Steve Hopwood's auto-trader for David486's H4 trendline scalping system | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Trendline Scalping System 1,907 replies

Trendline Scalping System Renko DAX 10 replies

Auto Hedge Commodity Trader System 29 replies

1M & 5M Scalping System (Scalping Paradise) 21 replies

Trendline, Trendline and Trendline 4 replies

Attachments: Steve Hopwood's auto-trader for David486's H4 trendline scalping system
Exit Attachments

Steve Hopwood's auto-trader for David486's H4 trendline scalping system

  • #41
  • Aug 29, 2010 11:25am Aug 29, 2010 11:25am
HI Steve No worries. I realize that the new idea is much more complex than a simple code tweak. One for the back burner when you have time. If you could add code for the pending trade idea from post 31 and 35 to the current bot first, that would be really great as I think it will really increase the quality of entries. Cheers mate and enjoy the weekend. Invisible
  • #42
  • Aug 29, 2010 12:52pm Aug 29, 2010 12:52pm
Grath sent me an auto-trendline ind, thinking it might be better for our purposes than zigzag.

I do not think it is, but it did give me an idea. I have re-coded the trendline calculation to look at High[] or Low[] instead of calling zz. This seems to me to pick up the most recent un-pierced highs/lows; have a play and see what you think.

Post 1 as usual.

Thanks Grath.

  • #43
  • Aug 30, 2010 10:36am Aug 30, 2010 10:36am
Steve,

New version had no trades on H4 overnight (with my own set), but did have one on the M15 which I am running with lower targets. Quickly shot off to the TP, so very nice trade there. Anyway, the point being that it seems to work.

I will be setting up another 2 demos to run in parallel on H4 and M15 with a slightly tweaked set file, plus with these I will not be using any trend detection. (On the current two I have been using ADX 14 on the next higher TF.)

Cheers

Invisible
  • #44
  • Aug 30, 2010 1:18pm Aug 30, 2010 1:18pm
Quoting SteveHopwood
The latest version is in post 1.
[*]I have removed all the mptm-type management features, for simplicity whilst we develop the robot. I have not removed any code; I have commented it out so that I can re-instate it later if necessary.[/list]
Put one of the management feature. JS or TSL. My timings are such that I cannot monitor the trades when market is in full swing. I am on Pacific time.
The EA takes good trades but by the time I wake up London is almost coming to a close.
And I trade it live.
  • #45
  • Aug 30, 2010 1:44pm Aug 30, 2010 1:44pm
@Vinesh

I am using Steve's MPTM to manage the trades as I also cannot monitor them myself.

My understanding is that Steve has removed this code from the latest version in order to make sure the actual trading code is working correctly. Presumably, it would be a very simple addition to then go back and add it later once the bot is finalized.

Cheers

Invisible
  • #46
  • Aug 30, 2010 4:04pm Aug 30, 2010 4:04pm
Quoting vinesh
Put one of the management feature. JS or TSL. My timings are such that I cannot monitor the trades when market is in full swing. I am on Pacific time.
The EA takes good trades but by the time I wake up London is almost coming to a close.
And I trade it live.
Don't be so stupid.

From post 1: This robot is under development and must not be used on live accounts.

Which bit of this do you not understand?

Take it off your live account NOW
  • #47
  • Aug 30, 2010 6:58pm Aug 30, 2010 6:58pm
Done!
  • #48
  • Aug 30, 2010 7:01pm Aug 30, 2010 7:01pm
Quoting SteveHopwood
Don't be so stupid.

From post 1: [color=Red]This robot is under development and must not be used on live accounts.
But Steve, maybe you should reconsider. Please have some concern for the poor brokers. They have families, children, mortgages, pensions, rent, rates, etc. We all contribute to their profits so they can meet their commitments. Some people are just kinder to them and if they want to trade highly experimental EAs with hard earned Cash then they should be allowed to do so.

Kind regards
  • #49
  • Aug 31, 2010 1:01am Aug 31, 2010 1:01am
Quoting astral77
But Steve, maybe you should reconsider. Please have some concern for the poor brokers. They have families, children, mortgages, pensions, rent, rates, etc. We all contribute to their profits so they can meet their commitments. Some people are just kinder to them and if they want to trade highly experimental EAs with hard earned Cash then they should be allowed to do so.

Kind regards


Sorry to be so harsh to one of the good guys Vinesh; I just needed to make sure you took the bot off live trading immediately.

  • #50
  • Aug 31, 2010 10:51am Aug 31, 2010 10:51am
Hi Steve

A couple of observations from watching the latest version:

 

  1. It seems that a trendline is drawn and then is left there by the bot. Even though the eye suggests there should have been a lot of trendlines drawn after that point, for some reason they are not. Perhaps there is an issue with deleting defunct trendlines?
  2. Just from eyeballing the charts, it seems that the old ZigZag version drew trendlines more in a way that they would be drawn by humans.
  3. There may be an issue with the code that looks for the trend: My two H4 demos have one difference - one takes the trend into account and the other does not. The former has traded, but the latter has not. I would have thought that the latter would have traded more, as it does not care about trend, so the fact that it has no trades seems to suggest an issue with the code that looks for the trend.

Since I prefer the way that the ZigZag version was drawing trendlines, I am going to switch back to that one for testing for now. The M15 set up with trend detection had 3 nice trades before I switched over to the latest version of the bot. I think this has some good potential!

Thanks Steve!

