Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

Trading the Sixths with Snowballs 68 replies

Gann High-Low System with Laguerre/Fisher/VSA/Sixths and others 124 replies

My Sixths trading journal with TOB EA 50 replies

sixths indicator 1 reply

Steve Trading Journal 3 replies

Attachments: Sixths trading - an EA by macman, Bob and Steve
Exit Attachments

Sixths trading - an EA by macman, Bob and Steve

Quoting astral77
Quite right, I agree totally. As you know how it works in UK is you fill in a questionnaire given to you by the court and you put all your income and living expenses on that sheet, whatever is left (i.e. income-living costs = disposable income) that is your weekly payment. I would say it would be around 10 Euros a week for the next Billion weeks!!! All the credit to French legal system.

Kind Regards
Well, he wont have any income for the next 3 years and I doubt he will get another job working for any bank or financial institution.

Although, maybe he could sell his story to hollywood for a film.

Scoobs.
Quoting macman
You can count on my vote Steve -
Over to asjeff for a photo of a gorgeous young thing with my head superimposed on top.
Ok, this post has no point whatsoever. Over in the newbs thread, murphy man pointed out that I am at nearly 5,000 posts.

So, this is post no 5,000.

Three and a half years ago, when I joined FF, I never even began to dream I might actually have something to offer.

So, 5,000 posts later and I am still waffling drivel to International standards.

Set up an International Drivel-wittering team to take on The Rest Of The Universe, and I will be the automatic first choice as captain, coach, selector and manager.

Now for a thought to make you all slash your wrists in despair:

Here is to the next 5,000.
Back to business.

I have managed a couple of loosing backtests today. It took an effort, but clearly is possible.

One of the characteristics of these backtests was that there were long strings of trades taken out that were filled just before the market resumed its march against them. So, there is a new input: MaxTradesAllowed. TB will only send new pending trades when the total of pending and filled trades is < MaxTradesAllowed. This will allow there to be MaxTradesAllowed + 1 open, so the default of 5 will allow there to be 6 open. I leave you to work out how for yourselves.

To facilitate clarity in back/forward testing, I have removed the pending trade deletion at the start of each new candle. Now, they are only removed if the sixths lines change. Once we are sure of our future direction, this feature is easily re-introduced.

Scoobs, the BarCount auto-calculate function appears to be less accurate than macman's hard-coded estimates. For now, I am leaving the code as it is but am open to change in the future. Many thanks for offering up the code.

Quoting duffypratt
So I take it you haven't looked into the rhythmic similarities between Wagner's Leibestod (the Liszt transcription) and Elton John's Honky Cat?
I play the transcription. Never come across the EJ. Not consciously anyhow, and am highly unlikely to take any notice if I did.

Is anybody testing on FXCM UK? I've setup a demo account with FXCM UK. Added latest EA and Sixths indicator and editing lot size to minimum .10 required by FXCM. No other changes. I'm not seeing any trades being taken. It's been running since noon Eastern time. I did see a few cancelled pending orders in the History for the A/U, but not other trades. Is this EA setup for ECN trading?
Quoting newera
Is anybody testing on FXCM UK? I've setup a demo account with FXCM UK. Added latest EA and Sixths indicator and editing lot size to minimum .10 required by FXCM. No other changes. I'm not seeing any trades being taken. It's been running since noon Eastern time. I did see a few cancelled pending orders in the History for the A/U, but not other trades. Is this EA setup for ECN trading?
Newera, Im running on Gomarkets and trades have been entered. They require ECN compatability in EA's. I have updated a few versions so it may be a version issue along the way

rgds

Peter
Hi Steve

not sure, but I think the trade management part isn't working.
So I set up a separate chart again with MPTM on it.
The rest works perfectly though!

Others have this as well?
"The only successful substitute for brains is silence."
Quoting DXTrade
Hi Steve

not sure, but I think the trade management part isn't working.
So I set up a separate chart again with MPTM on it.
The rest works perfectly though!

Others have this as well?
I can't tell yet whether the trade management is working or not. I'm running a demo with about 10 pairs. All of them are in drawdown to one extent or another and as far as I can tell, none of them have ever been profitable since I put on the new version of the EA this afternoon. For the day, I'm down $500 dollars on a $10,000 demo.

Maybe in the morning I will have a clearer idea. (Or maybe the demo account will blow up
  • Post #50
  • Quote
  • Edited at 4:03am Oct 6, 2010 3:45am | Edited at 4:03am
Quoting SteveHopwood
Back to business.

I have managed a couple of loosing backtests today. It took an effort, but clearly is possible.

