Steve Hopwood's Epsilon Megastar fib auto-trading robot | Page 6 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Time dependent auto fib? 1 reply

Auto fib bar count 0 replies

My Daily Fib Trading Journal 2 replies

Daily hi/low fib trading 35 replies

Steve Trading Journal 3 replies

Attachments: Steve Hopwood's Epsilon Megastar fib auto-trading robot
Exit Attachments

Steve Hopwood's Epsilon Megastar fib auto-trading robot

  • #101
  • May 14, 2010 3:20pm May 14, 2010 3:20pm
Last night I was testing with 10 pip fib length on the 1 min and was able to catch a lot of trades. I didnt get pulled into ranges because the trades wouldnt kick in until the full candle was outside either 0 or 100. The main differance is the the fib updates on more zigs.

At the time that steve took those screen shots the zigs just before that big swing were around 10 to 14 pips. So i had EJ, GJ, and UG jump into that big swing. At the bottom I was in profit $12 with just a single micro each. But the first full candle under 0 brought me in half way down the swing. So when it retraced none of them were in profit. As it continued back down the fib update putting my stops just neg of my buy in. And shortly after they all closed in the red.

Making fib scale smaller with the new jumping stop buy in might pull you into bad ranges. But will ensure good position in a big swing. There is most likely resistance at the top of the zig so setting your stop there may be a good idea sometimes because it shows a continuing trend. But there is no way of knowing if it will just go 5 pips high then come down. So i will be sticking with 50 pip SL and using MTMP for my exit expecting to just catch the next swing as it comes.

Have a good weekend everyone.
  • #102
  • May 14, 2010 4:49pm May 14, 2010 4:49pm
Quote
The suffix "et Al." behind someones name is the abbreviation of the Latin " et Alii", meaning the contributors on an (mostly academic) article. I consider Steve and his esteemed colleagues of academic level.


Sorry, Ludo. It was joke, hence the smiley face. Steve always encourages good ideas to be contributed by others which is one of his strengths.

Have a good weekend, all!
  • #103
  • May 14, 2010 5:21pm May 14, 2010 5:21pm
Hey guys. Good to see the discussion continuing here.

I posted V2 just before going out today for my usual Friday teaching away from home. On return, it looks as though the demo V2 account took no trades. If anyone else has anything to report then that will be a help, otherwise it will be time to look at the code again tomorrow with a view to demoing again on Monday.

Re V1, this has been my experience this week, both live and on demo:

 

  1. once I got the trade exit thingy worked out, trades stopped hitting their stop loss because they were closed when the market traded the wrong side of the confirm level.
  2. Consistently, the accounts would have a number of trades open on several pairs. These would go into massive positive upl, only for all this lovely profit to disappear when the market went into its next corrective phase. Eventually, candles would trade between the opposite direction confirmation and the trades would be closed. Fantastically profitable positions would be turned into merely slightly profitable ones, or, worse, losing ones.

This suggests to me that the markets are now dramatically different to those when the strategy document was written in 2005. This does not mean that the strategy is no longer relevant, but does suggest that it needs adapting.

I suggest a three-pronged approach, which should work with both V1 and a finished V2:

 

  1. Dump the scale-in stuff and the follow-the-successive-wave stuff. Take out the full trade at the first entry point, then use mptm to follow with a jumping stop until the trade finally closes.
  2. Trade lots of pairs. Treat them all as a basket, and use mptm to manage them and close them when they hit their basket tp.
  3. Continue to trade as per the original strategy, but widen the min fib size and play with the zigzag/rsi settings.

My instinct is to favour 2 but I need to give this a bit more thought.

  • #104
  • May 14, 2010 5:30pm May 14, 2010 5:30pm
Quoting Invisible
This is the other reason I read your threads: Your writing is hilarious.
Ok, enjoy this one.

Some kids work out stuff by muttering the working out process to themselves. "This one was that, so the next one must be....", that sort of thing.

Years ago, I was giving 8 year old Alistair a piano lesson. He had not been playing for long.

He played his prepared piece to me and hit a wrong note. What happened subsequently went thus, all muttered sotto voce:

Al: You played an E again. It is a D. It has been a D all week. I keep on telling you, it is a D, not an E.

