Please Help with coding a Trading System

Hi Everyone! I have been trading commodity futures and stocks for a while now and after blowing up accounts few times using widely available strategies over the internet, I decided to backtest strategies before commiting any capital for leveraged trading.

I have recently manually backtested a strategy on XAUUSD that looks great. The system is based on previous week high and low on daily close. The system can be traded without the need for EA however I would like to backtest on multiple asset before injecting capital for trade.

The backtest on XAUUSD starts from April 2006 to end of year 2019. The backtest took me significant amount of time so i guess a backtester/EA should significantly reduce the time it takes to see the trading result on different asset class.

Here is the screenshot of the year end profit and loss on XAUUSD on the strategy applied. Please note that the profit or loss amount is calculated like if Gold is currently trading at 1570 and our buy entry price was 1560 and we close the trade then that is profit of $10. I am not considering the pip value for the trades. Please note that the year 2007 and 2018 are in loss.

If a backtester can be developed, It will be highly appreciated and useful for everyone. The trade rules are quite simple and i will post the trade rules as soon as possible with examples. Thank a lot.

P.S. Since this is my first thread, please consider my error if i have posted in wrong heading.

The system uses high and lows of a candle and a modified fibonacchi tool which looks like this.

I take weekly highs and lows and plot the 0% of fib tool on low of the week and 100% on high of the last week. For the upcoming week direction biasness is bullish if the last week close is above the 50% level and bearish below 50%. The system is quite similar to pivot trading system if someone relates to it. Basically i am taking high and low only while pivot takes close data as well.

Now, if i take last week data for Gold and plot the fib tool the upcoming week i am bullish on gold. But my entry price is at $1564.67 + $1 only, no less not higher.

The stop loss situation is on the change in trade biasness i.e. when price closes below 38.2% level on daily timeframe. The zone between 38.2% and 61.8% is neutral sentiment to account some room for error.

Three separate entries of same lot size is opened using 10% of capital at the entry price. Take profit levels are 1st 100%, 2nd 150% and 3rd 200%.

I have modified the fib tool so i dont have to keep moving the tool for short positions so the TP levels become 1st 0%, 2nd -50% and 3rd -100%.

This is the system for when the asset trends.

Here is an example of winning trade.

Plotted the high/low of the week starting 25th november and ending 29th november 2019 with high and low of 1466.46 and 1450.18.

The buy limit order would be at 1460.21 + 1 which was triggered on 2nd December 2019.

Next day TP1 and TP2 was triggered at December 3rd and TP3 was triggered on December 4th.

Here is an example of a losing trade.

Plotted the high/low of the week starting 28th October 2019 and ending 1st november 2019.

The buy limit order would be at 1502.86 + 1 which was triggered on 5th November 2019 and the close of the day was well below 38.2% level where the position would be closed at loss.

The system catches good profits when the asset trends and loses when there is trend reversal or when in range. The RR ratio stands at minimum 1:1 to max 1:13 which depends upon close of the day.

Only have to monitor the trade once a day at close to see if we have to take stop loss. The entry order are limit orders with TP levels set at start of the week or after the stoploss taken where the trade biasness has changed and the new biasness side limit orders are entered.

Yoges,

Nice idea, thinking outside of the box. Would this idea work on a daily basis. From a very quick look, it seems yes, but further testing needs to be done. The modified Fib indicator your using is more of a guide than the traditional indicator such as MACD or RSI. The daily idea is just a thought I had after reading your thread. I wish you the best on your idea.

The system incorporates reversal signal as well but price should warrant enough reversal move for the reversal trade to be valid. If price reverses from 200% to 150% or from 150% to 100% level then we have an entry.

Sell Stop orders will be sitting at 150% or 100% levels after price hits the 200% or 150%. Loss taken if price fails to close beyond the entry level. TP levels at consecutive 50% price intervals.

Here is an example.

