Correlated Currency Divergence Strategy

Hello Traders,

I'm new on Forex Factory.

I have been "studying" trading for several years, and I'm still trying to find a strategy that suits me in order to become a more serious trader.

Recently I come up with an idea of an hypothetical strategy that I want to share with you, to have an opinion from more experienced traders, and eventually in case of some of you will help me to crack the information that I'm missing in order to make it real.

Basically Correlated Pairs (Gold/Silver) (EURUSD/GBPUSD) (AUDUSD/NZDUSD) (S&P/Nas100) (DAX30/CAC40) (USDCAD/OIL) (BITCOIN/ETHEREUM) move like DNA strands: they will cross each other, and then they will move away one from another like an elastic, to move together again later, and this sequence is repeated over time.

How can we take advantage of this sequence?

When they are far away from each other, at the maximum extension of the elastic, we can Sell the one above and Buy the one below.
No matter what, they will later cross each other again, and we will close both the positions at the crossing.
At the closing we will have 3 scenario:

1. If the cross occurs between the Buy and the Sell, both trades are in profit.
2. If price closes above the Sell, the Buy will be in profit and Sell in loss, with the profit of the Buy bigger than the loss of the Sell.
3. If price closes below the Sell, the Sell will be in profit and the Buy in loss, with the profit of the Sell bigger than the loss of the Buy.

So, no matter where the price closes, we will have a NET position of the 2 positions in profit.

Basically it is like an arbitrage on the divergence of correlated pairs.

A drawdown may occur after we place the trades, if the elastic enlarges even more, so we will have for some time a drawdown. In this case we may also think to add a position, (still minding the money management), as no matter what, soon or later, the 2 pairs will cross each other again.

It can be used on any time frame. I expect more trades on smaller time frame with small profit (smaller pips) each, and I expect fewer trades on bigger time frame, but with bigger profit (bigger pips) each.


What will we need?

1)
We will need an Overlay Indicator which puts the 2 pairs on the same chart. This Overlay Indicator for MT4 I found it for free on the web. But Overlay indicator come with the problem of the scale. There is also the "compare" option in TradingView.


2)
This point 2 is the most difficult part for me to solve. We will need an indicator which indicates something like the strength of the divergence/elastic, which will help us to filter and to place the 2 trades at their optimal/maximum extension, in order not to get too much drawdown.

For example, if this indicator has a scale of 0-100, we may consider placing the trades everytime the indicator goes above 80. You are more experienced, so my question is:

DO YOU HAVE ANY IDEA HOW THIS DIVERGENCE STRENGHT INDICATOR CAN BE ACHIEVED? WHICH FORMULA?

THIS POINT 2) IS THE MOST DIFFICULT CHALLENGE OF THIS STRATEGY WHICH NEEDS TO BE HACKED TO MAKE THE STRATEGY REAL.

If this point is solved we will need just an EA to make the dirty job and catch trade opportunities at day and night.


3)
We will also need a calculator which give us the equivalent position sizing of the trades to be placed for the 2 pairs. I make an example to try to explain me better:

Let’s say there is an Arbitrage Divergence on Gold / Silver and I want to Sell Gold and Buy Silver.

If I sell 1 Lot of Gold, how many lots of Silver should I buy? Which is the formula to sell/buy the same amount on 2 different pairs?



If you kept the time to read my post I would love to hear your thoughts. But most importantly, do you believe it can be done?

Please, I don't have your same years of experience so please avoid to criticize if this strategy eventually will be an unuseful strategy. I'm just doing my best. All contructive critics are accepted, expecially I will need to figure out that Point 2 of the Strategy.


Thanks in advance for answering,

Hi, interesting, can you attach the overlay indicator please.

Thanks for the ideas sharing here, and hope we could gather ideas from FFers for improvement of this strategy.

I did trade with Correlated pairs before but I have encountered one problem that would

Did you forward test or trade real account with correlated pairs of this strategy before? It is because Overlay Chart indicator repains badly that visual backtest is not trustable. It maybe what you said "the problem of scale" in your point 1. The exit at crossed pair can sometime be misled when backtest.

for your point 2, shall we set a distance in % between two pairs? e.g. High and Low of one pair for a certain number of candles vs High and Low of another pair for a certain number of candles


I am an exprert of EA developer that made a hundred of EAs using online tool.

Hello Pollylin!

Thank you for your post.

Not yet. I have to find a mechanical way when to place trades and closing them. But I may have found one. Keep reading below.

It's not a repaint in strict sense of repainting, where repainting indicator first they give signal and then they change it.

Have you in mind when you scroll orizontally a chart in metatrader that candles became longer or smaller according how the chart evolve? That's a scale problem, and it is present in the Overlay Indicatori as well. So, I agree with you, it cannot be used to find the right moment of entry, or exit.