Al: Only one thing for it.

At which point, Al made the universal child's revolver shape with his fingers, turned said revolver to his head and made the universal child sound for pulling the trigger.

On not toppling over dead, Al looked at his revolver-shaped fingers and muttered in disgust:

Al: Damn. Not loaded.

  • #105
  • May 14, 2010 7:51pm May 14, 2010 7:51pm
HI all,

Thanks to Steve for creating this system and I would like to show you my results and my current beliefs.

1. Here are my results for this week:

http://www.mt4i.com/users/bluemele/stats
Version 1, standard settings

2. Every week is different:

I believe that every week is different and am a big believer of Fibs and use them daily while trading futures as well as currencies. They WORK! And, make me money. My comment is more in line with the forward testing that has been done and the need to 'scrap' the method so quickly. As you can see, version 1 was profitable using forex.com demo and yes, there were a lot of pips lost due to retracements, but on top of that, I believe we started already through a much larger trend and missed the big start last week.

I highly advise more testing (forward) be done before a complete scrapping of this system to something completely different as the strategy is sound and probably one of the most used on the planet by people who know how to trade with discretion. If you add trailing stops, you miss so much of the move and it becomes a different strategy. One which will require thorough testing and effort and in the end may not be any better than before.

If you compare all the other systems out there, entries are not the challenge, but exits are and assuming using the MPTPM (sp?) then you are ending your involvement in the trend much more quickly. You can say, well, just get back in then. TRUE, but not always as easy especially on a big move as it will be filtered out right? I believe you have to be OK w/ leaving pips on the table...

Some of the best books I have read about trend/swing trading have all mentioned how OVER-OPTIMIZATION is the #1 reason why Mechanical systems don't work. KISS is always the best.

I am profitable on 2 system EA's currently, NanningBob 60x90 (not the one on my signature as I used too small of an account balance on demo) and 2 MACD's by Steve which are both simple simple simple. I prefer something w/o Martingale and I believe if you keep this SIMPLE and let it breath, it will prosper.... Just my 222 cents...


3. ATR

I believe ATR is brilliant (Oktomaniac) for sizing of fib wave! Great thinking.

I hope this helps any/some...
We are our own best indicator.
  • #106
  • May 14, 2010 8:06pm May 14, 2010 8:06pm
Quoting SteveHopwood
Hey guys. Good to see the discussion continuing here.

I posted V2 just before going out today for my usual Friday teaching away from home. On return, it looks as though the demo V2 account took no trades. If anyone else has anything to report then that will be a help, otherwise it will be time to look at the code again tomorrow with a view to demoing again on Monday.

Re V1, this has been my experience this week, both live and on demo:
[list][*]once I got the trade exit thingy worked out, trades stopped hitting their stop loss because...
These would be my favorites... but that's just me:

- No Hedging and no new scale-ins after the trade goes the wrong way.
- Personally I don't scale in as well if it goes the right way. I just want 1 trade... entries as they are now in V1
- I would prefer to have the possibility to stop the EA drawing the fibs and draw them manually once the trade is open. on/off in the settings.

So basically V1 with the choice to draw the fibs myself.

Have a great weekend guys...
"The only successful substitute for brains is silence."
  • #107
  • May 14, 2010 8:26pm May 14, 2010 8:26pm
Quoting bluemele
I believe that every week is different and am a big believer of Fibs...I highly advise more testing (forward) be done before a complete scrapping of this system to something completely different as the strategy is sound.
I agree. I'm not so certain the existing system has been thoroughly examined. The original exit strategy may be better than a MPTM exit strategy and may result in better R:R. I believe we need to better understand if there are tweaks that will enable us to use the "good" aspects of the exit strategy, but avoid losing the positive UPL only to see the trade close at a loss on a reversal close. Is there a way? I don't know, but I believe there are some brilliant, clever minds here on this thread that might have some ideas. (I respectfully understand I am not of the brilliant, clever category.)


Quoting bluemele
I am profitable on ... and 2 MACD's by Steve which are both simple simple simple.
I'm profitable on Steve's 2MA-Macd as well, but I spent many, many hours testing and tweaking to come up with settings that are working well in forward test. Steve does great work and his commitment is remarkable. I suspect many have abandoned the 2MA-Macd too soon.

