BigBob: Nanningbob V10 auto-trading robot by SteveHopwood | Page 4 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Nanningbob 4h trading system ver. 5.0 4,432 replies

Nanningbob 60x90 EA 3,920 replies

Require help amending Nanningbob v10 EA 0 replies

Help regarding NanningBob EA 27 replies

Nanningbob 4H trading system 2,017 replies

Attachments: BigBob: Nanningbob V10 auto-trading robot by SteveHopwood
Exit Attachments

BigBob: Nanningbob V10 auto-trading robot by SteveHopwood

  • #61
  • Aug 18, 2011 8:56pm Aug 18, 2011 8:56pm
Quoting SteveHopwood
Glad so many of you are having fun.

There is something brilliant on its way. Last week, El Dee contacted me to code an ea that works like this (imagine a sell trade):
[list][*]take the sell[*]at the start of each subsequent candle, take a new sell if the market has set a new low by x pips. Move the stop of the previous trade to breakeven.[*]at each subsequent new low, take a fresh trade, move the stop of the previous one to breakeven and trades prior to that have their stops moved to the high of the candle one to the right of where their...
Like the sound of this. It's exactly what I had imagined for pyramiding a trend using TrendStat. I shall eagerly await the code!
  • #62
  • Aug 19, 2011 3:01pm Aug 19, 2011 3:01pm
Hello Traders,

first i like to thank Steve for his good job and i hope that i could contribute my knowledge to this thread.

first i will discuss about the volatility. i think that the volatility says nothing about if a pair is trending or ranging. i attached three pictures of different currencypairs. all of them are in Daily Timeframe. but every pair is listed in another volatility group (1,3 & 6).
now i ask everyone, who could name the volatilitgroup of the pairs i attached?

be fair to us & yourself and don't look and compare which is the underlying quote.

i will give the answer later.

Mediator
Attached Image(s) (click to enlarge)
Click to Enlarge Name: pair1.gif Size: 12 KB
Click to Enlarge Name: pair2.gif Size: 10 KB
Click to Enlarge Name: pair3.gif Size: 9 KB
  • #63
  • Aug 19, 2011 3:21pm Aug 19, 2011 3:21pm
Quoting Mediator
Hello Traders,

first i like to thank Steve for his good job and i hope that i could contribute my knowledge to this thread.

first i will discuss about the volatility. i think that the volatility says nothing about if a pair is trending or ranging. i attached three pictures of different currencypairs. all of them are in Daily Timeframe. but every pair is listed in another volatility group (1,3 & 6).
now i ask everyone, who could name the volatilitgroup of the pairs i attached?

be fair to us & yourself and don't look and compare which...
As I understand it, Bob's point about volatility is not whether the more volatile pairs are trending and less volatile are ranging, but whether they are more likely to exhibit their expected behaviours, i.e. that more volatile pairs are more likely to go off on a strong trend and that the less volatile ones more likely to remain within a range.

I could have miss-understood.

On top of this, the assumption is that we accept Bob's definition of a 'trend'. I do not, but am happy to work on this project within his definition and so use terms relevant to it.

  • #64
  • Aug 19, 2011 3:32pm Aug 19, 2011 3:32pm
yes, maybe i'm wrong too. what i read and understood is, that he likely use CTR EA's on the low volatile pairs and made trendtrades on high volatile pairs.

i can understand your point of view as a coder to have defined rules for coding, but on top the rules are Bob's definitions and defined from MA's & PA.
  • #65
  • Aug 19, 2011 3:42pm Aug 19, 2011 3:42pm
Hi Steve,

thank you very much for your EA(s). So far I've learned a lot of MQL4 just by studying your very good readable code.

While backtesting I was wondering why the trend trades didn't close as they're supposed to do.

I found the following which could be the reason:

In the start() function first the TrendDetectionModule() function is called.
So by definition if the low or the high are inside the lines there is no trend. And this is also the definition of a closing trend.

Later on in the OpenTrades section LookForTradeClosure() is called - but only when in trend-mode:

if (trend != ranging) LookForTradeClosure();

But the TrendDetectionModule() before allready triggered the ranging period, because the price allready touched the lines.

So the LookForTradeClosure() would never be called....

Could this be the reason or am I getting this wrong ?

thanks a lot

osthafen
  • #66
  • Aug 19, 2011 3:54pm Aug 19, 2011 3:54pm
Quoting Mediator
yes, maybe i'm wrong too. what i read and understood is, that he likely use CTR EA's on the low volatile pairs and made trendtrades on high volatile pairs.
That is my understanding too, and is how BB is coded to trade.

Quote
i can understand your point of view as a coder to have defined rules for coding,
Yes. These rules may change with use and experience. It is worthwhile remembering how little experience I have of trading this system; BB's code reflects my experience and understanding and may well change with time - in fact, it would be somewhat stunning if not.

  • #67
  • Edited 4:12pm Aug 19, 2011 4:01pm | Edited 4:12pm
Quoting osthafen
Hi Steve,

thank you very much for your EA(s). So far I've learned a lot of MQL4 just by studying your very good readable code.

