Slopey Peaky Bob

Priyesh · Post by **Priyesh** » Mon Sep 17, 2018 8:37 pm

Hi Luke
Autotrading definitely enabled.
Priyesh

Priyesh · Post by **Priyesh** » Mon Sep 17, 2018 9:05 pm

Looks like the problem has been resolved. I simply removed the 1v EA and recompiled. All is well again.

Thanks
Priyesh

Post by **billv** » Tue Sep 18, 2018 4:45 am

Priyesh » Tue Sep 18, 2018 7:37 am wrote:Hi Luke
Autotrading definitely enabled.
Priyesh

Is your chart history up to date?
Perhaps the EA is busy downloading historical data?

Does the EA open new charts when you click on a symbol?
If it doesn't, perhaps it has crashed.
Save your EA settings
remove the EA
Restart Empty4
Apply the SPB EA to the chart
Load your saved settings

Priyesh · Post by **Priyesh** » Tue Sep 18, 2018 5:47 am

billv » Tue Sep 18, 2018 4:45 am wrote:
Is your chart history up to date?
Perhaps the EA is busy downloading historical data?

Does the EA open new charts when you click on a symbol?
If it doesn't, perhaps it has crashed.
Save your EA settings
remove the EA
Restart Empty4
Apply the SPB EA to the chart
Load your saved settings

Hi BillyV
Thanks very much, I had followed the steps as per your response and all is well.
Kind Regards
Priyesh

Post by **dclayw** » Tue Sep 18, 2018 7:10 am

SteveHopwood » Wed Sep 05, 2018 8:41 pm wrote:Those of you trying the Experimental version and Recovery already know that it is not working.

I returned to the code and found lots that was wrong. Here is a fully commented copy of the updated code. Would you codeslingers run your eyes over it, please, and see if you can spot anything obviously wrong? THe function is called just after the call to CountOpenTrades(symbol, PairIndex), so the bot has a picture of the open position.

Not too much activity in this thread. Just wondering if anyone has managed to have a look at the recovery code because I think it's definitely needed. I'm not really a codeslinger so can't add much value there.

I've continued tweaking my test setup during this drawdown period. During my extensive testing on native SS a while back I found that some pairs just did not seem suited to this indicator. Pairs with higher ATR seem more suited, and when you think about it this makes some sense doesn't it? Excluding pairs is sought of curve fitting in a way but that's what I've been doing. That together with "normalising" the ATR based distance between trades in the grid.

Most of you will probably think this is poppycock but for what it's worth I'll explain further. Steve uses DistanceBetweenTrades of 30 pips, seems as good as anything else. I'm using an ATR based grid. What I did was choose a GridAtrDivisor so that the average of each pair's ATR/GridAtrDivisor came out to as close as possible to 30. In this case GridAtrDivisor came out to be 2.55. I then exclude any pair where ATR/GridAtrDivisor < 30.

There's no concrete scientific principle behind this, it's just a means of excluding lower ATR pairs using Steve's arbitrary 30 pips as a guide. Anyway right now this basically excludes pairs with ATR < about 75.

What I did then is "normalised" these cutdown list of pairs, i.e. choose a new GridAtrDivisor so that the average of each cutdown pair's ATR/GridAtrDivisor came out to as close as possible to 30. In this case GridAtrDivisor came out to be 3.5.

In the whole scheme of things it probably does not matter whether you use hard coded DistanceBetweenTrades or one based on the ATR, but I can't help thinking the ATR approach is better. Difficult to know without months and months of testing but even then you would never know because there are too many other variables coming into play.

One more thing, during my native SS testing I felt that using SS H4 was the better way to go because the drawdowns with D1 were unmanageable. So my SS settings are H4 0.8,14,50. With these SS settings it makes sense to drop Peaky timeframe down to 30 minutes so you're looking at a smaller area of the chart.

I'm currently running some tests with and without doing that second normalisation I mentioned above and also with and without using ImmediateMarketOrders. The results are Ok but only 1 1/2 weeks in so meaningless really, but I hopefully heading down the right track. Do need code recovery though because it's just a matter of time before the account gets blown. I do use CloseMarketTradesOnOppositeColour so as to have at least some form of protection. Higher ATR pairs are also generally highly correlated.

SilverTrader · Post by **SilverTrader** » Wed Sep 19, 2018 11:55 am

Its a little way down the road now since SPB had it's July hiccup, but it has been left alone and is slowly getting back from its DD figure of just over -£1300 at one stage. Hopefully in time, it makes a full recovery.

