EMA 25-50 Trading System (Detailed)

Well, ST is only 5% done optimizing and already there is 60% better result with V1.13.2 than with V1.12. This is what I mean "Now results look really good"

But as you wish here is V1.13 without additional parameter. Very fast view on results (EURUSD 2008.01 - 2008.08):
V1.12
6350 pips including spread (ATRMulti = 2.8)
V1.13
7840 pips including spread (ATRMulti = 3.1, TP1Multi = 3.6)
V1.13.2
10200 pips including spread (ATRMulti = 3.9, TP1Multi = 3.4, ATRSLMulti = 2.4)

Jesus man, you've been partying too much
I'm kidding... yes, minutes after I posted indicator V1.13

now the latest to download is the pavk one?

OK... I wouldn't have noticed it, because I'm trying to achieve a better curve for a longer period (since 2006), and with MM=on (to simulate my real trading).

I use 2.5 to 4% of balance equity (actually working with 4%) - that's because the plan is to work with a maximum of 6 pairs simultaneously.

I tell you what:
Keep working around the first one you've posted today (V1.13.2), and I will work around this one... later we'll share results, OK?

There's only pavikogm's EAs

1) What I posted earlier today was the Indicator with those changes I've been mentionning;
2) pavikogm changed the EA to it (EA_testMM_V.1.13.2). He called it ".2" because he first made another version with issues around SL;
3) Then now he released EA_testMM_V.1.13 (it's supposed to be "the image" of the indicator I released earlier).

how can I make it work? It has not the smiling face, but the "wrong" cross...

Well, it seems impossible to optimize parameters in EURUSD to obtain an interesting and more stable growing curve.
There's too much drawdowns, and it stays around horizontal until the start of this year, where it shoots up.

I have some other ideas to try when I'll have the time.
But anyway after that, I will move on to other pairs, and see what happens on those.

I'm working these days, so it will advance slowly...
Anyone who comes up with some good setups, feel free to share, and save some testing times!

anyway imho we have to optmize just the last 6 months for every pair to keep them always up to date. shouldn't we?

There are systems that can work smoothly for a medium/long period of time... for now, ours don't seem of that kind.

So we might be forced to try that... it's one of my other ideas, though with a little nuance... I'll talk about that later.

the EA seems ok, it has just a problem with the SL/TP trailing... it is always different from the indicator segnals...

There's a reason to it:

First of all, remember that the indicator starts at the start of your chart, while when you test the EA, you fix a starting date;
Obviously, there's a good chance that at the start time of the EA (when it starts waiting for a signal), the Indicator is already in a trade (and in this time, it doesn't wait for signal).
So you have to follow both, until both are out of trades at the same time. When that time arrives, from that moment on, they both take trades and go out of them at the same time.

But sometimes, a mismatch appears:
That's because in the Indicator, you never have a yellow cross in the same candle of a TP (for calculation reasons, it has to be signaled in the next candle), but in reality it can happen - and when it does, the EA activates the SL in the same hour;
Mismatch then occurs when there's immediately an enter signal in the next candle (in the indicator, it's the SL candle).

I obviously noticed it immediatly and paid the required attention.

Anyway the bug is in the trail I definitly see that we have miscalculations between the real entry point and the BE settings at TP1.

example when TP1 is reached the SLs don't go to BE but little below. and TP2 is farer in the EA than in the Indicator.

This is pretty strange there's something wrong the difference is pretty big and I checked by my own eye that the entrance was exactly at the opening of the candle. so the indicator and the EA MUST have worked the same.

Again, I would suggest to change the "4 order thing", what you think about this system:

1) a single trade with all the 4 lot with no TP but with a SL placed where we want (SLAtrMultiplier * ATR)

2) a "limit" order opposite to the sign that is placed at TP1 and when triggered takes the SL of the whole "big" trade at the BE

and so on for the other TPs.

The "limit order system" would avoid the AWFUL slippage we run into with the current system

And this is 1 other thing I noticed.
I started playing a little with the settings of the Indicator and found these to be the best config for the 3 pairs I'm working with:

1) EURUSD
ATR Multiplier 2.2
TP1 Multiplier 1.0

2) GPBUSD
ATR Multiplier 2.0
TP1 Multiplier 0.5

3) EURJPY
ATR Multiplier 1.5
TP1 Multiplier 0.5

I noticed that with GBPUSD we get a lot of TP1 and low TP2, TP3 and TP4. I would suggest to cut TP4 and to increase of the double the level of "sell" in TP1. That would increase greatly the profits imho. Now a loss is 4 TP1 with the "modify" we would have a SL=2TP1 and since we have a Win/loss counter that is 1329/268 with these settings...

Apart from this when the EA will work properly and we have added just the "limit order system" imho we're ready to work live.

It's because of the spread, and pavikogm told us that, when he made the first EA:
that's because the Indicator only looks to one price, the chart price (by default it's bid price), while an EA, after entering a trade (long/short -> ask/bid), immediately starts looking and taking the exits based on the other price.
He told he made it that way (for now, obviously) to compare the signal timings with those of the indicator.

mm then it is very important that TPs trigger when they trigger in the indicator regardless to the spread... (we're going to gain a little less...) else all the tps/indicator system is ruined.

in this moment I'm following the trade from of this afternoon on EURUSD and in the indicator TP2 was reached at the second candle, with EA it has still to be reached even now...

the spread thing has to be shut down just not paying attention to it. we'll earn little less. Or anyway this is just my opinion.

I think there would be an easy way to keep the EA performing as the indicator, while not having those little losses at TP1.
In the EA, we would have to make our stops:
SL - spread -> when we are short;
SL + spread -> when we are long.
That way the EA could look at bid/ask when it's supposed to, and still have exits as in the Indicator.

But I don't agree with keeping it like this... if we evolved to an EA building, the Indicator will be/is secondary, like a first scratch.

Soon, we'll just change the code of the EA to make the stops equal to normal value of SL, while looking to bid/ask, in order to eliminate those nonlogic little losses.

Of course all optimization procedures will have to be made again.
(For now I focuse in learning the best procedures, results are secondary).

I mean that it would be GREAT if we could set "how much to sell" for each of the TPs.

I'm pretty sure it would be an easy modify, but I don't know

it is also important the spread thing, your idea is probably what we need.

There shouldn't be so difference, because spread is included in TP as it is with indicator. EA for TP2 has the formula Tonirai posted:
entryprice +/- TP1 +/- ATRMulti
I've checked this in ST "Result" tabs and everything looks ok.

I'll check indicator if it has the same calculations.

This code in EA is just commented out. You can uncomment it and get proper calculations - BE will be exact for Sell signals.

Of course, Tonirai and Corre, you can make any changes to the EA as you wish

again, checked manually in front of pc:

USDJPY 11.00 long
opening price 110.18
entry price 110.23

5 pips wasted no one knows why.

yes, the system was configured and actioned perfectly

Hi guys! after 2 weeks of training... I'm finally back to civilisation.
so.. what did I miss?

regards,
Zen

a lot
welcome back zeon!

I have had another (imho) BIG idea to gain much more.

Let's use 5 lots, and sell 2 lots on TP1 it would be the 40% of the trade and with 2,5TP1 we would reapay a SL.
Since we get heavier on TP1 and TP1 is the most reached (obviously) TP... that should work very very well...