While backtesting I was wondering why the trend trades didn't close as they're supposed to do.

I found the following which could be the reason:

In the start() function first the TrendDetectionModule() function is called.
So by definition if the low or the high are inside the lines there is no trend. And this is also the definition of a closing trend.

Later on in the OpenTrades section LookForTradeClosure() is called...
Thanks for this. Very good point, and one that I had missed. Cheers.

I figured that it would become necessary at some stage for BB to know whether it had sent a trend or range trade. Unbeknownst to the rest of you, I have added code to a yet-to-be-released version that sends a Global Variable with a new trade. The name of the GV is the chart symbol, and its value is 0 or 1. 0 represents a trend trade; 1 represents a range trade.

I have not released this yet because I could not see a use for this; you have just come up with one. For the coders, it is attached here. Non-coders, ignore it as it does nothing new.

I shall adapt the code in time for market opening after the weekend, and add instructions about setting up missing globals to the user guide.

Attached File(s)
File Type: mq4 BigBob.mq4   90 KB | 400 downloads
  • #68
  • Aug 19, 2011 4:12pm Aug 19, 2011 4:12pm
Going back to the volatility stuff, BB's calculations appear to be reasonable.

Those of you who were with me when I first developed this module a few months ago, will remember my description of GJ as 'psychotically deranged'. 'psychotically deranged' applied to any pair with a D1 pips range > 250.

The Beast landed me with a GJ buy on 8th April that went a long way into the hole, and that I subsequently hedged 100% whilst waiting for the market to decide which way to go to allow me to sort out the position. The damn market has barely moved since. Look at a W1 chart; the blasted market has barely blipped.

GJ's volatility calculation is now down to 159, only 9 points above that of the 'ranging' pairs.

Now, this raises questions about where we put the range/trend boundaries but that is a separate issue. The point of this post is to suggest that the volatility calculations are both revealing and reactive, responding to current market conditions.

I could be wrong - for sure I am no statistician. Nobody has argued with the basic premise yet, so I figure that no argument = agreement.

  • #69
  • Aug 19, 2011 4:25pm Aug 19, 2011 4:25pm
Quoting SteveHopwood
...
I shall adapt the code in time for market opening after the weekend, and add instructions about setting up missing globals to the user guide.
While updating, maybe you could also have a look at the TradeClosure code itself:

I attached a chart with an arrow where the trade should have been closed, but due to the "open"-part it wasn't:
PHP Code
 if (OrderType() == OP_SELL)
   {
      if ((Bid > FastTrendMaVal || Bid > SlowTrendMaVal) && (open < FastTrendMaVal && open < SlowTrendMaVal) ) CloseTrade = true;
   }//if (OrderType() == OP_SELL)
What was the intention for the "open < FastTrendMaVal..." ?

cheers

osthafen
Attached Image (click to enlarge)
Click to Enlarge Name: trend_trade_closing.jpg Size: 177 KB
  • #70
  • Aug 19, 2011 4:36pm Aug 19, 2011 4:36pm
Quoting osthafen
While updating, maybe you could also have a look at the TradeClosure code itself:

I attached a chart with an arrow where the trade should have been closed, but due to the "open"-part it wasn't:
PHP Code
 if (OrderType() == OP_SELL)
   {
      if ((Bid > FastTrendMaVal || Bid > SlowTrendMaVal) && (open < FastTrendMaVal && open < SlowTrendMaVal) ) CloseTrade = true;
   }//if (OrderType() == OP_SELL)
What was the intention for the "open < FastTrendMaVal..." ?

cheers

osthafen
This would be to ensure that range trades are not closed incorrectly. Trend trades by definition would have originated above/below both lines.

This is something else the latest code can help with - this function will only be called when the trade is a trend trade.

I think.

  • #71
  • Aug 19, 2011 4:48pm Aug 19, 2011 4:48pm
Quoting SteveHopwood
This would be to ensure that range trades are not closed incorrectly. Trend trades by definition would have originated above/below both lines.

This is something else the latest code can help with - this function will only be called when the trade is a trend trade.

I think.

I see. So this part could be dropped if we just call it when the trade was a trend-trade (I think because of the daily-MA calculation sometimes the line could "jump" on the other side of the open, also weekend-gaps could lead to an open inside the bars...) .

Do you want to hide every hint the trade could be an EA-trade from the crims ? Otherwise the trade-comment could maybe also be used to mark it as range/trend ?

cheers
  • #72
  • Aug 19, 2011 4:57pm Aug 19, 2011 4:57pm
Quoting SteveHopwood
Going back to the volatility stuff, BB's calculations appear to be reasonable.

Those of you who were with me when I first developed this module a few months ago, will remember my description of GJ as 'psychotically deranged'. 'psychotically deranged' applied to any pair with a D1 pips range > 250.

The Beast landed me with a GJ buy on 8th April that went a long way into the hole, and that I subsequently hedged 100% whilst waiting for the market to decide which way to go to allow me to sort out the position. The damn market has barely moved...
Hi Steve,
You and I have been around this shed for over a year….volatility needs to be based on ATR….Setup on each pair…..14 days/20 days/35 days, your choice….
This is the only way of getting a reading that is honest for each pair….
Please don’t send another missile over here, I’m still repairing from the last one….

