Steve Hopwood's Epsilon Megastar fib auto-trading robot | Page 3 | 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

  • #41
  • May 13, 2010 3:32am May 13, 2010 3:32am
Quoting Chip4Pips
Just a point of information:

The third bullet point on page 3 in the Epsilon Megastar user guide refers to the following template file which is not included in the zip file:

fib 75 25.tpl

I am going to assume that it meant to refer to the "epsilon megastar.tpl" file that is in fact included in the zip, but thought that I should bring it to everyone's attention. Please clarify at your leisure.

Thanks again for the wonderful toys!
Cheers. I will put it right to avoid confusion.

  • #42
  • May 13, 2010 3:43am May 13, 2010 3:43am
Quoting rfking2
Wow you quoted me in the code! That's awesome!
It is a good place to store points like the one you made and which I intend acting upon - makes sure I do not lose them.

V1 is on its way soon. I have had an email suggesting that the D1 Rsi is not sensitive enough in today's market - copyright on the pdf dates from 2005.

Also, there is Scoobs advice early on in his own thread, that your higher tf should be x3 your trading tf. It could be that the H1 Rsi is more appropriate to M5 trading. I am going to make the Rsi tf a user input.

In an email to me during the development of the bot, Scoobs pointed out that 'modern' thinking is that an up trend is at 55+, a down trend is 45- and that in between is no-man's land. At the time, I thought the H4 and H1 Rsi filters of having to move in the right direction would fulfil the same function for filtering out bad trades. This is no longer appropriate in an inputted Rsi system, and so I will be introducing use of Rsi in the 'modern' way.

  • #43
  • May 13, 2010 4:12am May 13, 2010 4:12am
Whilst I code V1, would you guys like to get your thinking caps on in preparation for V2.

camcvco quotes the first post in Jacko's thread; timely reminder this, thanks. I read the post a long time ago and have been meaning to revisit for ages. What he says made sense then; in light of what I know now about market movement, it makes total sense.

To remind you all, here is the point rfking2 made:
"I find it really profitable to add more lots at the tip of a retrace, usually around 50% or 61% (which is normally my entry point in a hard moving market to catch a trend). Also because it makes the market a little more predictable, you can time large hedges well to break even open
losses fast. Might be some good addition ideas for you"

So, what I am thinking is this. Take a down trend:

 

  1. the robot trades the trend resumption at the break of the lower confirm level as usual.
  2. during a corrective phase, send a pending sell at x pips below market at the 50, 61.8 and 100 levels.

Any thoughts, guys?

  • #44
  • May 13, 2010 4:49am May 13, 2010 4:49am
It takes longer to update the user guide, zip everything up again, update post 1 and let you all know a new version is available than it does to code the new version.

For those of you who want it now, here it is.

Attached File(s)
File Type: mq4 Epsilon Megastar V1.mq4   31 KB | 396 downloads
  • #45
  • May 13, 2010 5:27am May 13, 2010 5:27am
Here are the updates I made to V1.

 

  1. Made Rsi time frame and period user inputs. Leave the period alone unless you know what you are doing.
  2. Removed the extra Rsi filters as they are no longer of any use.
  3. Removed the need for Rsi to be going in the right direction before a trade is taken.

Everything is uploaded to post 1. There are no changes to the version a few posts back. If you have read the most recent posts, you do not need the updated user guide.

  • #46
  • May 13, 2010 7:32am May 13, 2010 7:32am
Hi Steve,

Thanks to Steve for this new initiative and for letting me post this to the thread.

You CAN backtest this. If you start the EA in visual mode, pause it, load your template and then delete the EA from the graph it works fine. Doesn't produce very good results though at present it would seem.

From what I can see I think there's perhaps two reasons: a) when there's a trade in progress the EA probably shouldn't create subsequent Fibs which overlap the Fib on which the trade was opened (as the trigger for the close often happens a lot sooner than would have happened with the original fib) and b) the trades might be being closed before you have a full candle above the 75% confirm line (which I believe is the rule in the original method).

