Backtesting
The process of backtesting has been a long and difficult one. For the purpose of reflecting to this at a later time, I decided to write it down a bit. My setups aren't perfect, nor do I believe they are going to work forever without tweaking or adjusting. Looking at the DAX, I have to take in account that the average daily range is a lot smaller than 6 months or longer ago. If markets can change over months, they can certainly change more over years.
The data I checked were about 200 intraday charts. I was looking for a setup that allowed me to take 3-5 trades a day. I realize this objective was not based on experience but on a personal subjective need willing to make an arbitrary number of trades each day. Some people argue that you should focus only on one market at a time, but I've found it illuminating to compare different markets. I've actually learned a lot from looking to the NQ where imo price moves where much cleaner and volume had more significance than on the DAX. A lot of people will probably argue with me on this, so let's just stick to the fact that I've been able to "read" the signs better at NQ. Anyway, I happened to find some setups that all together give me 3 to 5 opportunities each day,...
I won't ignore nor deny the fact that I've learned a lot from others and for developing a plan the steps described in the Trading Journals thread (posts #5 and #8) were quite helpful. One of the most important "insights" I've acquired is I've let go of the idea wanting or needing to know why price is doing this or that. I just needed to find a way to profit of it.
For backtesting I took batches of about 100 trades. I wrote almost everything I could see about the situation in the chart: what happened before, what happens after, what kind of volatility, around what price level, the volume it showed, etc... leaving indicators totally out of the equation. I looked at the results and compared them to another batch, and another one, and yet another one. After a while you get to see things, you find that price isn't moving totally random and there are ways to profit from. Ofcourse, all this isn't found in backtesting, but backtesting is proving (or disproving) certain hypotheses I had. The next step was in looking where you could find the best entry, find a suitable stop size and a possible target...
After forward testing it for a while, I found it actually did hold a lot of possibilities. I just needed to clear my mind of all the distractions and letting go of the need to understand why some trades were losers. Some trades are just bound to be, it's part of the probability a system generates.
I'm at the point where I'm still looking for small refinements but am confident the system shows a positive expectancy and hence will automatically lead to an increased equity over time.
Currently, I am implementing the setups in real time using one contract but trading only one contract is difficult. You don't have the luxury of taking a less-then-perfect entry and you can't let your profits run to its maximum (no scaling entries nor exits). Still, the results show potential and basically I know that if it works with one contract, the system will be more flexible using two.
The DAX has been moving around about to 100 points at some days, but the last couple of weeks the daily range has been quite small (30-50 points) but with a 25EUR per point, I started to focus on NQ which is only 20$ per point but shows less relative volatility. The road is long and winding, but I believe in the outcome and am confident I've chosen the right direction.