Spready

[SteveHopwood]

Thanks for your efforts fiox.

Would it help if I add adr to Spready? I had been considering it anyway, so it would be no trouble.

[fiox]

Hi Steve. I would have no idea how to incorporate ADR into this trading strategy, and honestly with many other EAs in the past I have had no luck with ADR.

Rather, I am looking for ideas to protect the account from situations like this one attached, where both sides hit all the buy/sell orders, and there is no way to exit except with a really strong trend or a loss. In the example the grid is 20 pips, XAU/USD.

I will try to increase the number of orders and slightly the grid distance, but sooner or later you will reach the limit of funds. Perhaps it would make sense to implement a close-all with lose system, in the unfortunate case that all positions are triggered.

[fiox]

I continued to study the EA's orders in the last few days, and I noticed that in two cases (10 buys / 6 sells once and 10 buys / 5 sells another) the overall take profitt was around 200, instead of 100 as per the settings. All other circumstances (even 10 buys / 4 sells) are respecting the TP of 100. I am using 0.1 lots on XAU/USD, so around 100 pips is the target for $100. Am I missing something?

Taking this opportunity for another report, as expected the crosses with a large ADR seem to work better, at the moment they are all in profit

[SteveHopwood]

I continued to study the EA's orders in the last few days, and I noticed that in two cases (10 buys / 6 sells once and 10 buys / 5 sells another) the overall take profitt was around 200, instead of 100 as per the settings. All other circumstances (even 10 buys / 4 sells) are respecting the TP of 100. I am using 0.1 lots on XAU/USD, so around 100 pips is the target for $100. Am I missing something?

Taking this opportunity for another report, as expected the crosses with a large ADR seem to work better, at the moment they are all in profit

Strange. I will have a look at the code to see if I can find anything.

Courtesy of wolfman, I am starting to fall in love with xauusd and will be posting about this soon.

[fiox]

I continued to study the EA's orders in the last few days, and I noticed that in two cases (10 buys / 6 sells once and 10 buys / 5 sells another) the overall take profitt was around 200, instead of 100 as per the settings. All other circumstances (even 10 buys / 4 sells) are respecting the TP of 100. I am using 0.1 lots on XAU/USD, so around 100 pips is the target for $100. Am I missing something?

Taking this opportunity for another report, as expected the crosses with a large ADR seem to work better, at the moment they are all in profit

Strange. I will have a look at the code to see if I can find anything.

Courtesy of wolfman, I am starting to fall in love with xauusd and will be posting about this soon.

Thanks Steve

Another quick update: after about 1 and a half months Spready reached the feared situation, 10 buys and 10 sells hit at the same time. I will proceed to a close-all and continue the test. At the moment Spready has still produced a clean 2k, also considering this loss

I have already started a demo with a maximum of 20 orders and one with a maximum of 6 orders. The idea continues to be interesting to me, but I am convinced that a stop loss / protection of some kind is needed for these situations

If there is interest at the end of the month I will make a more detailed report of all the demos (there are 7 at the moment) to study their progress

[trader689]

great going fiox, please keep us posted

trader

[SteveHopwood]

V 1b is in post 1. Scroll down to the grid inputs and you will find a new one - MaxFilledTradesOnBothSides. The default is 8.

Spready will close the basket if there are >= 8 market buys and >= 8 market sells. I have no idea where the 'sweet spot' is for this input; I just figure that we are in trouble by the time we reach this point.

Re fiox's issue where the positions hurtled through the basket TP without closing the basket - no idea why this happened. The code must be some of the simplest across the whole of SHF. There is a function (CountOpenTrades()) that builds a 'picture' of the open position.

Straight after the call to COT is the call to the function (lookForBasketClosure()) that compares the cash upl with the profit target and closes the position when the profit target is met/breeched. I can see nothing wrong with the code, so this is a mystery.

[fiox]

V 1b is in post 1. Scroll down to the grid inputs and you will find a new one - MaxFilledTradesOnBothSides. The default is 8.

Spready will close the basket if there are >= 8 market buys and >= 8 market sells. I have no idea where the 'sweet spot' is for this input; I just figure that we are in trouble by the time we reach this point.

