Multi 10:4 Trader EA

[glabbeek]

Hi Andy,

Don't get me wrong. ADR or ATR are both the right way to go. My background is from stock trading and there ADR is hardly ever used. Since FOREX is a 24 hours business ADR might be OK.

[glabbeek]

I also think that it is VERY important to stay with the NB10.4 system and not be tempted to add new filters or indicators.

If anybody thinks he can improve Bob's system, then do so in Bob's thread. If your idea get's addapted, then move it into the EA.

Great job so far!

[MrLong]

I also think that it is VERY important to stay with the NB10.4 system and not be tempted to add new filters or indicators.

If anybody thinks he can improve Bob's system, then do so in Bob's thread. If your idea get's addapted, then move it into the EA.

Great job so far!

Hi glabeeek,

I agree with you, this system does work, but we do need to filter some things like:

1) Don't take a trade too close to the 240 MA, user can set the distance.
2) Don't take a trade when the pivot is miles away from the 240 Ma, user can set the distance.
3) If a trade has been stopped and the conditions become valid again, user can set the filter to ignore until tomorrow(No more trades with this pair today).
4) Don't allow more that a specified number of trades of a specific currency to enter orders, user can decide the amount of say, JPYs open trades.
5) If we are too close to the 240 MA, then reduce the lot size, user to select distance and reduced lot size.


I've noticed these trades as a basket, drift up and down from 7% profit and back to zero, maybe we need to add a routine to grab profits when above ?%.

Thanks for your feedback
Andy

[snailbeard]

Hi Andy,

1.07 is rocking, with a basket of currencies and after a sleepy day the pips are rolling in fast!

That's without any more fixes since yesterday.

I am still looking for signs of any outstanding bugs...

The only one I came across so far was on the FXDD account it duplicated some pending orders.
It was when I loaded a template that also loaded the EA.
I'll see if I can reproduce this when I am board with watching the open trades

10% growth in 24 hours - this can't be real - I expect on a bad day those pips could roll out just as fast

This is the first time I have see an EA making more than it loses - still early days - mustn't get carried away!

Awesome EA!

Brian

[MrLong]

Hi Andy,

1.07 is rocking, with a basket of currencies and after a sleepy day the pips are rolling in fast!

That's without any more fixes since yesterday.

I am still looking for signs of any outstanding bugs...

The only one I came across so far was on the FXDD account it duplicated some pending orders.
It was when I loaded a template that also loaded the EA.
I'll see if I can reproduce this when I am board with watching the open trades

10% growth in 24 hours - this can't be real - I expect on a bad day those pips could roll out just as fast

This is the first time I have see an EA making more than it loses - still early days - mustn't get carried away!

Awesome EA!

Brian

Hi Brian,

Sounds like when you used the template, you may have also changed the MagicNumber, so the EA wouldn't be aware of the trade.

Best Regards

Andy

[mjws00]

It is definitely a fun one to watch as a basket. One thing to remember is that the 2% risk is vastly underestimated. You may have 3+ almost perfectly correlated trades running. Andy's latest suggestions address some of the concerns I have.

It is also remarkable just how many of these are profitable at some point during the week. Might be at the beginning or the end of the week. The entries typically provide a place to get out BE+ or BE- on the retest of the pivot.

One of my goals, was to see if we had positive expectancy taking every signal if we manage the trade correctly. The win rate on non-filtered signals is very high with good exits. With good human discretion on these, 85% hit rate isn't that hard. If they can be simplified we can code em. This allows good gains and low drawdown for what are sometimes low R:R trades (but possibly low MAE). But we need more data.

Also if treating it as a basket and taking the correlated trades improves results. We can easily control risk and drawdown with position sizing (i.e 3 JPY trades setting up. Allocate 1/3 each). It may be that the benefit of being in the one cross that runs outweighs those that stall or stay flat.