One of the characteristics of these backtests was that there were long strings of trades taken out that were filled just before the market resumed its march against them. So, there is a new input: MaxTradesAllowed. TB will only send new pending trades when the total of pending and filled trades is < MaxTradesAllowed. This will allow there to be MaxTradesAllowed + 1 open, so the default of 5 will allow there to be 6 open. I leave you...
Hi Steve,

Hmmm .... I can see the idea behind limiting trades, but IMHO it may be counter productive.

There is a post here http://www.forexfactory.com/showpost...2&postcount=79 by a guy who has done a nice backtest which shows a clump of trades on GU - as he says "nasty period from 5/20/2009-6/4/2009".

But in real life that is a 2 week + period, so had you been trading live you would not just sit and watch a position going against you for weeks and do nothing about it. We all know it is impossible to build a set & forget profitable bot - the major banks can't with all their funds, so we have no chance. It is essential to accept that we have to intervene to help the basically stupid bot.

So what are the intervention options? :

1. Adjust MPTM so that any slight profit is grabbed (this is why limiting trades may be counter productive)

2. Adjust the look back period of the Sixth's indi to cover a much smaller range

3. Turn off Primary and/or Secondary trades

4. Adjust the Primary / Secondary lot sizes

5. A considered combination of any of the above ...

Any ideas, comments always welcome. Between us all, we should be able to produce something robust and fairly easy to manage ... and profitable
  • Post #51
  • Quote
  • Edited at 4:57am Oct 6, 2010 4:22am | Edited at 4:57am
Hi Steve,

I have been testing TB a lot over the past 24 hours and particularly on GU H1 and keep coming up with that issue of May 09 that Greg was talking about over on the newbies thread. That type of prolonged run is going to wipe this system out. If you look at H4 over the same period the account survives.

I think the problem is that TB over trades. You added the PipsBufferFromTradeLine to make sure we didn't just touch the line which has cut down the trades somewhat.

However what is still occurring is that if the sixths lines move ever so fractionally and we retrace through again a new trade is entered despite it only being a few pips from the last. If then we have Start_Recovery_at_trades=3 then most parcels of trades end up in recovery and breaking even with those in May09 never getting a chance to break even and therefore destroying the account balance.

So firstly I think we need to limit the number of trades and then secondly redefine what recovery is.

I would suggest that what we really want is one trade only occurring on the retrace back through the gold and or the green waiting for either of those to close in profit before entering a new trade.

However if the sixths did move significantly (and then we are obviously in DD) then I would see that we would want to enter an additional trade on the retrace. Therefore I think a new variable MinPipsFromLastEntry makes sense, allowing the trades before recovery to be used at very significant levels rather than all 3 around the same price point just because we are dipping backwards and forwards across a sixths line. I think this would cause many less trades to enter recovery just because the limit on trades is breached despite the market not having really moved.

Regarding recovery (as far as I understand) at the moment as soon as our Start_Recovery_at_trades is breached the T/P is moved to the average entry price and then TB continues to enter trades as normal (up to the limit imposed by MaxTradesAllowed) each time adjusting the T/P to the new average entry price.

My first suggestion here would be add another new variable RecoveryBreakEvenPips that enables us to modify the breakeven level by x pips in our favour to cover fee's brokerage etc etc. Many of these recovery trades are not in massive drawdown when occurring but rather have just used up the limit on the number of trades allowed therefore grabbing an extra 5-10 pips each time makes sense.

My second suggestion on recovery is that given this systems core is based around Bobs retracement type trading why then are we not implementing a true Bob style 2.4.2 or 1.1.3.3 recovery? i.e. as soon as Start_Recovery_at_trades is breached then the T/P is moved to the average entry price (+ RecoveryBreakEvenPips from above) but then no new trades are entered until the market moves a further x pips against us (in the case of GU I think Bob suggests 250?) from the breakeven line at which time the first recovery trade will be taken which depending on the system being used will be a certain multiple of the sum of ALL the current open positions (not just 1 x our lot size as unlike in Bobs system we may already have 2 or three trades open before the recovery kicks in).

I have attached a screen grab below that hopefully illustrates all these points.

This is the start of the May09 GU problem. Here you can see sells taken in this order

1. 1.47637
2. 1.47788
3. 1.48648

Recovery kicks in

4. 1.48672
5. 1.4902

Now MaxTrades is breached and the portfolio is on its way to a 1000 pip DD.

I think it is fairly clear that we really wouldn't have wanted to enter trade 1 & 2 (just one would be fine) and also neither 3 & 4.

