Hedge Trading the US Market using 2 symbols

Hi Anthony,

The smallest lot size for US30 with Pepperstone is 0.1 lots, which means I can go 0.1 lot US30 and 0.9 lot US500. I understand a lot of brokers only offer minimum lot size of 1 for US30, which means the minimum lot size for US500 is 9, which has a significantly higher margin requirement.

I would only place these trades at or after US market open (0930 New York Time), since the spread for my broker is at its narrowest during this time, and I assume will be the same for most.

Try in a demo account, you can do the same thing in MT4 you will just need to manually open both positions, and then close them when you're enough in profit that slippage will not turn your trade negative. For me, that would be roughly $10-$20 using a 0.1 lot size, or $100-$200 using a 1 lot size with the US30 symbol. The way you can tell if lot sizes are correct is that one position should turn positive, and one should turn negative by roughly the same amount almost instantly during US market open hours. If one position starts moving much faster than the other, then you know you need to readjust your lot sizes. I believe for most CFD brokers the 9:1 or 10:1 ratio of US500 to US30 is correct though. Not sure how it would go with futures / e-minis etc as I have never traded these.

Hi Mr4x86
I will try in the US session
I hope this strategy is the opportunity to collect green pips
Thanks for share

Thanks much

Hi Mr4x86
Testing on demo today, the result is quite satisfactory.
Still running

Its really GOLDEN EGG - but only for bigger account (i think 10 k is the minimum).
Do you know someone broker which allows fractional lots for this indexes?

I am with Pepperstone in Australia which allows 0.1 lot size of both US30 and US500 with 1:200 leverage. margin requirement is about 20 bucks each and can safely trade with 2k minimum account size.

I went 0.1 short and 0.9 long yesterday, being silly and cautious but market bounced and US30 actually rose slightly more than US500 (which on average doesnt happen) so my trades are still sitting at -$15 from yesterday. I can hold them open swap-free for 10 consecutive days before an admin fee is applied - plenty of time to exit the trade.

If DD increases too much i'll add another position the same, or 1.0 US500 instead of 0.9 buy if the market doesn't tank on open tonight. Sometimes you get positive equity and can close the trades out in minutes or hours, sometimes it can take a week.

I created a second script that auto opens and auto closes trades, and it produces profits even backtesting as far as Jan 2016 with 30% DD using lot sizes as per above on a 2k account.. Now of course its just an estimate as you wouldnt necessarily run exact same lot sizes in all market conditions, but its nice to know you're safe from blowing an account even if it was left on auto-pilot, especially considering the ups and downs the market has had since Jan 2016.

If anybody finds a way to further optimise my code / settings / idea, please feel free to share with the class

Good work - I should have done 10:1 ratio too lol...

One thing I forgot to mention, you may as "why not go long 1 US30 and short 10 US500?". You can try it, last night it probably would have worked, but I backtested that too and the drawdown ends up being much larger. I think in the bigger picture over many years, US500 slightly outperforms the US30 so long US500 is the way to go

A very interesting system. I have tried once Today witb 10:1 ratio and at some time point the profit went up to 40 dolllars. I would love to See an EA for This system so I can make backtest on MT4.

Since I don't use MT4, you'll need to find someone who can do a cAlgo to EA converter. Code below is for the auto trader with variable parameters for optimisation.

cTrader has only in its recent versions allowed backtesting with multiple symbols, so up until that point it was all trial and error forward testing with demo accounts. I've literally spent months on this trying to figure out the best parameters and now with multi-symbol backtesting in cTrader you can literally backtest as far as the broker has been offering that symbol. 10+ years.