Yesterday I came up with an idea. To use the Standard Deviation 1.5 on the Ratio.

Ratio: EURUSD divided by GBPUSD

Trades are triggered when Ratio close above 1.5 Standard Deviation, or when Ratio close below -1.5 Standard Deviation.

Trades are closed when Ratio cross the Ratio AVG.

So we don't use the Overlay indicator to find entries/exit, but simply the historical data of EURUSD and GBPUSD with SD.

So today I backtested the strategy on a short time range: 1 Jan - 31 Jan 2022.

The result is positive, so even if we cannot assume that the strategy will work always on all correlated pair, I can start to be a little optimistic.

Unfortunately doing such back test manually with excel it's extremely time consuming. It took me almost 2 hours to complete 1 month worth backtest. But given the positive outcome, I'm more than motivated to do backtest till 30 Jun for EURUSD/GBPUSD, and also to backtest all other correlated pairs.

FYI I did the backtest using 1 lot for all trades with a stoploss at 50€. Stop loss was triggered only 5/6 times.

Speaking about Lot Size, in my original post above, I asked a question. Which is the formula to allow me to know how many lots of Silver should I trade to match 1 lot of Gold?

I'll appreciate a lot if you could answer.

As you can read in my other post in this thread the Overlay Indicator has a scaling problem.

Strategy can be more clear if you use Second Chart Indicator instead.

Second Chart Indicator will place the second pair on the same chart, but on a different window.

Please see both of them attached.

You might find this thread interesting:
https://www.forexfactory.com/thread/...-similar-stuff

Similar, if not identical, to the idea that you're presenting...the problem is that correlations tend to go through cycles of de-coupling which translates to deep drawdowns if you don't have a sound money management program in place.

https://www.forexfactory.com/thread/...ation-strategy

Best,

Dave

Thank you. I'll go through it.

Hi Orion,
do you mean something like this?

It is basically just EURUSD und GBPUSD on one chart.
To make it comparable, I plotted both in relative returns.

You can read more about rel. returns here:
https://www.investopedia.com/terms/r/relativereturn.asp

'Speaking about Lot Size, in my original post above, I asked a question. Which is the formula to allow me to know how many lots of Silver should I trade to match 1 lot of Gold?'
Here you can read more about on how to calculate this:
https://www.litefinance.com/blog/for...-lot-on-forex/
https://www.investopedia.com/terms/p/pip.asp

Since we want to buy/sell if distance between a & b is higher > average, here is a chart showing distance (a-b) vs average of distance.

Distance > average
EU > GU

Means we sell EU and buy GU

Yes, 7bit's work on this strategy is the best I've seen anywhere. Hmmm...7bit disappeared not too long after he finished up the thread. Maybe he hit on the right formula and retired from public life Old Dog followed up with some excellent work on this subject also, especially regarding cointegration and its applications.

Regarding lot size weighting...you'll need to normalize the pairs by comparing pip value and volatility for sure and probably a few other things that I can't remember.

Yes, I think relative returns works too. Probably applying standard deviation as well. For now I'm experimenting with the Ratio (i.e EURUSD/GBPUSD with 1.5 Standard Deviations and I'm having profitable backtest.

Uhm... I read these links. But I don't find the explanation of weighting 2 Pairs.

I mean, If I have 1 lot of gold, I know how much € moves if goes in profit of X pips.

What I don't know is how much lot of Silver I must open to move the same amount of € for same number of X pips.

Speaking about "normilaze" I did this reasoning:

I make an example.

Let's say Gold move 1% profit, and Silver move 1% profit.

The formula must give the output lot size, where the the Profit on Gold is the same of the Profit on Silver, because both moved 1%.

Let's make a numerical example.

Gold quote 1700.00$
Silver quote 18.00$

Gold moves 1% to 1717.00$
Silver moves 1% = to 18.18$

If we buy 1 lot of Gold the profit is 170$

How many lot of Silver we should buy to have 170$ on Silver from the 1% move?

Does this reasoning make sense?

https://www.mataf.net/en/forex/tools/volatility
https://www.mataf.net/en/forex/tools/pip-value

These two links at Mataf will give you the pip value of each currency pair and the average hourly/daily range for all of the primary pairs.

Once you normalize the pip values (so that they are roughly equal) you can measure the ADR of the pairs that you are interested in. Example...typically the GBPUSD is more volatile than the EURUSD (130 ADR to 103 ADR respectively over the last 10 weeks) This means the lot size of the GBPUSD must be lower than the EURUSD so that the GBPUSD doesn't run off and ruin your model because it is more volatile than the EURUSD.

This is a great tool for correlated pairs trading:. https://www.forexfactory.com/thread/...rading-advisor

What you mean by the standard deviation here