Quoting bluemele
I believe ATR is brilliant (Oktomaniac) for sizing of fib wave! Great thinking.
OMG, did I come up with something smart? Thank you I just try to put some thoughts out there and let the truly smart ones respond if there is any kernal of value.
  • #108
  • May 14, 2010 8:30pm May 14, 2010 8:30pm
Quoting DXTrade
These would be my favorites... but that's just me:

- No Hedging and no new scale-ins after the trade goes the wrong way.
- Personally I don't scale in as well if it goes the right way. I just want 1 trade... entries as they are now in V1
- I would prefer to have the possibility to stop the EA drawing the fibs and draw them manually once the trade is open. on/off in the settings.

So basically V1 with the choice to draw the fibs myself.

Have a great weekend guys...
What do you think of scaling out of trades, closing half at a profit and moving the rest to BE plus some pips? It seemed like a good idea when Schlossberg said it.
  • #109
  • Edited May 15, 2010 1:21am May 14, 2010 9:11pm | Edited May 15, 2010 1:21am
Quoting octomanoak
What do you think of scaling out of trades, closing half at a profit and moving the rest to BE plus some pips? It seemed like a good idea when Schlossberg said it.
Absolutely... and thanks for reminding me... forgot that one.
-Close half at the 61.8 extension of the original fib and move SL to BE+5

Actually I wouldn'd use ATR to scale the Fib. Fibs are drawn on swings, nothing else.
But SL calculated through ATR is a good the way to go though.
Cheers
"The only successful substitute for brains is silence."
  • #110
  • May 14, 2010 9:14pm May 14, 2010 9:14pm
Quote
Al: Damn. Not loaded.

Both funny, and also explains who this "Al" guy actually is.

Cheers mate. Enjoy the w/e.
  • #111
  • May 15, 2010 3:46am May 15, 2010 3:46am
Quoting bluemele
Every week is different

I believe that every week is different and am a big believer of Fibs and use them daily while trading futures as well as currencies. They WORK!

I highly advise more testing (forward) be done before a complete scrapping of this system to something completely different as the strategy is sound and probably one of the most used on the planet by people who know how to trade with discretion.
I agree. I manually trade fibs on the Russell every day. I like the original strategy. But I also like how customizable this EA is. So I come at this with a scientific approach. I have control groups then I am also testing different fib lengths using different zzs, on currencies with different spreads, with different time frames, different tweaks in the code, and different exits. I want to know how it will perform most every situation and what the best settings are before I go live. Not everyone runs a bulldozer over a mattress before they buy, but those who do know for a fact the it will stand the test of time. This thing traded for me over 60 times Thursday night. Ended up slightly in the red but I got a lot of data really fast. And that's all I could want right now.
  • #112
  • May 15, 2010 4:47am May 15, 2010 4:47am
Quoting DXTrade
- I would prefer to have the possibility to stop the EA drawing the fibs and draw them manually once the trade is open. on/off in the settings.
Done, and added to V3.

I want the functionality from V1 to be available in subsequent versions, and I want functionality from later versions to be available in V1.

To save myself a lot of copy/pasting code from one to the other, I am combining them both in V3. There will be two trading modes: one to trade as per the original strategy; the second to get into the trades earlier as per V2.

I shall work on this today and have it ready for release in time for market opening tomorrow night (UK). The user guide might not be ready by then.

  • #113
  • May 15, 2010 6:12am May 15, 2010 6:12am
Here is V3 for those of you who like playing with strategy tester.

This combines V1 & V2 and adds the extra facility Uli wants. It should be possible to do all your fibbing manually and let the robot enter the trade; from there, manual traders can either leave it to manage the exit or turn this feature off and manage the trade manually.

I want those of you with more trading experience to put forward your thoughts about the minimum fib size, and automating it in line with the volatility of different pairs.

What I am thinking is this:

 

  1. Use Atr to determine the size of the fib, but if so:

    1. which time frame?
    2. use the returned value, or does it need a multiplier/divisor

Or, could one of you indi coders perhaps modify zigzag so it also returns an average swing size over x number of swings?