//Use ONLY US30 Symbol with this bot
//Recommend using default parameters, only change lot sizes
//SAFE Drawdown lot sizes for US30 with 200:1 leverage per account balance:
//$AUD2000: 0.1
//$AUD10000: 0.5
//$AUD20000: 1.0
//$AUD50000: 2.5
//$AUD100000: 5
//$AUD200000: 10
//$AUD1000000: 50

//In bear markets, use a 9:1 ratio of Buy US500 to Selling US30 Volume
//In bull or consolidation markets, use 10:1 ratio of Buy US500 to Sell US30 Volume

using System.Linq;
using cAlgo.API;
using cAlgo.API.Internals;

namespace cAlgo.Robots
{
[Robot(TimeZone = TimeZones.UTC, AccessRights = AccessRights.None)]
public class SellUS30 : Robot
{
[Parameter("US30 Volume", Group = "Basic Setup", DefaultValue = 0.1, MinValue = 0.1, Step = 0.1)]
public double InitialVolume { get; set; }

[Parameter("US500 Volume", Group = "Basic Setup", DefaultValue = 0.9, MinValue = 0.1, Step = 0.1)]
public double US500Volume { get; set; }

[Parameter("Pip drawdown new position", Group = "Basic Setup", DefaultValue = 110, MinValue = 10)]
public int NewPosition { get; set; }

[Parameter("Target profit for each group of trade", Group = "Basic Setup", DefaultValue = 30, MinValue = 1)]
public int AverageTakeProfit { get; set; }

[Parameter("Max Spread", Group = "Basic Setup", DefaultValue = 2.0)]
public double MaxSpread { get; set; }

public double balance;
private bool cStop = false;
private double sp_d;
private string thiscBotLabel;

protected override void OnStart()
{
// Set position label to cBot name
thiscBotLabel = GetType().Name;
}

protected override void OnTick()
{
sp_d = (Symbol.Ask - Symbol.Bid) / Symbol.PipSize;
if (MaxSpread >= sp_d && !cStop)
Grid();

RCN();

if (Positions.Count(x => x.Label == thiscBotLabel) > 0)
{
if (Positions.Where(x => x.Label == thiscBotLabel).Average(x => x.NetProfit) >= InitialVolume * AverageTakeProfit * Symbol.PipSize)
{
foreach (var position in Positions)
{
if (position.Label == thiscBotLabel)
ClosePosition(position);
}
}
}
}

private void Grid()
{
if (Positions.Count(x => x.Label == thiscBotLabel) == 0)
{
Symbol symbol1 = Symbols.GetSymbol("US30");
Symbol symbol2 = Symbols.GetSymbol("US500");
ExecuteMarketOrder(TradeType.Sell, symbol1, InitialVolume, thiscBotLabel, 0, 0);
ExecuteMarketOrder(TradeType.Buy, symbol2, US500Volume, thiscBotLabel, 0, 0);
}
{
if (Positions.Where(x => x.Label == thiscBotLabel).Average(x => x.NetProfit) <= InitialVolume - (InitialVolume * NewPosition * Symbol.PipSize))
{
Symbol symbol1 = Symbols.GetSymbol("US30");
Symbol symbol2 = Symbols.GetSymbol("US500");
ExecuteMarketOrder(TradeType.Sell, symbol1, InitialVolume, thiscBotLabel, 0, 0);
ExecuteMarketOrder(TradeType.Buy, symbol2, US500Volume, thiscBotLabel, 0, 0);
}
}
}

private void RCN()
{
ChartObjects.DrawText("pan", A_cmt_calc(), StaticPosition.TopLeft, Colors.Tomato);
}

private string A_cmt_calc()
{
string gc_78 = "";

if (sp_d > MaxSpread)
gc_78 = "MAX SPREAD EXCEED";
else
gc_78 = "SPREAD OK";
return (gc_78);
}

protected override void OnStop()
{
// Close all open positions opened by this cBot on stop during backtesting
if (IsBacktesting)
{
foreach (var position in Positions)
{
if (position.Label == thiscBotLabel)
ClosePosition(position);
}
}
}
}
}