Recovery.PNG

snip · Post by **snip** » Thu Sep 20, 2018 10:51 am

Hi, first time I have posted..

I have been using this EA and the experimental version for about 6 weeks now, with Gareths set file slighty modified. I have a problem with both EA's where I get the message 'WE HAVE REACHED OUR WEEKLY PROFIT TARGET. I HAVE STOPPED TRADING UNTIL NEXT WEEK'. I have tried to stop this from happening by turning off(making false) the input 'UseTradingDoneForTheWeek', but with no effect. I have read the manual and looked at the other inputs, but cannot find anything that will stop this happening. Can anyone offer any help please.?

Cheers Snip

Post by **SteveHopwood** » Fri Sep 21, 2018 3:44 pm

Slopey Peaky Bob v1w is in post 1. I had a look at the code in response to snip's post above and cannot see anything wrong with the code apart from one possibility. You do not need the update if you have not experienced the problem.

DIYers, go to the top of the Autotrading() function. The first block of code checks for trading done for the week. I have added this to the bottom of the block:
else
DoneForTheWeek = false;

so the entire block looks like this:

Code: Select all

   //Trading done for the week stuff.
   //This allows DoneForTheWeek to reset to 'false' if the bot was left running over the weekend.
   if (GlobalVariableCheck(DoneForTheWeekGV))//Is there a GV?
   {
      //Extract the time.
      datetime WeekStartTime = (datetime) GlobalVariableGet(DoneForTheWeekGV);

      //Is it a previous week?
      if (WeekStartTime < iTime(Symbol(), PERIOD_W1, 0) )
      {
         DoneForTheWeek = false;//Reset the bool
         GlobalVariableDel(DoneForTheWeekGV);//Delete the GV
      }//if (WeekStartTime < iTime(Symbol(), PERIOD_W1, 0) )
      
   }//if (GlobalVariableCheck(DoneForTheWeekGV))
      else
         DoneForTheWeek = false;

-----------------------------------------

Experimental version 1x is in post 1, inspired by how I am managing those trades that went deep into DD during the summer, using triple SuperSlope. Along came Thomas' TDesk and the whole task instantly became 100% easier.

I am lucky in that the original trades are attracting positive swap apart from one pair so there is not the usual pain of negative swap - in fact the swap is excellent on some:

TDesk: triple SS TD D1, H4, H1
Leave the trades alone when SS lines up. Some are in mighty DD but SS is showing the right direction. NJ was at c. -$400 at one point and entered positive territory yesterday.
Hedge when there is a full SS signal in the opposite direction.
- All three values must be >= 1 (buy hedge) or <= -1 (sell hedge).
- Close the hedge trade when the H4 and H1 SS values have changed back again with the same minimum values as used to open the hedge.

This has worked well so far. Hedge trade losers have not been massive and I have made some cash out of them. Current DD is half of what it was at its height. Hedging has stopped my margin percent from dropping below 1,000.

SPB has no hedging; the point I am making is that a triple SS setup is immensely powerful when all three line up, so I have added this to SPB - a highest, medium and lowest time frame. Traders can use any combination of them.

I have also added the ability to turn Peaky off in case anybody fancies basket trading just with SS.

This was not easy coding and there are bound to be bugs, so still not on your live accounts yet folks.

snip · Post by **snip** » Fri Sep 21, 2018 4:51 pm

Hi Steve

Thanks for the update. Unfortunately it will not compile because it's missing a header file include\TDesk.mqh. I can only assume when I originally downloaded the EA I must have downloaded a ex4 file so compilation was not required. I cannot remember compiling anything 6 weeks ago, but there again I must admit my memory these days is not as it was..

I did notice you have a 'My shell EA code' section, where I guess the header file resides. Do I need to install any of this stuff to be able to compile successfully.?

Sorry to be a bother..

Cheers Pete

Post by **SteveHopwood** » Fri Sep 21, 2018 4:54 pm

snip » Fri Sep 21, 2018 4:51 pm wrote:Hi Steve

Thanks for the update. Unfortunately it will not compile because it's missing a header file include\TDesk.mqh. I can only assume when I originally downloaded the EA I must have downloaded a ex4 file so compilation was not required. I cannot remember compiling anything 6 weeks ago, but there again I must admit my memory these days is not as it was..

I did notice you have a 'My shell EA code' section, where I guess the header file resides. Do I need to install any of this stuff to be able to compile successfully.?

Sorry to be a bother..

Cheers Pete

Ye Gods, how the blue blazes can you have missed this? http://www.stevehopwoodforex.com/phpBB3 ... 106&t=5522

Get a grip.