So if we had implemented MinPipsFromLastEntry=30 for example then we would have the following

1. 1.47637
2. 1.48648
3. 1.4902

Therefore when recovery kicked in our breakeven price would have been 1.48435 rather than the earlier 1.4802. Which actually would have been hit a few bars later when the market moved down to a low 1.4826.

But that was probably just luck so let's presume we didn't get that spike down and instead remaining in recovery.

In the first example (the current system) you can see we have used up all our trades and our breakeven rate is 1.48355....we were in the 1.60's before this was ever touched again.

However if we followed true Bob style 1.1.3.3 then we would be entering 3 x Lot Size (we have 3 open positions) at breakeven + 250'ish (the exact level would be determined by another retracement back through sixths. i.e. after market is 250 pips higher recovery allows another trade based on original trade entry logic but this time is same size as total open position) so in this example roughly 1.5093 bringing our new breakeven to 1.4968 which was once again breached on the 7th May giving us a second chance to get out. If once again this spike lower was just luck then if you run the same chart in your system you will see Level 3 recovery further down the track would have also resulted in getting us out of the position.

Whenever I test a system I come from trying to prevent blowing up first and then seeing what effect those changes have on the remaining already profitable areas. I think in this case these changes have shown that on 3 occasions we would have been able to get out of the DD that would have shut down the account otherwise ....they can't all just be luck, while at the same time I would imagine possibly having a beneficial effect on the system as a whole through reducing the number of trades going into recovery prematurely through over trading.

Anyway I have been rambling on long enough and perhaps this is not the direction you wanted to take TB but I do think you have something worth pursuing here. I hope these comments provoke some thought.

Jono
Attached Image (click to enlarge)
Click to Enlarge Name: example186.jpg Size: 82 KB
Quoting jm2110
[color=black][font=&quot]However if the sixths did move significantly (and then we are obviously in DD) then I would see that we would want to enter an additional trade on the retrace. Therefore I think a new variable MinPipsFromLastEntry makes sense, allowing the trades before recovery to be used at very significant levels rather than all 3 around the same price point just because we are dipping backwards and forwards across a sixths line. I think this would cause many less trades to enter recovery just because...
Your brilliant contribution highlights neatly why I make my work freely available her at FF - doing so attracts so many fabulous minds.

I have adopted your MinPipsFromLastEntry thingy immediately; this is in post 1.

One of the loosing backtests I managed to run yesterday was UC from 5th Jan to 28th June of this year (Alpari UK demo). Adopting the MinPipsFromLastEntry has turned this into a winning streak, making an overall difference of several thousand pounds on a 10K starting balance and trading 0.3/0.1, so a clear success here.

Note to coders: I have added a new function to the bot - bool IsTradingAllowed(). This returns true or false and is called at the start of void CalculateTrades(). In turn, void CalculateTrades() aborts if the call to IsTradingAllowed() returns false. This means you can quickly and easily add and test any new filter that takes your fance. Have fun.

Re Recovery, you come up with great ideas here also Jono.

One point I did not make originally in post 1 is that, whatever the Start_Recovery_at_trades input, Recovery only starts if the combined upl of all the pairs' trades is negative. Merely having x trades open does not necessarily trigger Recovery.

Quote
My second suggestion on recovery is that given this systems core is based around Bobs retracement type trading why then are we not implementing a true Bob style 2.4.2 or 1.1.3.3 recovery?
Only because macman's system has a form of Recovery built in, especially now that we have the MinPipsFromLastEntry thingy. No reason why we should not try to enhance this so once Recovery kicks in, I am thinking this:

 

  1. use the pips range supplied by Max in the NB thread as the pips dd required to trigger another trade
  2. only send primaries
  3. only send them under normal trading conditions for the system i.e. pending trades at the gold line

Let's leave enhancing the breakeven to take a few more pips until we have the basics in place. In the meantime, we have moving the breakeven line manually to adjust the tp.

Incidentally, the manual line-moving thingy is not working properly. I have some live UC trades in dd at the moment and when I moved the line, TB adjusted the tp's, then moved the line back and re-adjusted them.

The only way to persuade TB to leave the line alone was to disable Experts, move the line and re=enable Experts; this appears to have worked.

I have to go out soon and will be out for most of the day. If any of the ace coder-proofreaders can see why the code doesn't quite work, it would save me some pain somewhere down the line.

Thanks again Jono. Wonderful contribution.