In our country, there is one person who trades in a similar way - but currencies (instead of indexes) based on a closed circuit between currencies according to their strength. Unfortunately, I do not know this system exactly. It uses its own indicator to assess the output.
---------------------
Automatic translate:
FX - ie a closed circuit of currencies with the consequent formation of synthetic pairs (euro, dollar can be bought or exchanged in a bank or exchange office, but it is difficult to buy a currency pair EUR / USD, GBP / USD or GBP / JPY). The selection of sub pairs for SOK inputs is governed by what has been mentioned several times earlier - mutual correlation on currencies and pairs, volatility, price range comparison, amplitude and pip value of the pair. Each time you enter (see figure), a minus state occurs, depending on the amount of the spread and the number of lots that are input in the game. Subsequent market movements create a certain "monetary" sinusoid (even this can be calculated plus / minus bus), which oscillates irregularly between profit and loss. E.g. we start with a start position of -150 euro and the oscillatory volatility of SOK is between +/- 300 euro. The sinusoid itself thus oscillates approximately between +150 euro gain and -450 euro loss, depending on the degree of correlation and volatility. If the sine wave reaches a certain profit threshold (above level zero), all positions are closed at once and the process is repeated. It depends on the correct selection of pairs, how many individual "rounds" in a certain time period.

Similarly to FX, where there is a big advantage of the stationary currency circuit, the similar principle of SOK can be applied also to stock markets using probability, statistical arbitrage and internal correlations, for example by selecting certain titles from the most traded indices - DJI, NQ-100, SP500 but also from FTSE, DAX etc. However, in this case the stationary circuit must be replaced by a moving circuit whose range is given by the value of the index, which is formed by the shares themselves.
By: SID

I have to rectify the terminology a bit from the end of the last paper, where the oscillatory sinusoid of SOK is written. As a colleague from the TFG chat pointed out to me correctly, it is more of an oscillatory curve that oscillates between zero, resp. between gain and loss of set SOK. Thanks to the skill of the same colleague and the creation of his SOK indicator, it is possible to indicate quite precisely the course of this curve and with the help of another indication even set a preset profit limit at which all positions contained in the SOK automatically close at the same time. On the example of the curve of the SOK indicator at the end of Friday's session March 12, 2010 with preset levels +200 euro and -400 euro, which are marked with colored crosses. So next time I will no longer use the name oscillatory sinusoid, but the oscillation curve.

By: SID
-----------------
PS.:I think choosing the couple you did is crucial to profitable trading.

I would like to know more about this system and how entry trades are decided within FX. I have tried for over a year to come up with something that works with correlated currencies as I love correlations but eventually the correlation changes and you're s**t out of luck.

The reason this one works so well is that US30 and US500 correlation holds near perfect 100% on every time frame. You can bring NAS and RUS into this but their positions are hard to size, they also don't open and close at the exact same time as each other in the way the US30 and US500 do - potentially leaving you exposed during a market gap for a few hours.

The other symbol I have tried this with is oil - XBRUSD and XTIUSD. A few FX brokers pay swaps for going long XBR and short XTI, and theyre closely correlated too but not correlated enough. the spread between the two prices can go from $3 up to $20.

If anybody has any research on correlated pairs that is profitable, I'd be interested. I would even attempt to make a cAlgo bot to go with it.

I think even somewhat-correlated pairs will have their place in trading again, when central banks start jacking interest rates up again so that we can at least make money on positive swaps.

Still early days but i see a potential here.

I tried the same a few years ago, but it was WTI vs Brent oil.
My impression is this: it has potential, it takes a big account and averaging down or SL because there will be long, long periods drawdown.
Good luck.

Yep I tried the same, and I had positions open over the weekend of the Saudi facilities bombing. Come Monday morning, WTI had exploded higher 2 hours before XBR opening, and guess which instrument I had short? Scary times

Last night I set my cBot to auto-trade while I went to sleep, completely missing the 50bps rate cute. Woke up this morning and my $2k account is up by $100 - 5% is not bad. I made a bot similar to the above but with buying US30 and selling US500. Both made money overnight. I used a ratio of Buying 9 US500 to selling 1 US30 in one bot and Selling 9 US500 to buying 1 US30 in the other, just so the margin was as equal as I could get in both long and short positions. I'll see how it goes tonight

Any update on this?

