Steve Hopwood's Epsilon Megastar fib auto-trading robot

Re redrawing fibs, this seems like simple common sense.

There is not point in leaving a fib in place when a subsequent swing leaves it looking silly.

Trust me to know about this. Looooooonnnnnnnnggggggggg before this thread, I had a fib trading robot that relied on hedging instead of stops to deal with failed trades. The prob with this approach was that it relied on leaving the original fib in place.

Any idea how silly a three week old fib looks? If not, try it.

Fibs have to move with the latest swing if they are to continue to be meaningful.

Phew! The bot has a trade on EURUSD. I didn't screw it up after all. It was embarrassing to see my bot languishing there on demo while all of your are becoming multi-millionaires running it live.

New v1 going on now. Thanks Steve!

This keeps on coming back to me (saved to my source code so I do not lose it, from the original post):
from rfking2:
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).

I am doing a kind of thinking aloud here, so bear with me. I find posting like this is a help.

What I am seeing with this system as currently posited is this: by the time trades are sent, the move from the corrective wave back in favour of the trend has already had to make a lot of movement. Unless the trend is very strong, this trend resumption is already starting to run out of steam.

The last para can also describe the behaviour of a ranging market.

So, as currently coded, the robot is getting into trades a little on the late side. This does not matter in a strong trend, but is less effective in a weaker trend, or a range that is big enough to kid the robot that a trend is in place.

So, taking a short trend as an example, V2 of the robot will act thus:

 

1. take shorts as per when there is a break of the lower confirmation level.
2. during a corrective wave, we are looking for a resumption of the trend so:
   1. market rises to 38.2, send a pending sell at 31
   2. market continues to rise to 50, so send a pending sell at 44
   3. market continues to rise to 61.2, so send a pending sell at 50

This would get is into the market much earlier in the resumption of the trend.

Any thoughts, guys?

Hi Steve,
Had 2 good trades waiting for me when I got in the house tonight.....EU and EJ.....both steadly marching south. That's a good start...I'd like to see the scaling work on the 4H timeframe...........Have a great Day

whip

Wow, you guys have been busy today. I removed the ascending/ descending rsi fail safes last night to make it trade more. Its nice you took them out of V1. 4 total trades. Lost 11 pips with A/C, G/U shorted and got up to 94 pips then closed at 34 and scaling never kicked in. U/C and E/U are both open and up 22 pips.

A jumping stop to get into the market sounds like a good idea to me. I am going to start testing the latest version now. Looks like its already evolved a lot.

Me too, on demo.

I see you have an easy switch to zigzagme. Have you noticed any differences in trading with this? Did you ever test any of the fibo zzs?

Thing with the zzme is that it picks up the latest swing very quickly. This means the fibo will never become validated, so the robot will never trade.

I use zigzag because it picks up the latest decent sized swing. I know it misses a lot of the intermediate swings that fractals would pick up, but I don't care about those. The robot rejects <40 swings, so I think we can rely on zz to pick up on the latest best one.

I could hand code the robot to pick up all the most recent swings and then decide which one is the one to trade. With a lot of effort. And a lot of bugs. And success, finally, in about a month's time.

Possibly.

So, I am sticking with zz until someone comes up with definitive evidence that x will give a better swing measurement.

Ok, so the attached is never going to set the world alight, but this is live trading with actual cash.

Even better is this: the account has not seen draw down for long after the trades were taken out.

Outstanding are open eg and ej trades with a combined upl of c. +100.

Looking better than most of the stuff I have coded, so far.

The one thing I hate is floating DD. Why? Because it tempts me to move in and close trades out of fear rather than letting an EA handle it for me. (Although perhaps a bit more fear would have been a good thing when I was using PID4. Sorry guys, I should not give you flashbacks!)

The statement may not set the world alight, but the quick movement of trades into profit is a very positive point for me.

Cheers, Steve!

I think most of us here could give ea flashbacks of terrifying intensity. Mine involved falling for the usual stuff - "After x years struggling to find the best trading method and failing......tried every system going.....finally found for ourselves this system......we are licensing it because we are sorry that so many others are following the same losing route....myself and my dedicated band of fellow-strugglers......we want others to share in our success....."

Yeesh. When I think back, I want to shoot myself. When I think how much it cost me, I want someone else to shoot me.

So, I take EA development here very seriously. Ditto my responsibilities.

Just like everyone else here. This is a wonderful community.

This is the other reason I read your threads: Your writing is hilarious.

Steve,

I like the idea of what you stated in your post (copied below) as a way to get into the market. I was thinking about your intent on doing this, do you want to make a full order at one area or spread your order out? If you find that you go to profit quick it might increase the RR.