Quoting SteveHopwood
use the pips range supplied by Max in the NB thread as the pips dd required to trigger another trade
Yes, but please don't use the hard coded old version
this one is better :
Inserted Code
      int multi;
      DailyATR = iATR(Symbol(), PERIOD_D1, 14, 0) ; // in 4 digits
      if(Digits == 5 ) multi = 100000;
      if(Digits == 4 ) multi = 10000;
      if(Digits == 3 ) multi = 1000;
      if(Digits == 2 ) multi = 100;
      DailyATR            = DailyATR*multi;
      Deviationfast  =iStdDev(NULL,0,15,0,MODE_EMA,PRICE_CLOSE,0);
      Deviationfast  =Deviationfast*1.5;
      Deviationslow  =iStdDev(NULL,0,40,0,MODE_EMA,PRICE_CLOSE,0);
      if (Deviationfast>Deviationslow) Deviation=Deviationfast;
      else Deviation=Deviationslow;
      //if (Deviation<DailyATR) Deviation=DailyATR;
      if(Digits == 3 ||Digits == 2) Deviation*=100;
      if(Digits == 5 ||Digits == 4) Deviation*=10000;
      //Deviation   =NormalizeDouble(Deviation  ,0);
      ReEntryLinePips    =NormalizeDouble(Deviation*ReEntryPips_DeviationMult  ,0);
      ReEntryLinePips    *= multiplier;
      bool CalmMarket=false;
      if (ReEntryLinePips <DailyATR)  {ReEntryLinePips =DailyATR; CalmMarket=true;}     
 
      // adapt for TF
      if (Period()==PERIOD_M15)
      {  ReEntryLinePips   =NormalizeDouble(0.70*ReEntryLinePips  ,0);
               }
      if (Period()==PERIOD_H1)
      {  ReEntryLinePips   =NormalizeDouble(0.85*ReEntryLinePips  ,0);
              }
Quoting maxou888
Yes, but please don't use the hard coded old version
this one is better :
[code]...
Cheers max. You are a star.

I arrived home to find the bot still happily churning out trades within a few pips of one another.

There was a flaw in the IsTradingAllowed() logic which I have attempted to correct. I am still not sure this is correct. As I so often say, any clown can do the coding. It is working out what to code that is the clever bit.

Whatever, I found a bug that was causing pending sent during Recovery to have their take profit adjusted to the current breakeven. I have squished this. Fix in post 1.

Onwards and, as ever, round and round in circles.

I have had another go at the MinPipsFromLastEntry thingy. Still not sure it actually works.

Post 1 for the odd passing individual still vaguely interested and has the foggiest idea what I am talking about.
Quoting SteveHopwood
I have had another go at the MinPipsFromLastEntry thingy. Still not sure it actually works.

Post 1 for the odd passing individual still vaguely interested and has the foggiest idea what I am talking about.


I'm still interested and do know what u r talking about. I too have several trades open about 3-10 pips apart. I'm no millionaire trader but I don't think this is the approach we want to take. :-). The minpipsapart idea is a must. Too bad I'm on iPhone and can't change Ea till I get to work in 2 days. Should have put it on vps I guess

Steve between your eas and bob v6.0 my brain is tired and weary and no my eyes go easily out of focus. hopefully it will pay off soon!!
Steve,

Fantastic work on "The Beast" looks like it has real potential. Thanks for sharing this with us.

I have noticed during forward testing that the mptm that is built into the EA is setting the SL to BE at +40 but it does not seem to be adding BE profit of +30. As iIundeerstand this should be happening and then jumping stop set to +30. Not sure if I have something set incorrectly but I have double checked everything.

I'm no coder so I could be wrong!

Thanks again.
Quoting indy360
Steve,

Fantastic work on "The Beast" looks like it has real potential. Thanks for sharing this with us.

I have noticed during forward testing that the mptm that is built into the EA is setting the SL to BE at +40 but it does not seem to be adding BE profit of +30. As iIundeerstand this should be happening and then jumping stop set to +30. Not sure if I have something set incorrectly but I have double checked everything.

I'm no coder so I could be wrong!

Thanks again.
That could be because I uploaded TB with the default mptm settings wrong. Look at the inputs and see if adjusting them sorts out the problem.

I am just knocking together a quick ea for a FF member, then I will start looking afresh at the Recovery thingy.

Thanks Stev,

I have checked and my mtpm settings are:

BreakEven: True
BreakEvenPips: 40
BreakEvenProfit: 30

JumpingStop: True
JumpingStopPips: 30
AddBEP: False
JumpAfterBreakEvenOnly: True

I had a GU trade open yesterday that went +47, the SL was moved to BE but the BE profit of +30 did not work, it has since retraced and hit SL at BE.

Where am I going wrong??

Cheers.
0 traders viewing now