Hi Peluq,

This week the US500 fell slightly more than the US30, so as such I am slightly in drawdown by roughly 5% for the week. This is keeping in line with backtesting of my automated script that suggests a maximum of roughly 20% drawdown over the past year.

Screenshot supplied for proof with my live trading account. Expecting equity to turn positive in the coming days. I am using 9 lots of US500 for every 1 lot of US30 just FYI. Will post updates later on this week

Cheers

The problem may be that the 10 days (swap-free) are not enough to make a profit. With the EA you can determine how long the drawdown was in days?

Hi Peluq,

If you do a backtest on any of the algo's, cTrader 3.7 and above allows backtesting with multiple symbols so its quite easy to see how long each position is held. Every now and then a position is held longer than 10 days but generally not more than 20. The profits on all the other closed trades make up for the occasional admin fee that must be paid. Alternatively you can close positions out manually on the 9th day. After some experimentation I have come to a few conclusions:

1) It is best not to trade while the market is still crashing. A few times this past couple of weeks the market has been limit-up and limit-down where my broker has stopped allowing buy positions of US30 but allowed them on US30, or simply just stopped the ability to trade them altogether. Once the circuit breaker was lifted, my script saw the positions were net positive but the market reacted so quickly and violently then by the time the positions were closed I made a small loss.

2) The second script that I created that behaves as somewhat of a grid (adds positions when drawdown increases) will bring faster profits and handles these turbulent times better. I'm attaching this cBot file now with the best tuned default settings for Pepperstone at this point in time. I am providing the code as open source in case anybody can build upon it and improve it, then share it with the rest of the class. I cannot post cAlgo files as attachments so you will have to copy / paste and build your own cAlgo with it. If you have used cTrader its not hard to do, otherwise google it.

Best settings for below are as follows:

Markets stable: Use default of 10:1 ratio of US500:US30
Markets volatile / trending down: Use 9:1 ratio of US500:US30

//Use ONLY US30 Symbol with this bot, best to use h1 timeframe
//Recommend using default parameters, only change lot sizes
//SAFE Drawdown lot sizes for US30 with 200:1 leverage per account balance:
//$AUD2000: 0.1
//$AUD10000: 0.5
//$AUD20000: 1.0
//$AUD50000: 2.5
//$AUD100000: 5
//$AUD200000: 10
//$AUD1000000: 50

//In bear markets, use a 9:1 ratio of Buy US500 to Selling US30 Volume
//In bull or consolidation markets, use 10:1 ratio of Buy US500 to Sell US30 Volume

using System.Linq;
using cAlgo.API;
using cAlgo.API.Internals;

namespace cAlgo.Robots
{
[Robot(TimeZone = TimeZones.UTC, AccessRights = AccessRights.None)]
public class SellUS30 : Robot
{
[Parameter("Sell US30 Volume", Group = "Basic Setup", DefaultValue = 0.1, MinValue = 0.1, Step = 0.1)]
public double InitialVolume { get; set; }

[Parameter("Buy US500 Volume", Group = "Basic Setup", DefaultValue = 0.9, MinValue = 0.1, Step = 0.1)]
public double US500Volume { get; set; }

[Parameter("Pip drawdown new position", Group = "Basic Setup", DefaultValue = 110, MinValue = 10)]
public int NewPosition { get; set; }

[Parameter("Target profit for each group of trade", Group = "Basic Setup", DefaultValue = 30, MinValue = 1)]
public int AverageTakeProfit { get; set; }

