Pippo: a Murrey Maths retrace auto-trading EA by SteveHopwood | Page 6 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Pippo: A Murrey Maths Retrace EA 188 replies

Maths (ratio) problem 16 replies

Do i need to be a maths genius? 12 replies

How important is a knowledge of maths? 96 replies

Stupid maths 19 replies

Attachments: Pippo: a Murrey Maths retrace auto-trading EA by SteveHopwood
Exit Attachments

Pippo: a Murrey Maths retrace auto-trading EA by SteveHopwood

  • #101
  • Jun 16, 2011 1:33pm Jun 16, 2011 1:33pm
Steve.

Another small anomaly I found is the following:

On the Bands on the chart, if you hover over the line with the curser, you get the Band name such as mml4 and mml5 etc. Looking at the code, you call the price value of these lines to use in your if statements.

The anomaly is the following:

If you hover over a SOLID line such as mml4 then hover over the next SOLID line up (not the dashed line), it shows as mml5. If you hover over the dashed line between mml4 and mml5, you get mml5.5. Shouldn't that be mml4.5. It is the same for the others as well. It goes mml5, mml6.5, mml6.

Looking at your code, it seems you have adopted the code to work around this anomaly so probably no big deal but I wanted to bring it to your attention in case it could lead to other, unforeseen problems in the future.

Dave R.
  • #102
  • Jun 16, 2011 1:47pm Jun 16, 2011 1:47pm
Quoting droland
Looking at your code, it seems you have adopted the code to work around this anomaly so probably no big deal but I wanted to bring it to your attention in case it could lead to other, unforeseen problems in the future.

Dave R.
The code that generates the MM x.5 lines is mine, so I understand it and how it works.

  • #103
  • Jun 16, 2011 2:15pm Jun 16, 2011 2:15pm
Steve. Another small observation.

When the market is very confined and the Bands are close together, the EA is set up for example to use Bid on closing the short trade. For Example:

if (OrderSelect(TradeTicketNo[3], SELECT_BY_TICKET) > -1 && OrderType() == OP_SELL && OrderProfit() > 0)
{
target = ObjectGet("mml11.5", OBJPROP_PRICE1);
if (Bid <= target) return(true);

As you can see in the attached picture, when the bid is <= target, it closes the short trade but the close price is the Ask price of a short position, not the bid price.

In this case, we lose the spread. As you can see, my closes are above the dashed line in the attached picture. No big deal when the Bands are far apart but more significant when the Bands are tight together.

Dave R.
Attached Image (click to enlarge)
Click to Enlarge Name: pippo xauusd.gif Size: 43 KB
  • #104
  • Jun 16, 2011 2:34pm Jun 16, 2011 2:34pm
Quoting Pippopotamus
Another triumph for mediocrity! Congrats. Really you hope sway all to accept the BOT in its' current state. You are doing a great job at it. Hope you haven't engineered any bridges, or anything else around NYC. Leaving little flaws, unsure of their ultimate effect, as you are wont to do, could lead to disaster. Well, I am sick to death of this Confederacy Of Dunces and leave you all to your FOLLY. Steve, please remove all reference to me in any way, shape, or form from the intro to this thread. I insist. Most especially, change the name of the BOT...This...
You are kidding, right?
  • #105
  • Jun 16, 2011 5:23pm Jun 16, 2011 5:23pm
Quoting droland
Steve. Another small observation.

When the market is very confined and the Bands are close together, the EA is set up for example to use Bid on closing the short trade. For Example:

if (OrderSelect(TradeTicketNo[3], SELECT_BY_TICKET) > -1 && OrderType() == OP_SELL && OrderProfit() > 0)
{
target = ObjectGet("mml11.5", OBJPROP_PRICE1);
if (Bid <= target) return(true);

As you can see in the attached picture, when the bid is <= target, it closes the short trade but the close price is the Ask price of a short position, not the bid price.