Pyramiding needs to be optimized. It is so simplified for concepts sake that it can completely kill the edge when the market trends. Once again with intervention it is fantastic, but we can measure and test with an ea and easily make this component much stronger.

There is also no reason why averaging in, at say 1, 2, or 3 weekly pivots couldn't lead to very very nice average costs for either longer term trades, or just the midweek reversal style trade. Once again position size can easily control risk. It is human fear and greed that makes this hard.

It is also just as strong on the trade back towards the 240. There is tremendous opportunity as a pure reversal system.

Most systems have that one killer condition that just pummels them. This seems pretty robust even in chop. In trendy markets it crushes if you can get the pyramiding right. I have seen some weakness on swinging around the 240 and on market turns... but that is manageable. I'd still like to see 10,000 trades so we could play with optimizing when it struggles.

Just a few of the thoughts and concepts this thing triggered. Very very few rules based systems catch my imagination. Even fewer laggy indicator based ones. So please forgive the verbiage, I type fast and love to see good ideas evolve quickly. I like adaptability in my trading and testing so I'll hack on the outlier concepts and share the good stuff.

Awesome progress, Andy.

[glabbeek]

I also think that it is VERY important to stay with the NB10.4 system and not be tempted to add new filters or indicators.

If anybody thinks he can improve Bob's system, then do so in Bob's thread. If your idea get's addapted, then move it into the EA.

Great job so far!

Hi glabeeek,

I agree with you, this system does work, but we do need to filter some things like:

1) Don't take a trade too close to the 240 MA, user can set the distance.
2) Don't take a trade when the pivot is miles away from the 240 Ma, user can set the distance.
3) If a trade has been stopped and the conditions become valid again, user can set the filter to ignore until tomorrow(No more trades with this pair today).
4) Don't allow more that a specified number of trades of a specific currency to enter orders, user can decide the amount of say, JPYs open trades.
5) If we are too close to the 240 MA, then reduce the lot size, user to select distance and reduced lot size.


I've noticed these trades as a basket, drift up and down from 7% profit and back to zero, maybe we need to add a routine to grab profits when above ?%.

Thanks for your feedback
Andy

It's almost impossible to translate a system, that based on eye-ball entry's and exits, to a full working EA.

For sure you will need some filters that are not described in the original by Bob, but the other 99% has to be the same. (My opinion)

I'll be running the EA for some time before I'll make any comments. I don't want to be addressing issues that are on a micro level. First get the big picture working correct.

Some people are happy with the profit's that it is making and I agree it is looking very nice. But that doesn't mean it is working as Bob has meant it to work.

I love the visual layout you made and the programming is also clean and easy to follow. I'm very impressed with your skills!

[Sojourner]

Closed my first basket (manually). Think I'm gonna like this thing. Fantasmical work guys! =)

[MurphyMan]

Closed my first basket (manually). Think I'm gonna like this thing. Fantasmical work guys! =)

Me too! I haven't been watching my V1.07 demo very closely, but at one point it was down about 200 pips. Take a look a day-and-a-half later and I was up about 150 pips, so I manually closed the basket.

Pretty cool.

Sam

[slipshod]

Yep, so far so good Closed a basket as well, but left some small trailing entries in (from my hack of the entry code) to keep positions open if they choose to run.

Andy, I'm not sure that reducing the lot size on trades near the 240 is the answer - the issue is the close proximity of the stop which makes any trade highly likely to stop out. Better to have a minimum stoploss distance based on ATR, which will both protect the trade and also lower the lotsize. I know its not 10.4, but it is common sense (imho).

Also I'm seeing a very odd looking sell stop on EU right now. The entry's quite a distance from the price, with a huge 330 pip stoploss and a teensy 40 pip TP. By contrast there's a sell stop on GBPCAD with a 90 pip SL and a whopping 280 pip TP. Most of the time the trades opened by the EA look right, but you have to keep manually checking for oddball cases like these...