Invisible

  • #51
  • Aug 31, 2010 1:18pm Aug 31, 2010 1:18pm
Hi Steve

Here are some more observations, this time based on backtesting which complement observations of the demo trading in the previous post:

Zig Zag version

 

  1. Backtest period Jan 13 to May 28 2010 on EURUSD M15 with 24 hour trading: 3 trades taken right at the start and then nothing for the rest of the period.
  2. If auto trend detection is on and is set on ADX (I have not tried the others) it trades. If trend detection is off, then there are no trades.

Current version

 

  1. Backtest period Jan 13 to May 28 2010 on EURUSD M15 with 24 hour trading: 11 trades taken right at the start and then nothing for the rest of the period.
  2. If auto trend detection is on and is set on ADX (I have not tried the others) it trades. If trend detection is off, then there are no trades.

The above results seem to indicate there may be some issues in the code. The lack of trades after a very early spurt makes me wonder whether old trendlines are being canceled and redrawn or not. The issue with the trend filter makes me suspect there is some issue in the code there.

I hope this feedback helps. The trades I have seen have been well taken, and the underlying system is good, so I think there is exciting potential here.

Cheers

Invisible

  • #52
  • Sep 1, 2010 4:33am Sep 1, 2010 4:33am
I have done some work on the trend line drawing code. I have coded a minimum 8 hours between legs 1 & 2 and attempted to stop a trade following a re-touch of the trend line during the leg 2 candle. I am not sure if the latter is quite correct yet.

  • #53
  • Sep 1, 2010 8:29am Sep 1, 2010 8:29am
Quoting SteveHopwood
I have done some work on the trend line drawing code. I have coded a minimum 8 hours between legs 1 & 2 and attempted to stop a trade following a re-touch of the trend line during the leg 2 candle. I am not sure if the latter is quite correct yet.

Good Morning Steve,
Put in latest update and the bot has taken a GJ and EJ both short. They look good....

whip
  • #54
  • Sep 1, 2010 4:50pm Sep 1, 2010 4:50pm
I have squished a couple more bugs.

Also added the pending versus market trade options. I have not updated the user guide, but the inputs should be obvious.

The default settings have resulted in an eu backtest with a tiny profit - not enough to be worth running the bot, but better than any backtest with previous versions.

Have a play and see what you think.

  • #55
  • Sep 1, 2010 4:57pm Sep 1, 2010 4:57pm
Hi Steve

Thanks for that. There are a lot of different settings to play with so I will see if there is a set that looks good. (Also in demo I am using MPTM to lock in profits so that may improve things also.)

Cheers!

Invisible
  • #56
  • Sep 1, 2010 4:59pm Sep 1, 2010 4:59pm
Quoting Invisible
Hi Steve

Thanks for that. There are a lot of different settings to play with so I will see if there is a set that looks good. (Also in demo I am using MPTM to lock in profits so that may improve things also.)

Cheers!

Invisible
Me too, a la NB at 25/10.

  • #57
  • Sep 4, 2010 4:40pm Sep 4, 2010 4:40pm
Hi Steve

All my backtesting of this bot seems to indicate that what really dictates the probability of a profitable trade is not the entry, but the trend filter. The best trend filter is RSI, which allows it to do well with 2010 data, but it does not really do well for 2009, for example.

Invisible
  • #58
  • Sep 4, 2010 5:33pm Sep 4, 2010 5:33pm
Quoting Invisible
Hi Steve

All my backtesting of this bot seems to indicate that what really dictates the probability of a profitable trade is not the entry, but the trend filter. The best trend filter is RSI, which allows it to do well with 2010 data, but it does not really do well for 2009, for example.

Invisible
Maybe this is a system that is only profitable under recent conditions. This would make sense. No trader would invent a system that worked 5 years ago but fails now.

Here is the nub of the problem as I see it, bearing in mind that I am trying to code David's system as is, not as might be if x,y,z

 

  1. legs 1 & 2 have to happen after the start of the new week
  2. they can happen both within falling and rising trendlines (a squeeze).
  3. the bot has to be able to take trades in response to market moves in either direction.

Looks like back to the drawing board for this particulat bot.

Hey ho

  • #59
  • Sep 6, 2010 9:37am Sep 6, 2010 9:37am
Quoting SteveHopwood
Maybe this is a system that is only profitable under recent conditions. This would make sense. No trader would invent a system that worked 5 years ago but fails now.
HI STEVE

I do not know what happened in 2009 (my memory capacity does not allow it). What I know is forward testing on demo since 26th August (i.e. 8 trading days) the BOT has had 1220 pips profit with current DD of 520 pips. For people who are weak in maths that is profit of 75 pips per day.

I am going to contribute some money to my poor broker by sticking this on a live penny account for couple of weeks and see what happens. Will update in due course. It would be useful if by the end of this week others let us know how their test results.

Kind Regards
  • #60
  • Sep 6, 2010 12:05pm Sep 6, 2010 12:05pm
Quoting astral77
HI STEVE

I do not know what happened in 2009 (my memory capacity does not allow it). What I know is forward testing on demo since 26th August (i.e. 8 trading days) the BOT has had 1220 pips profit with current DD of 520 pips. For people who are weak in maths that is profit of 75 pips per day.

I am going to contribute some money to my poor broker by sticking this on a live penny account for couple of weeks and see what happens. Will update in due course. It would be useful if by the end of this week others let us know how their test results.

Kind...


that's great result astral

can you tell me your setups please, coz I want to test it alive too...


cheers

ps/ and you're right...doesnt matter what was last year or any year before, coz wount be the same for sure....
0 traders viewing now