Have a great evening

whip
  • #73
  • Aug 19, 2011 5:04pm Aug 19, 2011 5:04pm
Quoting osthafen
I see. So this part could be dropped if we just call it when the trade was a trend-trade (I think because of the daily-MA calculation sometimes the line could "jump" on the other side of the open, also weekend-gaps could lead to an open inside the bars...) .
Thingies to sort out as the bot develops.

Quote
Do you want to hide every hint the trade could be an EA-trade from the crims ?
Yes.

  • #74
  • Aug 19, 2011 5:09pm Aug 19, 2011 5:09pm
Quoting hiredwhip
Hi Steve,
You and I have been around this shed for over a year….volatility needs to be based on ATR….Setup on each pair…..14 days/20 days/35 days, your choice….
This is the only way of getting a reading that is honest for each pair….
Please don’t send another missile over here, I’m still repairing from the last one….

Have a great evening

whip
Thing is, it doesn't matter how many times I Google and read about how atr works, I cannot get my head around the maths, or what atr is actually trying to tell me. I don't understand it, in other words.

I can get my head around the formula I have for calculating volatility based on the daily pips movement of x days.

It seems to me that a pair that has moved an average of 300 pips a day for the last month has been quite volatile, whilst one that has moved 20 pips a day in the last month has gone to sleep. Don't really need an indi I don't understand to tell me this.

  • #75
  • Edited 6:13pm Aug 19, 2011 6:00pm | Edited 6:13pm
@steve & whip

you are both right, it is very important to use volatility for every pair, but not for determining trend or range. the same move at 8th april with gbpjpy happens with another pair but lesser volatile and lesser piprange but same characteristica.

what i mean is, volatility give us the behavior of market for a possible price movementrange. it is true that volatility shrinks in rangingphase if a trend happens before (in terms of normalization over the whole period).

trendtrading in volagroup #1 has the same right to exist like in volagroup #6 but with smaller piprange.

all this vola-stuff is important here because the system uses fixed lots and not risk-based lots determined at a SL-tag. if we would determine the lots with risk-based calulation the vola-calculus would then be unimportant, because 1% loss is 1% loss. i hope you could follow my thoughts.

Mediator
  • #76
  • Aug 19, 2011 6:26pm Aug 19, 2011 6:26pm
Quoting osthafen
While updating, maybe you could also have a look at the TradeClosure code itself:

I attached a chart with an arrow where the trade should have been closed, but due to the "open"-part it wasn't:
PHP Code
 if (OrderType() == OP_SELL)
   {
      if ((Bid > FastTrendMaVal || Bid > SlowTrendMaVal) && (open < FastTrendMaVal && open < SlowTrendMaVal) ) CloseTrade = true;
   }//if (OrderType() == OP_SELL)
What was the intention for the "open < FastTrendMaVal..." ?

cheers

osthafen
Here is the updated code if you want to play with it over the weekend.

Attached File(s)
File Type: mq4 BigBob.mq4   90 KB | 483 downloads
  • #77
  • Aug 19, 2011 6:45pm Aug 19, 2011 6:45pm
Quoting SteveHopwood
Here is the updated code if you want to play with it over the weekend.

Thanks !

I'm allready playing :-) trying to find a better exit for the trend trades. What me annoys is the price just touching the line and then continuing to trend. I'm just experimenting with a veryFastMA instead of the actual price to touch the lines....
  • #78
  • Aug 19, 2011 7:02pm Aug 19, 2011 7:02pm
Quoting osthafen
Thanks !

I'm allready playing :-) trying to find a better exit for the trend trades. What me annoys is the price just touching the line and then continuing to trend. I'm just experimenting with a veryFastMA instead of the actual price to touch the lines....
Go for it, Tiger.
  • #79
  • Aug 19, 2011 7:10pm Aug 19, 2011 7:10pm
Quoting osthafen
Thanks !

I'm allready playing :-) trying to find a better exit for the trend trades. What me annoys is the price just touching the line and then continuing to trend. I'm just experimenting with a veryFastMA instead of the actual price to touch the lines....
If you don't wanna lose a trend move, when price is just touching line,Ea should not close order.Onlu if the bar close is inside the lines...my view,detailed in post #32...

I think we should "teach" the EA to take into consideration gap between MA's...(don't kill me, Steve!)
I don't say
  • #80
  • Aug 19, 2011 7:56pm Aug 19, 2011 7:56pm
Quoting dragosd1
If you don't wanna lose a trend move, when price is just touching line,Ea should not close order.Onlu if the bar close is inside the lines...my view,detailed in post #32...
With veryFastMA I mean just look at the present (not yet closed) candle and
take the Median price, (high+low)/2 with the low being the price touching the line.
In the EA this would be:
veryFastMA=GetMa(PERIOD_D1, 1, 0, MODE_SMA, PRICE_MEDIAN, 0);
And it does filter the touching points without closing too late.
0 traders viewing now