[Parameter("Max Spread", Group = "Basic Setup", DefaultValue = 5.0)]
public double MaxSpread { get; set; }

public double balance;
private bool cStop = false;
private double sp_d;
private string thiscBotLabel;

protected override void OnStart()
{
// Set position label to cBot name
thiscBotLabel = GetType().Name;
}

protected override void OnTick()
{
sp_d = (Symbol.Ask - Symbol.Bid) / Symbol.PipSize;
if (MaxSpread >= sp_d && !cStop)
Grid();
RCN();
}

private void Grid()
{
if (Positions.Count(x => x.Label == thiscBotLabel) == 0)
{
Symbol symbol1 = Symbols.GetSymbol("US30");
Symbol symbol2 = Symbols.GetSymbol("US500");
ExecuteMarketOrder(TradeType.Sell, symbol1, InitialVolume, thiscBotLabel, 0, 0);
ExecuteMarketOrder(TradeType.Buy, symbol2, US500Volume, thiscBotLabel, 0, 0);
}
}

private void RCN()
{
ChartObjects.DrawText("pan", A_cmt_calc(), StaticPosition.TopLeft, Colors.Tomato);
}

private string A_cmt_calc()
{
string gc_78 = "";

if (sp_d > MaxSpread)
gc_78 = "MAX SPREAD EXCEED";
else
gc_78 = "SPREAD OK";
return (gc_78);
}

protected override void OnBar()
{
sp_d = (Symbol.Ask - Symbol.Bid) / Symbol.PipSize;
if (MaxSpread >= sp_d && !cStop)
if (Positions.Where(x => x.Label == thiscBotLabel).Average(x => x.NetProfit) <= InitialVolume - (InitialVolume * NewPosition * Symbol.PipSize))
{
Symbol symbol1 = Symbols.GetSymbol("US30");
Symbol symbol2 = Symbols.GetSymbol("US500");
ExecuteMarketOrder(TradeType.Sell, symbol1, InitialVolume, thiscBotLabel, 0, 0);
ExecuteMarketOrder(TradeType.Buy, symbol2, US500Volume, thiscBotLabel, 0, 0);
}
if (Positions.Count(x => x.Label == thiscBotLabel) > 0)
{
if (Positions.Where(x => x.Label == thiscBotLabel).Average(x => x.NetProfit) >= InitialVolume * AverageTakeProfit * Symbol.PipSize)
{
foreach (var position in Positions)
{
if (position.Label == thiscBotLabel)
ClosePosition(position);
}
}
}
}

protected override void OnStop()
{
// Close all open positions opened by this cBot on stop during backtesting
if (IsBacktesting)
{
foreach (var position in Positions)
{
if (position.Label == thiscBotLabel)
ClosePosition(position);
}
}
}
}
}

Wouldn't the market rebalanced or become uncorrelated
At some point. Both move in the same directions. Only by rebalancing lot size to hope the correlated move in your way.

Hi Jensen,

If you check here https://www.myfxbook.com/forex-marke...ion/US500-US30 you can see the two indices stay near perfect correlation across all time frames. The only thing that changes is the value of each lot sizes over time as each market goes up or down. This is why the system works better if you "grid" it a little bit by adding another hedged pair if the pair you're carrying falls out of correlation by x amount of pips, because they will swing back the other way at some point, its just a matter of time. Unlike pairs such as EURUSD / USDCHF which can become uncorrelated for a very long time, the US30 and US500 swing backwards and forth within a tight range.

The main issue at the moment is with limits up and down (mostly down) in recent days, theres a good chance brokers will stop trades being placed. Or even worse in my case, still allowed you to place trades short on the S&P500 but not on the US30 in one instance, making you no longer hedged when the algo places the only trade its allowed. My advice is to possibly wait until volatility and also spreads have calmed down in the market before activating this bot, but once done it is safe to go for an 8:1 ratio if the market is looking overall bearish then switch back to 10:1 once its bullish again.

For me I'm not bothering right now because US30 spread during US market open is roughly 12 which is downright robbery considering it is normally 1.6 - almost a factor of 10. Which means you have to overcome a silly spread before you can break even, never mind profit