE to download an html file of all the original trades from my Metatrader platform. There is also more interesting data in the file at the top. If it doesn't work let me know, but you should just be able to open in right up in your browser.

So... Although this strategy did end up at a 9% profit after 2 months, 289 trades is too small a sample size and too narrow a range of time for me to feel comfortable enough to put any stock in this just yet. I'd love to have at least a thousand trades, or run the numbers on a different position sizing method for the same trades. If I had a robot that could do it all for me it would be so much easier but damn... I am even willing to pay the guy and he just won't respond.

I'll keep you in the loop of any other data I happen to figure out. Until then I suppose I'll be hard at work.

-Francisco

Sunday night fight

heh, I'm just trying to lure people in with a catchy post title did it work?

Anyways, I set the robot up and set the take profit for 1.5 times risk. As I mentioned before the robot doesn't have the ability to adjust risk for a martingale strategy, so I will just have to let it take the trades and figure out the results later. At least things will be more accurate this time with actual entries and exits so there will be no guessing this time.

I have no idea how this will work out, maybe it will go bust when I apply the martingale to it, maybe it will have mixed results. Either way, I am trying to gather important data and I am so happy I have this robot to do it for me. It would be practically impossible for me to manually trade this on the road. This tool really helps me out so very much!

Here is a chart of the robots setup.

It occurred to me the other day that my upcoming post called "Making it all back, the worst idea ever" is basically contradictory to the martingale approach in the first place. It kinda makes me laugh, but contradiction and paradox is life itself. Stay tuned :)

Promised update

During the last time I was out the robot lost 35 times in a row to bring a drawdown of  about 3.7%.

This dramatic string of losses came as a result of my choice to simply not take profits the entire time I was out. It was difficult, as there were plenty of chances to take a good amount of profits, but I stuck it out and the market ended up going down and reaching the profit ratio of 20:1 that I had entered into the robot long ago. I would leave it empty, but the parameters require a number.

Anyways, I said when I got the time I would go back and look at all the profits I could have accumulated during that week out so I took the time today. It was far too difficult for me to look at the charts and feel comfortable saying "I would have taken profit here and here and here." The truth is, that would be what we call "Hindsight Trading" and it's basically bullshit. So, what I decided to do instead is go over a scenario where I left the trading zone exactly where it actually was the whole time and simply take profits that where 1.5 times the risk.

The zone is 10 pips wide, so I would take profit at +15 pips. I also decided to see what it would look like with the martingale risk profile in place, meaning I double my position size (risk) for every loss.

Here is a chart of the all the action.

For you, it will be difficult to see all the entries and exits so you will simply have to trust my data. this chart is a 5min chart, but to get the entries I looked at a 1min chart and the actual account history. Now I did have to estimate some entries that the account simply did not have, but that is because IF the account were actually taking profit at 1.5 times risk, it would have entered a new position and therefor had a new loss in some instances. Below is chart I made with all the data on the right.

As you can see, the drawdown is sharp, but only dipped below the starting equity 3 times. The final result was that instead of losing 3.4% I would have gained 6.7%  $11,432.00

So, with that in mind here is what I want to do to help me acquire more data: Instead of leaving the profit ratio at 20:1 I'll simply set the robot to take profit at 1.5:1 so that I can get realistic results on my entries and exits instead of having to look at charts and guess what they would be. Since the robot does not have the ability to implement the martingale risk on it's own it will stay at one size and then i will have to add that in manually during the week.

It will be difficult to watch the account balance continue to diminish because who wouldn't want to see it go up, but I'll have to accept that while this fake account will more than likely be losing money, I am gathering important data in an effort to be profitable in real life.