In the “End of Day” swing systems testing, I have found some great points or topics for later discussion. The Head Blog will host such reports. Until more data can be collected, it will be premature to draw many conclusions. Therefore, this post is the only bit of information on the systems, plus anything you pick up if you are paying attention to this daily as I go. I post archives of all the data and assure you that nothing is cherry picked and nothing is done in hindsight. I want all the positive and negative data equally because it helps find the proof in the pudding to compare things that worked to things that didn’t. However, it is important to note that real money is not involved, and advice is not being given. I’m simply blogging about my process of determining if strategies work; and the results you see are part of that process.
The process is called quantitative analysis. Here is the google-fu:
noun ~ analysis of a situation or event, especially a financial market, by means of complex mathematical and statistical modeling
I tend to lean toward the statistical modeling side of it all. The complex math comes in for me when it is time to automate reactions and ideas. But, the decisions being made all have to do with the statistics.
Entries and Exits
I am currently testing 4 entry/exit strategy combos. They are all trading the same list of 100 stocks. The only rule is that you take the earliest signal in a direction and no adding to the position with new signals. If you get a second signal in the same direction, it is deemed late and may not trade the name. This will give the earlier signals the reward for that directional call. However, I will take a signal in the opposite direction. Then, you have a hedge zero net sum against the profit and loss, but you still get the statistics of each trade individually and the strength of each strategy may still be weighed against one another.
I have picked 2 trend strategies and 2 anti-trend strategies. Trends tend to continue, so, the first two strategies form the base of the data collection.
- Enter upon the alignment of mean deviation on 3 time frames with room to target at least the standard deviation on each timeframe. It must also be deviating from the 4-hr-EMA(21) in the same direction. This is shown by a score of 7 or -7 in the journal.
- Exit upon score=10 meaning all timeframe targets hit (you target the biggest of them), or stay in trade only as the score stays at zero or above. If timeframes cycle before targets are hit, this can allow for some lengthy swings.
- Value in the Trend
- Enter upon proprietary strength meter equaling a certain level, all shown in the journal‘s ranking pics. This is the most prevalent of the signals.
- Exit upon score=10 meaning all timeframe targets hit (you target the biggest of them), or stay in trade only as the strength stays at 0.5 or above.
- Enter against the trend upon signal that near-term deviation is divergent from far-term deviation.
- Exit upon score=10 meaning all timeframe targets hit (you target the biggest of them), or stay in trade only as the score or strength are an improvement from entrytime. (Purposefully late exits! In later analysis I will discuss improving the model and how to find out where things should improve.)
- Overage (This system was discontinued; I set it up for failure to show some of the things that could have been done to improve the system. Also, I show what a poor trading strategy can do to your bottom and top lines. Perhaps tactics is all that is needed to improve this system, or perhaps time would be best spent elsewhere.)
- Enter against the trend upon the overbought or oversold signals. Any overage indicator will work, I have chosen 3.3% or more away from the 4-hr-EMA(21). This is arbitrary as it depends on what each trader defines as “overdone.” (You are trying to be first to the countertrend.)
- Exit the next day unless another signal can take it in transfer for the same direction (not an add). If that signal comes, all exit rules belong to the new signal. (Not often, but this is where you can pick up a value trade and have done so a day early.)
The entry methods are methods I have made work several times in several tests over several years. However, the exit methods are not the way I trade. I want the data to collect and stay in a trade. Then, I can go back and test all the ways exiting differently can help or hurt the system. IMHO, winning in trading is all about exit management. ~B3