If either of these methods can be used successfully, then MinFibSize can become a genuine acceptable minimum for the pair, rather than a one-size-fits-all.

Any ideas gratefully received.

Attached File(s)
File Type: tpl epsilon megastar v3.tpl   5 KB | 402 downloads
File Type: mq4 Epsilon Megastar V3.mq4   39 KB | 489 downloads
  • #114
  • May 15, 2010 8:54am May 15, 2010 8:54am
The early entries function needed a bit of extra code. Even so, I have not seen it take a trade in visual back-testing, so I am still not sure there is not a bug in the logic. Maybe you coders could take a look at void TradeEarlierEntryStrategy() and see if you can spot anything.

I have gone back to simply naming the bot Epsilon Megastar. This is not one of those robots that is going to have more and more and more stuff added, so we do not need version numbers.

The user guide is up to date. Everything is in the post 1 zip.

  • #115
  • May 15, 2010 9:00am May 15, 2010 9:00am
Quoting SteveHopwood
Here is V3 for those of you who like playing with strategy tester.

This combines V1 & V2 and adds the extra facility Uli wants. It should be possible to do all your fibbing manually and let the robot enter the trade; from there, manual traders can either leave it to manage the exit or turn this feature off and manage the trade manually.

I want those of you with more trading experience to put forward your thoughts about the minimum fib size, and automating it in line with the volatility of different pairs.

What I am thinking is this:
[list][*]Use...
Steve,

Great idea on the zigzag indicator update. Either way I think dynamic fibs are great.

Also having the functionality to do both on entry/exit's is smart.

I believe sizing of fibs has to be from the current timeframe you are investing with, otherwise your stops (unless static) would be too large possibly? Depends how we end up exiting.

One thing I do on the Russell and ES futures is, I use the smaller timeframe fibs and larger time frame fibs to add a discretionary position as they will typically match on a pivot.

For instance I will mark on my smaller timeframe chart the larger timeframe 50% 61% 76% Fib area's. Then, I will then wait for the smaller timeframe fib to 'match' the larger timeframe.

Typically, a 76 on the smaller timeframe will be the 38 or 50 on the larger etc.. So, I doubt you could program this in or make any sense of it mechanically, but that seems to work for me and many other traders that I know.

Scaling out to 'cover your trade' is never a bad idea in my opinion. Also, scaling in on a great trend is also a good idea. If you enter, and the fibs match again for another entry, why not...
We are our own best indicator.
  • #116
  • May 15, 2010 10:19am May 15, 2010 10:19am
I just received a really useful pm from fxdrive, who pointed out some faulty logic in the stop loss calculation code.

He (she?) also suggested a way in which I could alter the code slightly to make the bot run in strategy tester without having to be in visual mode.

I have done this, and it works - I am running a backtest right now. The update is in post 1.

This download is essential. The logic error needs fixing.

Nice one, fxdrive.

  • #117
  • Edited 1:06pm May 15, 2010 12:46pm | Edited 1:06pm
This one is not in post 1 yet. The new input UseDynamicSwingCalculation should give the game away - uses Atr to calculate the size swing. If smaller than MinimumFibSize, the calculation is adjusted to the minimum.

Attached File(s)
File Type: mq4 Epsilon Megastar.mq4   41 KB | 264 downloads
  • #118
  • May 15, 2010 12:58pm May 15, 2010 12:58pm
Quoting SteveHopwood
This one is not in post 1 yet. The new input UseDynamicSwingCalculation should give the game away - uses Atr to calculate the size swing. If smaller than MinimumFibSize, the calculation is adjusted to the minimum.

Steve,

No attachment to your post.

Lots of exciting new enhancements for Epsilon Megastar! Thank you, thank you.
  • #119
  • May 15, 2010 1:07pm May 15, 2010 1:07pm
Quoting octomanoak
Steve,

No attachment to your post.
Yes there is.

Quote
Lots of exciting new enhancements for Epsilon Megastar! Thank you, thank you.
You could not be more welcome.

  • #120
  • May 15, 2010 2:13pm May 15, 2010 2:13pm
Quoting SteveHopwood
Yes there is.

LOL. You mean yes there is NOW.
0 traders viewing now