One of the most important aspects to consider when testing or re-testing input signals on a CTA is that they have a different “edge” of the technique they use, for the time they are trading (short-term, swing, long-term, etc.).
Positive price action is when the market moves in the direction of trade. In other words, it is good when the market rises when you buy and bad when the market falls. It is good when the market goes down when you sell short, and it is bad when the market moves up against you. You also need to consider these situations, which are detrimental to trading when you buy and the price falls first, then reverse and move above your entry price and move higher.
In trade, a move in a bad direction is called a “maximum negative excursion” and a move in a good direction is called a “maximum favorable excursion” (MFE). You can use these 2 components to directly measure the “edge” of an input signal.
If a certain input signal produces a move that is higher than the average – maximum good move, medium – maximum bad move, it indicates that there is a positive edge. Conversely, if the maximum good action is higher than the maximum good action, it indicates a negative edge. This is not necessarily a bad thing, because you can use this “negative edge” input signal to conduct “reverse” trading – (Average Return Strategies).
Random access would be when MAE and MFE are approximately the same. For example, if you toss a coin, you would expect that the heads representing buyers and sellers would equal MAE to MFE after using this type of input method.
It takes a few more steps to turn this into a rigid measurement method for input signals. First, you need to have a way to identify price movements in different markets, and second, you need a way to determine the time period when you want to measure average – MFE and average MAE.
To organize MFEs and MAEs in different markets, CTAs should be able to compare averages and equalize them using the Average True Range or ATR. It is useful to compare the price behavior of a particular entry signal using different time frames to isolate the market activity of inputs in different markets. Use the following formula:
1. Calculate MFE and MAE for a specified period.
2. Divide each (MFE and MAE) by the Average Real Range (ATR) to adjust according to the variability at the time of entry.
3. Summarize each of these values separately and then divide by “total signals” to obtain “medium-wave-regulated MFE and MAE”.
4. The ratio is the division of the MFE regulated by the average volatility into the MAE regulated by the average volatility.
To determine the time frame used, use the #day you used in the ratio description, and indicate the number of days the MFE and MAE component were calculated. For example, R10 – ratio size, calculates MFE and MAE for 10 days, including the day of entry, R50 uses 50 days, and so on.
This ratio is used by the CTA to measure whether the input signal has a reliable edge. For example, if they did a random (coin-flip entry) test, they would probably look at the following results; R5- ratio 1.01, R10- ratio 1.005 and R50- ratio 0.997. These numbers are very close to 1.0, and if we tested more, the numbers would be closer to 1.0. This is because after entering a trade based on random access, the price will be in the same direction as it was.
Give an example of this using the Donchian Trend System. The entry rules for this system should only be “bought” when the price exceeds the highest level of the previous 20 days, and sold as soon as the price is below the lowest level of the previous 20 days. The results are as follows. For this example, the R5-ratio was 0.99 and the R10-ratio was 1.0. You think the R ratio should be greater with a positive edge in your input signal. This is true, but what you need to keep in mind is that the Donchian breaking system is the next medium and long-term trend system, so its entry should be an edge in the time frame, not a short one. For entry, the R70 ratio is 1.20, which means that buying and selling in the direction of a 20-day break means an average of 20 percent deviation from the opposite direction when looking at the price movement at 70. days after the input signal. This ratio definitely changes on different days, and this is one of the reasons why trade violations are psychologically difficult.
If you follow or own your own login method, you should take the time needed to research what “extraneous” your login system is in the markets during the period you are trading based on the above. If you do, I think you will be amazed at some of the results you can find. If you need help calculating or researching different login strategies that might be right for you, send me an email or don’t hesitate to call me directly.