In...
Thanks Dave. I propose worrying about this if it proves to be a problem. It didn't when I was testing the bot accidentally using a M15 tf, in the weeks before putting it up here. As I so often say, we shall see.

  • #106
  • Jun 16, 2011 5:36pm Jun 16, 2011 5:36pm
Quoting SteveHopwood
Thanks Dave. I propose worrying about this if it proves to be a problem. It didn't when I was testing the bot accidentally using a M15 tf, in the weeks before putting it up here. As I so often say, we shall see.

No Prob.

Just bringing it to your attention. I am going to play with the code myself especially to see how it affects the large spread pairs. See what happens.

Dave R.
  • #107
  • Jun 16, 2011 5:39pm Jun 16, 2011 5:39pm
Quoting droland
No Prob.

Just bringing it to your attention. I am going to play with the code myself especially to see how it affects the large spread pairs. See what happens.

Dave R.
Excellent plan.

  • #108
  • Jun 16, 2011 6:00pm Jun 16, 2011 6:00pm
Ok guys. In light of some unpleasantness here of late, I am temporarily closing this thread to give some contributors time to cool down and stop provoking one another.

I shall re-open it on Sunday night. Any contributors then wishing to carry on the row from earlier on, will find themselves banned from my threads without further warning.

I am also coming down with a heavy hand about the future development of this robot. There is nothing experimental about it. It was trading successfully before I uploaded it here.

So, I am sharing it with friends, not inviting 'improvements'. I do have some plans for future development in light of the ineffectiveness of the balanced trade filters, but these plans are mine and will be programmed in my own time.

The bot is named Pippo in honour of the contributor who provided me with the initial idea; without this idea, the bot would not have been born. There his contribution to the ea's development ended; every other aspect of the strategy it trades is mine. Pippopotamus does not like some of the aspects of this; I do not care. This is my robot and will continue to be coded as I want.

In between now and rejoining the thread on Sunday, some of you need to re-read post 1; there are those here who have shown that the information contained there has gone so far above your heads that your posts are ridiculous.

So guys, enjoy the rest of the trading week and have a great weekend. Continue raking in those demo pips with Pippo; it is brilliant to see others reproducing the results I was getting during the bot's early development.

  • #109
  • Jun 17, 2011 10:28pm Jun 17, 2011 10:28pm
Testing. Wanted to post my updates.

dr

OK Cool. It is still up. I will post my update next.
  • #110
  • Edited Jun 18, 2011 1:05am Jun 17, 2011 11:29pm | Edited Jun 18, 2011 1:05am
Here is my update on my H1 results. Looking really nice. I did do, what I always do and over leverage something. My XauUsd was so tight on the Bands that I increased the lot size for this pair.

Ouch.

I am now trying to recover but look at the rest. Gotta love it.

WARNING: I did re-program the EA to allow the EA to take a full 0.5 band pip profit (See my previous post). I found that using the H1 TF, if your profit was 0.5 Band value LESS SPREAD, on some pairs you had less that 1 or 2 pips profits (Pips value of 0.5 x Band Pip value LESS SPREAD) - (have a look at chfjpy on H1. Less than 10 pips on the full band width).

My change captured 100% of whatever the 0.5 Band pip value was. On some H1 pairs and on most H4 pairs it is not that significant. I do believe that it is a good change that Steve should incorporate.

Since I screwed up on my lot size on xauusd, I have to wait until I recover my margin to continue but this is why we do this on demo before we go live.

Finally. Steve. Don't respond to, react to, listen to, feel bad about, or in any other way feel responsible to or for the psychotic behaviour of people in your public forum.

I realize you don't need me to say it but I had to.

Just so we are square, I don't really agree with some of your remarks either, but it is your kitchen, your food, your cooking and I am your guest. My decision to be here is mine and I respect your house and intend to obey your rules (OK, to the best that I can).