Plotted the high/low of the week starting 30th december 2019 and ending 3rd Jan 2020. Price hits 200% level on 8 jan and reverses below 150%.

Sell stop orders at 150% would be triggered and TP levels of 100%, 50% and 0% would be set until hit or biasness changed in upcoming week.

Thank you acetrader, this idea works on my backtest on gold from 2006 to 2019 with losses on 2 out of 14 years on end of year basis. However, i cannot show you live trading results as it is in concept phase only. This backtest took me almost a month taking 5 hours a day and 6 days a week. If someone could help automate a backtester. I would like to backtest it on other assets as well. Therefore i am asking for help from someone who can code here on FF.

The modified fib tool is just for ease of use. If it is flipped from high to low on normal one, same levels are to be found. The system is based on the levels the fibs provide for entry and exits. No lagging indicators needed for the system like the macd or rsi or moving average. However anyone is free to use them in conjunction to filter trades for backtesting purpose.

Yoges,
After reviewing several pairs on the 15M time frame it appears your method would give good results on the daily basis. I did my reviewing using the MT4 provided Fib indicator & that seems to provide what your looking for (the 50%) level, but a custom made indicator for this would be even better if it would update at the end of each week. Of course there are losing days with this as there are with every method. Doubt I would trade this as I have my own established method, but always willing to look at what others are doing. Hopefully someone will come along to code the indicator your requesting. Wishing you the best.

_______________________________________________

Yoges,

In the words of the late great Steve Job "One Other Thing." Just a thought, but hookup a 200 EMA @ Median Price (HL/2) to your charts as maybe a filter of sorts. Just a thought & nothing more. I happened to have one running on one of the pairs when I tested your method. Seems to confirm or not the direction.

Acetrader, I doubt that there is a system without losses and if you already have a system that gives you positive returns on day to day basis or year on year basis whichever is your evaluation criteria, that is great.

During the backtest i found that this one will give losses during trend reversals and while the asset in in ranging mode. I am still tweaking the system manually and will continue to do so even if i do not find an automated way. I have already put the system in forward test phase to find the results and issues that can come along.

Also, i did use a filter of 50 EMA and the result was not good so dropped the idea. I do have few other ideas still to be tested but the time it takes to test the system has been the hard part.

The main idea of the system is closing prices determines the upcoming direction. So, entries will be on retracement within the range and exits on sentiment shift for losses and range extension for profits.

Do share if you find anything interesting.

Yoges, I will continue to look into this as my time allows. Your right it takes time to evaluate trading methods. I observed on the 15M time frame about a 60 - 65% win rate on two different pairs.

Yes your correct, all methods will incur losses, but its how we minimize those losses that counts.

The 200 EMA does seem to fit well from my observations

The trend reversals are going to have an impact on the results & that is why I looked @ this on the 15M time frame as a way to trade it short term.

I do have a system that is working well for me, but I always like to take a
look at what other traders are doing & try to help if I can. You are clearly on to a method that needs no indicators such as MACD or Stochastic & has the potential to produce many pips for you. You are thinking outside of the box. Best of trades to you!

Acetrader, which timeframe range did you take on the 15M chart? I have looked at it the same timeframe taking 1 hour range ... It works there as well but i do not know about the win rate. Above 30% win rate will be good for the system because the RR is higher. The backtest has 40% win rate in Gold so the result is good. I am looking to tweak it to around 60% before committing live funds.

I chose weekly range in daily timeframe because i do not have to monitor the prices frequently because the system works on limit orders and close basis. Only have to check if the close made a biasness shift where loss is taken i.e. just once a day at opening.

Also, i come from a country where i can trade futures and stocks only and the trading cost in spread and commissions might be one of the highest in the world. So the range has to be in higher timeframe.

I have thought about adding indicator for filter and will test it out most probably with volume/volatility indicator. The problem with adding Moving average was it filtered out losses but filtered the wins as well as the system best profits come from entering early on trend. For now testing out with different stop loss levels to curb the losses.