Anyway, looks interesting. Hope this helps. There are some icwr ea's around on the web from 2006/2007. Here is one. Might provide some pointers...

All the best,
David
Attached File(s)
File Type: mq4 icwr.mq4   11 KB | 415 downloads
  • #47
  • May 13, 2010 7:50am May 13, 2010 7:50am
That's really useful David, thanks, because it includes code that draws the fib; I can adapt this to the robots here.

I have a good live ej sell open right now - +150 in round figures. The first scale-in trade has been taken.

As the fib has swung lower, I have been moving the hard stop to 10 pips above the 100. If the market retraces to there, then my trades should have been closed by the robot. I don't want to risk a brilliant trade turning into a loser because my machine was disconnected from the criminal for any reason.

I think this is a good feature to add.

I have been sent a pm in which the writer points out that a 50 pip sl is not always appropriate and that some measure of flexibility is necessary.

What I shall do is code in a requirement that the hard stop should be minimum of 10 pips beyoud the hi-lo of the trade-triggering fib.

  • #48
  • May 13, 2010 8:20am May 13, 2010 8:20am
Hey guys, Steve added me to his buddy list so i could post here.
Earlier i sent him a pm with some little ideas about the hardstops and the minimum pips of a so-called "active wave".
I think that many of you already thought about it but anyway i'll qoute a part of the pm here, if it's not a problem to Steve (i'll delete it otherwise).

Quote
(blabla)

Anyway, i suggest to make it adaptive to the many pairs; we might find useful changing hardstops and active wave sizes based on average true range or some other kind of average range of bars.

Though, using this strategy manually i put my hardstops right behind the swing high (or swing low, depending on your position being long or short) of the active wave; in this way, if you find the right size of the wave for that pair, you'll find the right size for the stop too.

(blabla)

So here guys, any suggestion about the customization of hardstops and wave sizes? With customization i mean changing it for different pairs: if a pair is much more volatile, we should use wider stops.
Let us know whatcha think about it

P.S.
Oh and anyway, still in the PM i told Steve that this strategy is a nice way to put toghether 2 of the many tips about trading:
"trade with the trend" and "buy on dips and sell on rallies".
  • #49
  • May 13, 2010 8:24am May 13, 2010 8:24am
Hi Steve,

Thanks again for your tireless efforts.

After changing the trading times to what is suggested by the author of the PDf I have a GU and EU sell open at the moment both opened within the last hour and atm both in profit..

Thanks again

Grant

ps, its only a demo account so i left the GU in.
  • #50
  • May 13, 2010 8:49am May 13, 2010 8:49am
Hi Steve

Thanks for V1. I like the new ideas that are coming in - particularly the SL. I had the previous one on GBPUSD and EURUSD overnight. I have quadruple-checked everything I did, but there have been no trades at all, and I think it looks like there should have been. I have it on an FXDD demo.

Any ideas anyone?

Invisible
  • #51
  • May 13, 2010 8:52am May 13, 2010 8:52am
My live IBFX (M5, H1 Rsi) ej sell trades have been joined by an eurgbp buy.

The poor ole gbp has to be the only currency it is possible to but the euro against right now.

My demo trades with more pairs are storming away - even the same GU trade. Mind, about the safest bet in the universe right now is selling the gbp - don't really need fibs for that one.

  • #52
  • May 13, 2010 8:56am May 13, 2010 8:56am
Here is a version that uses the hi-low of the active impulse wave to set the hard stop.

Don't worry if this sometimes appears to set a stop a long way away if the impulse is a big one. The stop is only there to save our accounts in the event of cataclysm; it should not actually be reached.

I have not updated post 1 yet; I thought some of you might want this feature asap. Updating can wait until later.