Dave R.
Attached Image (click to enlarge)
Click to Enlarge Name: DetailedStatement.gif Size: 5 KB
Attached File(s)
File Type: zip DetailedStatement.zip   9 KB | 506 downloads
  • #111
  • Jun 18, 2011 1:26am Jun 18, 2011 1:26am
Quoting droland
Here is my update on my H1 results. Looking really nice. I did do, what I always do and over leverage something. My XauUsd was so tight on the Bands that I increased the lot size for this pair.

Ouch.


Dave R.
Hi,

Would it be possible to code a trade comment input so that we can run multiple instances of the EA on different time frames and easily see which time frame is more profitable?

Also, rather than the basket closing at breakeven, could there be a BE profit setting coded?

Just my 1c's work
  • #112
  • Jun 18, 2011 4:18am Jun 18, 2011 4:18am
Quoting droland
Here is my update on my H1 results. Looking really nice. I did do, what I always do and over leverage something. My XauUsd was so tight on the Bands that I increased the lot size for this pair.

Ouch.

I am now trying to recover but look at the rest. Gotta love it.

WARNING: I did re-program the EA to allow the EA to take a full 0.5 band pip profit (See my previous post). I found that using the H1 TF, if your profit was 0.5 Band value LESS SPREAD, on some pairs you had less that 1 or 2 pips profits (Pips value of 0.5 x Band Pip value LESS...
You have convinced me. Nice one.

OK guys, open for business again.

  • #113
  • Edited 5:41am Jun 18, 2011 5:35am | Edited 5:41am
Dave R, when I open your detailed statement from post #110 and look at the open GBPJPY trades I notice something weird.

At 9:15 on the same date a 1.2 and 0.2 lotsize are opened at the same price. Does someone have an explanation for that? Has it to do with your modification?
Attached Image (click to enlarge)
Click to Enlarge Name: droland open gbpjpy.JPG Size: 23 KB
  • #114
  • Jun 18, 2011 8:29am Jun 18, 2011 8:29am
Updated version in post 1.

I received a pm yesterday from a contributor wishing to take Pippo live next week; this prompted me to make some changes that were in the pipeline anyway. Please see the Trade closure and TB margin etc filters sections in post 1. Bear in mind that the code has yet to be tested and so might not be watertight.

I have made some alterations to the display. The lower level dashed target lines are now the same colour as the trade-trigger levels, whilst the irrelevant ones in the middle are removed.

  • #115
  • Jun 19, 2011 1:58am Jun 19, 2011 1:58am
Quoting JulesvH
Dave R, when I open your detailed statement from post #110 and look at the open GBPJPY trades I notice something weird.

At 9:15 on the same date a 1.2 and 0.2 lotsize are opened at the same price. Does someone have an explanation for that? Has it to do with your modification?
No, this had nothing to do with my modification.

If you go back earlier in this thread, you will see that I noticed some anomalies in the order of the appearance of the Lot sizes. I checked the code and the lot size is governed by where the current price is located within the various Bands. Steve numbers the Bands from 1 to 12. Band 1 is between -2/8 and -1/8 and has the Lot4 (largest) lot size.

The coding looked fine to me but where I believe the anomaly comes from is that every hour (on the H1 TF) the Lines are recalculated. If the recalculation moves the Bands enough, a different Lot size will occur for the next Pending Order and may be out of sequence with the open trades.

In the example you found on my statement, I already had 0.6 GbpJpy open. I expected that the next Pending order would be 1.2. This didn't happen. The EA opened a pending order for 0.2 Lots. I didn't understand what happened then so I deleted the 0.2 Pending order and entered my own pending order for 1.2 Lots at the same price as the 0.2 order. I didn't notice but the EA again posted a Pending order of 0.2. They both opened at the same time.

Subsequently I rationalized that since the coding looked fine the only factor that changed was the Line positions.

I haven't thought this through completely so I wasn't going to post but since you asked that is what I believe is causing the anomalies.

