CoralMAX v2.1 – New Features – Other – 11 July 2023

INTRODUCTION

In the Trade DDC section there is now an option for Hedging DDC:

Hedging DDC: Use a percentage of the gains of a position in one direction to perform drawdown control on a position in the other direction.

Trade DDC must be enabled for Hedging DDC to be used, as well as setting a value for the Hedging DDC Percent input.  When Max has positions open in both directions (i.e., is hedged), if one of them closes out in profit while the other is in drawdown, then the Hedging DDC logic is applied.  Let’s say that we’re in a EURUSD long and short, and the long position closes out with 0.8% profit while the short position is in drawdown.  If we have Hedging DDC Percent set to 50.0, then we will use 50% of the profit from this trade to perform DDC on the short position.  Max will find the trade that’s a part of the short position that is the farthest from average (i.e., the “worst” trade in the short basket) and will reduce up to the percent amount from the trade we just closed in profit, in this case 0.4%.  

If the short position has only one trade, then it will reduce 0.4% from that trade.  If there are multiple trades in the short position it will reduce up to 0.4% from the farthest trade in drawdown from the average.  If it is less than 0.4% in DD that means the entire trade will be closed.  Otherwise it will close the largest partial lots it can without exceeding 0.4%.

Using Hedging DDC is a great way to use profits that are being gained in one direction to help keep positions in drawdown in range, while ensuring that you are still banking profits.  

RSI AUTO-LEVEL

This is a very experimental feature.  I’m exposing it so others can play with it and see if there’s any utility in this idea, but I do not have strong evidence for or against it.  It’s just an idea.

There are two new inputs in the RSI OPTIONS section for “Auto-Level”:

When the RSI Filter is turned on but Auto-Level is turned off, the RSI Filter works as before.  It will use the RSI High and RSI Low values from the inputs to determine the market conditions for entry.

How do we know what the right settings are for RSI High and RSI Low, though?  In theory we want there to be a high probability that a trade that is entered above/below these levels will, as directly as possible, reverse and hit our trade targets.  We can experiment in the backtester with different values to see if there is some optimum setting, and as we detect changes in results over time we can re-assess these settings.  It’s trial and error, science and intuition.  But I was thinking, what if we could automatically determine the “optimum” high/low values in some automated way?

What I want to know is, at what RSI High value, when price has exceeded that level and reverses, will I be more likely for a trade to reach the trade goal first, before reaching the scaling distance?  Which RSI level has the best probability, over time and on average, of working out without needing to scale in?  That is the condition we are testing for.  RSI High and RSI Low are analyzed independently.  The RSI Auto-Level Strike Rate is the percentage of time that a reversal from that RSI level results in a winner, and we’re looking for an RSI level that achieves that strike rate.

To do this, we test every RSI level starting from the RSI High in the inputs, up to 90.0.  Working backwards, we find a candle on the RSI timeframe (e.g., H1) where RSI exceeded that level, and then on the chart timeframe (e.g., M5) we look forwards from there to see if price reached 0.5 ADR in profit before going 0.5 ADR in drawdown.  We record whether it was a winner, and then continue looking back to the previous RSI high value.  We go back to 100 times that RSI exceeded the high (or the end of available data, whichever comes first), and return the strike rate from that RSI level.  If the strike rate is greater than RSI Auto-Level Strike Rate then that becomes the new RSI High value.  If not, we then go up to the next RSI level (in steps of 2.0 at a time) and test again with this new level.  The result is that we will pick the lowest RSI level above RSI High at which, in recent history, the strike rate was at least  RSI Auto-Level Strike Rate.  And then the process is repeated for RSI Low.

This routine is run when the EA is first started, and then again every new day.  The RSI levels are constantly adjusting to current market conditions. When it sets new values it prints an INFO message to the log; prints a message to the status bar; and sets the current RSI levels to the hover text of the RSI Value cell in the UI, so you can check at any given time what the current values being used are.

What I would expect to see comparing the fixed RSI levels with the RSI auto-level is fewer entries overall (i.e., the RSI level is higher so we reach it less often); positions having a lower duration with less drawdown, and probably a decrease in overall profits and drawdown.  And over time, a lower probability of having to manually manage trades that went against us and needed (aggressive) DDC.

In this test run there were 75 total trades (positions) opened.  The average duration was 3 days, with a maximum of 25 days.  Average DD was 0.63%, maximum DD seen was 6.3%.  On average each position closed out with only a single trade, but on the high-end a maximum of 6 trades was needed.

Compare to the exact same settings, but with RSI Auto-Level turned on:

In this case the total number of entries went down (from 76 to 47).  The maximum duration went down to just 12 days, with a maximum DD of 4.77%, a reduction of around 1.5%.  The maximum number of trades needed was 5, compared to 6 from before. The total profit returned went down as well, from 22.15% to 13.5%.  This was expected, but it’s disappointing that the effect was so large.  However, if this reduces the likelihood of a trading needing some manual intervention, then it could be worth the reduction in gains.  It just depends on your comfort level in dealing with trades in drawdown.  You can experiment with different values for  RSI Auto-Level Strike Rate to find the right balance between strike rate and the returns observed in testing.

This is not a particularly scientific analysis; it is one test, using opening prices only.  But it demonstrates that the feature had the intended impact, to reduce drawdown, duration and number of entries from trades.