"So, taking a short trend as an example, V2 of the robot will act thus:

1. take shorts as per when there is a break of the lower confirmation level.
2. during a corrective wave, we are looking for a resumption of the trend so:
   1. market rises to 38.2, send a pending sell at 31
   2. market continues to rise to 50, so send a pending sell at 44
   3. market continues to rise to 61.2, so send a pending sell at 50

Steve

Using your example of a downward impulsive wave to enter. In order for it to then be deemed as a corrective wave it would have to pass back through the 75% level and the move would have to be less than 40 pips.

You could still add pending orders on the reversal but it would still be while the wave is a downward impulsive.

If the move goes against us 40+ pips the wave then turns to an upward impulsive.

Lux

Hi Steve

I love all the changes you are making however I wonder if it would be beneficial to still keep an emergency hard stop? I am merely thinking of the times a currency pair may move 200+ pips, pick us up and then head back the other way.
I may be overly cautious but I always like to know what my "worst case scenario" is when entering a trade.
Would it be possible to keep an emergency hard stop variable in the inputs as well as this new enhancement of moving the stop using the high/low of the active wave ie whichever hits first? That way those like me who are a tad conservative could rest in peace knowing a big bit of voladility isn't going to hurt too bad. Those who like to live a little could still set the emergency stop miles away.

Just a thought.
Cheers

Here you are. The new input is UseAdaptableStopLoss; set this to false to use the standard 50(0) pip stop.

Thanks for your thoughts re V2 folks. I have had a re-think, influenced by these.

The robot has not been as profitable this week as I had hoped. Currently, I have two open, profitable trades; were I to close them, I would end the week with a tiny loss.

I am as sure as I can be that this is because we are not getting into the trend resumption early enough, as rfking2 suggested.

Here are my thoughts. Taking a down trend as an example, because nearly everything is trending down at the mo:

 

1. there is a corrective that peters out and becomes a swing high.
2. from here, we take our scale-in trades at 5 pips below the 50, then the 38.2, then the 25 and finally the 0.

There is a danger of the first trade being taken out by a fresh retrace, but we are trading into a trend and we have to get in somewhere.

ZigZag is not quick enough to pick up the most recent swing high. ZigZagMe does this, as you can see from the two pics that I set up yesterday - ok it is a swing low but the pics illustrate Me's quicker reaction.

I am settling in to think about the program design for V2.

V2 is well under way. It looks as though ZigZag is still the best.

We can all spot a trending market by eyeballing the chart - at least, we should all be able to. Getting a robot to do this is much harder, which is why we need an indi like Rsi to do it for us.

Is there a demand for user inputs to tell the robot what the trend is i.e.

 

1. TrendIsLong=true;
2. TrendIsShort=false;
3. TrendIsRanging=false;

Allied with the ability to make the robot leave the fibs alone, this would give traders complete control over where to pull the fibs and still allow the robot to send the actual trades, so you do not have to sit at the console all the time.

Is there a demand for user inputs to tell the robot what the trend is i.e.

 

1. TrendIsLong=true;
2. TrendIsShort=false;
3. TrendIsRanging=false;

Allied with the ability to make the robot leave the fibs alone, this would give traders complete control over where to pull the fibs and still allow the robot to send the actual trades, so you do not have to sit at the console all the time.

[/quote]

Sure wouldn't hurt to add these inputs.

I think the EA has flexibility already as to how early we get into a trend by adjusting the fib size and the rsi timeframes. Getting these settings correct for each currency pair we are trading is pretty key.

Along with recognising a trend at the earliest possible time I believe this system will do well if we can keep the losses small and capitalise on the big moves. This EA sure allows the big moves to run. It's during the inevitable ranging markets that the EA will string a number of losses together. Keeping these stop losses small while we are waiting for the big moves is key to the success of this system (in my opinion).
Losses are inevitable. If we take 4 ranging losses of 50 pips each and then get a great 200 pip big move we have obviously only broken even. If however we could keep these losses to a max 30 pips each and still pick up the big moves then on the above example we would be nicely ahead. Can it be done? I don't know, but often the big strong moves pick us up and dont look back for a while - making a small hard stop possible.
I may be totally on my own wavelength in looney land but to me we need to test different hard stops for different chart timeframes, fib sizes, currency pairs etc until we have the best settings for 2010 market conditions.

Here is a first draft of V2.

This required some fresh coding, so don't put it on live accounts yet. Play on demo first.

It should send trades as the retrace ends and the trend resumes. Falling, at just beolow the 50, 68.2, 25, 0 etc

Could you coders please look at my logic. I am busy for most of the rest of today.

Cheers