I think that the code may have to be changed to look at the open trades and ensure that the next pending order Lot Size is the next in the progression rather than just what the Band calculation dictates.

In other words for what happened in my case, the 0.2 Pending Order should not have appeared. Since I had a 0.6 Lot size open, the new Pending Order should have been 1.2. At least I think that should be the case.

Dave R.
  • #116
  • Jun 19, 2011 4:17am Jun 19, 2011 4:17am
Quoting droland
No, this had nothing to do with my modification.

If you go back earlier in this thread, you will see that I noticed some anomalies in the order of the appearance of the Lot sizes. I checked the code and the lot size is governed by where the current price is located within the various Bands. Steve numbers the Bands from 1 to 12. Band 1 is between -2/8 and -1/8 and has the Lot4 (largest) lot size.

The coding looked fine to me but where I believe the anomaly comes from is that every hour (on the H1 TF) the Lines are recalculated. If the recalculation...
Just to confirm, all this is correct.

Quote
I think that the code may have to be changed to look at the open trades and ensure that the next pending order Lot Size is the next in the progression rather than just what the Band calculation dictates.
I have no intention of doing this. This bot should be traded on the H4/D1 tf; these will be subject to fewer MM recalculations.

  • #117
  • Jun 19, 2011 5:49am Jun 19, 2011 5:49am
Quick heads-up guys. I received this pm from Eamonn yesterday:
"Morning Steve hope you are well.

I have been following the Kevinator system this last two weeks. It is very profitable and looks like it could be coded into an EA. Have a look if you've got time.................."
Eamonn and I are cooking up the ea between us; I will post it when it is ready and if it looks to have legs. In the meantime, have a read at http://www.forexfactory.com/showthread.php?t=297744

This could hardly be a simpler system, looks like good fun and could be the idea solution for those of us who want to spend a few hours a day scalping. The thread is an easy read. I have read the first 10 pages or so, and have already learned to skip most of the posts apart from Kevin's. The rest are the usual mixture of, "Eh? Whut? Who am I?" and mewling about repainting indi's (don't get me started ).

  • #118
  • Jun 19, 2011 6:11am Jun 19, 2011 6:11am
Quoting SteveHopwood

This could hardly be a simpler system, looks like good fun and could be the idea solution for those of us who want to spend a few hours a day scalping.
Good news, Steve. The Kevinator system is solid - looking forward to it.

Rob
  • #119
  • Jun 19, 2011 12:24pm Jun 19, 2011 12:24pm
Quoting SteveHopwood
Just to confirm, all this is correct.

I have no intention of doing this. This bot should be traded on the H4/D1 tf; these will be subject to fewer MM recalculations.

I think I agree with you. I had already started an H4 demo and I like it better for a number of reasons not the least of which is that MM recalculations are fewer. I haven't seen the anomalies yet that you get on the H1 tf.

I haven't watched the MM recalculations on H4 but logic would dictate that the position change of the MM Lines would not be as great.

Also, the Band width appears to be generally wider on the H4. On my tests they are about 2x larger. This would lead to maybe fewer trades but larger profits per trade.

I have attached my one day run on H4 for reference. Actually very nice results but I obviously need more forward testing.

Dave R.
Attached Image (click to enlarge)
Click to Enlarge Name: DetailedStatement.gif Size: 5 KB
Attached File(s)
File Type: zip DetailedStatement.zip   3 KB | 380 downloads
  • #120
  • Edited 4:48pm Jun 19, 2011 4:19pm | Edited 4:48pm
Steve.

On market open, your latest update was opening and closing pairs in rapid succession. The worst was GpbJpy and EurJpy. It must have opened and closed 20 times.

I have attached a log file for your reference.

Dave R.

PS. I went back to the previous version and that stabilized it so some small bug I think.
Attached File(s)
File Type: zip 20110619 - Copy.zip   7 KB | 430 downloads
0 traders viewing now