Re fiox's issue where the positions hurtled through the basket TP without closing the basket - no idea why this happened. The code must be some of the simplest across the whole of SHF. There is a function (CountOpenTrades()) that builds a 'picture' of the open position.

Straight after the call to COT is the call to the function (lookForBasketClosure()) that compares the cash upl with the profit target and closes the position when the profit target is met/breeched. I can see nothing wrong with the code, so this is a mystery.

thanks Steve, sounds like a good additional feature. I will keep an eye out on the basket TP to see if that behavior recurs

Regarding the profit/danger conditions, in case anyone is interested or can add some ideas, these are some values that I extracted directly from traders in demo accounts, with grid settings 10, distance 20, TP 100$, 0.1 lot, XAU/USD

Note: the pips indicated are considered from the first order level (buy in the example).
e.g. the required price movement to have a basket TP in the case of 3 buy - 0 sell is approximately 55 pips

3 buy - 0 sell > 55 pips
5 buy - 1 sell > 90 pips
7 buy - 2 sell > 130 pips
9 buy - 3 sell > 170 pips
10 buy - 4 sell > 215 pips
10 buy - 5 sell > 305 pips
10 buy - 6 sell > 410 pips
11 buy - 4 sell > 215 pips

In my opinion the situation becomes difficult after 10 - 5, as a trend of 300 pips and above may not arrive. I'm wondering what loss there would be if we decided to close the grid at this point

Also, attaching a screenshot of the movement that generated 10 buys - 10 sells for further analysis

[fiox]

Here we are, at the end of another month. Brief summary of the results I had with 10 different Spready configurations.

Let's start with the bad news, the platform that has been running since the end of September with orders 10 grid 20 has returned almost everything to the market, and at the moment even XAU/USD is in loss. Second bad news, the test with maximum 6 orders produced terrible results, after almost 1 month of trading. Finally, the GBP/AUD and GBP/JPY crosses continue to perform inconsistently, even with different settings, and will likely be replaced.

Let's move on to the good news. The XAU/USD and GER30 (Dax) are profitable on all other platforms that have a greater number of orders (from 15 to 20). The platform with orders 20 and grid 20 is also in profit. The settings that are currently working best are 15 orders and grid 25 or 30 pips, respectively $1500 and $1900 in profit in about 1 month of trading XAU/USD with 0.1 lots.

In particular, the GER30 showed consistent profits across all platforms, even with fewer orders. This is perhaps not surprising given the persistent uptrend, but it is still noteworthy. In case anyone wants to try this EA with GER30, it is worth pointing out that GER30 on AfterPrime is never traded with a spread close to 0, so I had to make a stupid change to the EA, replacing the line of code " if (CloseEnough(spread, 0) ) " with " if (spread > 0) "

Personal conclusions: Spready has potential, as it is based on a probabilistic trend and not on indicators. However, at present it can produce excellent results as well as serious issues to the account. The solution to the problem does not seem trivial to me at the moment, and in any case it is not sufficient to reduce the number of orders, since at that point the EA is not able to produce profit. Increasing the number of orders could improve things, but the problem remains that as the number of orders increases the risk of burning the account increases. I'll do some more testing with the new MaxFilledTradesOnBothSides function

If anyone is interested in the details I can share the myfxbook links

Cheers

[SteveHopwood]

In particular, the GER30 showed consistent profits across all platforms, even with fewer orders. This is perhaps not surprising given the persistent uptrend, but it is still noteworthy. In case anyone wants to try this EA with GER30, it is worth pointing out that GER30 on AfterPrime is never traded with a spread close to 0, so I had to make a stupid change to the EA, replacing the line of code " if (CloseEnough(spread, 0) ) " with " if (spread > 0) "

V 1c is in post 1.

The new input is right at the top of the inputs - MaxAllowableSpread. Spready will send the initial orders if the spread is <= this figure.

If you want to save yourself the faff of downloading again, add this to the top of your inputs:
extern double MaxAllowableSpread=0;


Then go to the code snippet that you have already identified and replace it with this:

Code: Select all

   //Trading
   RefreshRates();
   double spread = Ask - Bid;
   if (CloseEnough(spread, MaxAllowableSpread) || spread < MaxAllowableSpread )
      sendOrders();