Yoges,

Here is what I have at this time. Load the GBP/JPY pair on the 15M & scroll back to 12-30-2019. Install a vertical timeline on the 00:00 candle, then load onto the chart the Fibo Retracement tool provided by MT4. Load the Fibo on the date of 12-30-2019 & go forward on each date until you arrive at todays date. Always adding the Fib on each day. Now what do you see? This pair is not the pair I got the 60% from as I closed that chart after I looked at the results. Now there are days that this fails completely & there are days that it works producing many pips, but there are days that it will produce only a few pips, then turn in the opposite direction. On this time frame there are short term pips to be had, but the reversals have got to always to be considered. This is what I see on just a very quick review, but more time is needed to study this.

This appealed to be for two reasons. 1. It could be used short term. 2. Needs nothing more than a Fibonacci tool freely provide by our good friends at MT4.

Unless you require an EA nothing more needed.

This may work better on some pairs than others.

Just a thought, but a High/Low indicator that will recalculate @ a time you choose that will also show the 50% level for the High/Low of the day would help.

This appears to me to be a method that could be used Short Term or Longer Term if some filter is found to addon. Again the 200 EMA comes to mind.

Best of trades to you.
____________________________________________________________________

Ok, I got your point that you are using 1 day range on 15 min timeframe therefore you might be getting different results. The issue will be too many losses when the asset is ranging. I used weekly range on daily candle, so essentially i am taking 5 candles only. If daily range is taken then I will use 4 hour candles. For hourly range i will use 15 min candles.

I have considered reversals for the system so trailing profits will be the previous levels. Say price hits 150% level then i will move my trailing stop to 100% level to lock in the profits.

Also, say in uptrend i have buy side open and price hits my 2nd target or the 150% level, and my order is closed. In other term, I can say that i executed a sell order at that level. It would not be wise to open a new sell order there because i would be calling the level a top.

Now lets assume it was the top and doesn't hit the 3rd target. Price reverses to 100% level and takes out my trailing stop loss. Again here, the orders executed were sell orders. But now i can open new sell position here because the price has reversed enough. If the price in next candle open is above my entry price i will close it and take loss. And the TP levels will be 50%, 0% and -50%.

Furthermore, it seems you have tested only 28 days data range. That is too short time to say the system will work or not work. If you chose a trending asset/pair then you will have good results. If you chose a ranging asset then there will be bad result. Using day range i would at least test 3-4 years, and review the monthly performance of the strategy.

For my backtest, i used week range and tested it since april 2006 and reviewed on year to year basis. That is almost 14 years or 700+ week data range which gave around 2000 closed trades. I perhaps would have tested more but that is the limit of free data tradingview charts give.

This took a lot of time that is why i was looking for someone who could help to code.

I can't trade forex so havent looked at it yet. Still i will have a look at what happens with day's range on 15 m candles. But i guess it need constant monitoring as we need to check every 15 min candle close. Cheers

Yoges, You have summed it up nicely. Yes 28 days is not enough data for a true determination on the method, but those 28 days gave me enough to know on a short term basis there are pips to be made. I use the 15M time frame because I'm a short term trader rarely holding positions overnight, or over the weekend. Another reason for the 28 days is my time is limited at the moment. Will keep looking into this.

Well, we all have our choices how we manage the positions. Short term or long term, the need is a profitable and consistent system. If this system does happen to do both then 5-10 minutes a day is enough to manage the entries and exits for one asset/pair. I just want to see if this works with other assets both positively correlated and negatively correlated ones.

There are lot of things in life beyond watching charts constantly and I want to make time for each and every one of them. Hope you are doing so as well.

Hoping to just employ the capital to do the work. But a lot of research and work needed beforehand. A backtesting EA could have made it a lot easier but seems manual work is what i have to do. I will update the thread if there are new findings.