Attached File(s)
File Type: mq4 Epsilon Megastar V1.mq4   31 KB | 303 downloads
  • #53
  • May 13, 2010 8:57am May 13, 2010 8:57am
Just finished the trading strategy pdf .. what a great, easy read, very well written and illustrated. Looks like a great system...not too familiar with it Lets see if the bot has legs

Thanks Steve the EA Machine
  • #54
  • May 13, 2010 9:38am May 13, 2010 9:38am
Hehe. I am just settling in to code the hard stop to trail the hi-lo of the impulse wave, and realised I did not mention the most significant change in V1 - an alteration to the trend decision making. Details in the post 1 updates section.

  • #55
  • May 13, 2010 9:53am May 13, 2010 9:53am
Here is the version that updates the stop as the fibs change in response to fresh swings.

Attached File(s)
File Type: mq4 Epsilon Megastar V1.mq4   32 KB | 414 downloads
  • #56
  • May 13, 2010 10:53am May 13, 2010 10:53am
This new robot has great potential Steve - Kudos!


Quoting SteveHopwood
Here is the version that updates the stop as the fibs change in response to fresh swings.

  • #57
  • May 13, 2010 11:53am May 13, 2010 11:53am
Your a busy man Steve...

I have been doing a bit of thinking (dangerous as that may be). Could be my paranoia, but there appears to be a real danger with this method and therefore with this EA for losses when markets are ranging instead of trending.

I am not sure what the answer is or even if there is an answer, but wanted to throw the thought out for discussion...

T...
  • #58
  • May 13, 2010 12:36pm May 13, 2010 12:36pm
Quoting tsquad
Your a busy man Steve...

I have been doing a bit of thinking (dangerous as that may be). Could be my paranoia, but there appears to be a real danger with this method and therefore with this EA for losses when markets are ranging instead of trending.

I am not sure what the answer is or even if there is an answer, but wanted to throw the thought out for discussion...

T...
If the size of the "active-wave" is the right one, that will prevent from taking false signals. At least most of the times.
  • #59
  • May 13, 2010 12:58pm May 13, 2010 12:58pm
I will have to find time to do more testing with this method, but my thoughts so far from this week's worth of testing:

1. there seems to be some serious lag from using the default zigzag parameters. in the doc they don't really define set rules for determining a hi or low to draw the move from (i'll call them zig points). There are different ways to do this obviously, basic definition of swing hi/lo, fractals, zigzag etc. If you watch the default zigzag in visual backtest, and also throw fractals on, u'll see plenty of moves that are ignored by zig but could be used with fractals. The pdf examples don't use either and obviously are very easy to draw in hindsight. I've played with using mtf fractals set to m15, and this could be a good possibility to implement maybe.
2. I've also played with only redrawing fibs on the move in the direction of your trade. IE if you are short only, only draw new fibs on down moves over 40 pips, ignore up moves. This seemed to help things alot, but i've only sampled a short timeframe of a couple of weeks so would need more testing.

Thanks for the changes u've made already steve. lightening up the trend filters was definitely needed, and having the ability to adjust these should come in handy as well.

Anyhow, just thoughts, take them as you will

P.S. this does indeed work in visual backtest fine, other than the scale in feature which I still have yet to see if it works live or not. I assume somebody must be forward demoing this to determine if it works as intended.
  • #60
  • May 13, 2010 2:06pm May 13, 2010 2:06pm
Quoting tazmaan
I've also played with only redrawing fibs on the move in the direction of your trade. IE if you are short only, only draw new fibs on down moves over 40 pips, ignore up moves. This seemed to help things alot, but i've only sampled a short timeframe of a couple of weeks so would need more testing.
I thought about this as well, but have not tested it. When reading the strategy document, most of the trade examples seem to do this for the M5 strategy, but not for the H4 strategy. I'm not sure why there was not consistency between the two timeframe strategies.

Since there is a distinction to be made between an impulse wave versus a corrective wave, perhaps the bot should have a configurable setting to choose whether to limit that fibs are drawn on the impulse waves so this can be forward tested.
0 traders viewing now