Forex Factory (https://www.forexfactory.com/forum.php)
-   Trading Systems (https://www.forexfactory.com/forumdisplay.php?f=71)
-   -   Feedback on my first EA, Giveaway to Forex Factory (https://www.forexfactory.com/showthread.php?t=570960)

Edorenta Dec 18, 2015 1:39pm | Post# 1

Feedback on my first EA, Giveaway to Forex Factory
 
3 Attachment(s)
Edited on the 12/31

Hi everyone!

I'm quite new to coding but trading for more than 4 years. I got into algo dev as I am studying IT aside my finance Msc, have fun with my ea.

For better/fewer trade opportunities please look toward enabling the ATR filter and the soon to come News/Session filter.

Strategies:
NB: Every single strategy includes confirmations brought by oversold/overbought and trends (given by the MACD and momentum) setups. No entry is random. You can configure your own setups depending on the pair you work with and the timeframe you work with. No set file is attached yet as I am not able to optimize the EA to any pair atm, so please stop sending me pm about set files.

(1)Stochastic Crossover:
The EA will simply look for a stochastic reversal crossover with proper market conditions aside to open positions.
(2)Double MA slope reversal:
The EA will look for a fast MA slope reversal given that the slope is in the direction of the trend, and that the slower MA slope is reaching its inflection point as well.
(3)RSI breakout through RSIMA:
The EA will look for RSI moves around its average. I advise to use this strategy only to scalp ranging markets with tight SL/TP.
(4)MACD Crossover:
The EA will look for MACD signal line crossing over/down its Main line. The EA is pickier on long when the signal line is below 0, and pickier on shorts when the signal is above 0.
(5)Volatility Hike (ATR breakout):
The EA will look for a setup in which the ATR quickly rises upon its moving average.
Tweak the Minimum ATR and ATR1 sets to change its sensitivity to volumes and amount of entries.

(1) and (3) are giving the best results in short term trading and ranging market swing-trading, (2) and (4) are better for mid-term trend trading.
(5) is really interesting if you're into news, volatility hike and breakout trading.

The main strength of this EA compared with milions of similar is the way it handles trades:
I spent a lot of time studying order management and I believe that the key to a successful trading is not in the entries but in the order and money management itself.
Mashup-Recovery allows you now to use hybrid SL/TP/TS calculations based on weighted ATR/momentum and fixed pips. This feature should not be used live yet as I am still working on it.

Version 2.2 Released:

Bug fixed:
-Dynamic SL/TP/TS were not working properly, all fixed now. Great backtests are back in town

New features:
-2 new kind of dynamic trailing stop, one by momentum/true range (hybrid), one by double true range (short term/mid term) weighted hybrid.
Have fun, happy new year!


Version 2.1 (old):

Bug fixed:
-Glitched opposite close

New features:
-New strategy modes (Entry on RSI/RSIMA cross) and MACD
-New hybrid ordermanagement functions
-New Maxspread allowed

Version 2.0 (old): Smart Order Management

You will now be able to chose:
-The main strategy (Entry on stochastic crossover/double MA slope change/volatility hike);
-The Takeprofit, Stoploss and Trailing calculation type (On ATR, Momentum, or classic fixed pips);
-More money management options: Classic / Martingale / Anti-Martingale.
-A dynamic risk calculation mode: Classic = fixed lots, Dynamic = % of your equity risked every trade.

Soon to come:
-Time/News/Session filter
-More strategies entirely configurable
-Hybrid order management for SL and TP calculation (Hybrid momentum/ATR and capped (i.e. in case of news if the atr is really high and momentum as well, the SL and TP will be capped at a given amount)

I've done a few backtests using Tickstory 99% data modeling, and I believe there is a true potential in this ea, as the worst chainloss I had in my backtests was about 10, which makes the recovery mode really interesting. If you can optimize the sets a little you can get to a W/R close to 60% and no more than 7 losses in a row over the last 2 years.
I'll try to run some backtests at work and upload them here as my computer seems not to be powerful enough.

I'm usually using Tickstory Pro 99.9% as historical modelling to backtest on MT4 but cannot atm as my computer is dump.
If someone could try some optimization and post the results it would also be great!
Merry xmas friends

It's my turn to give a little back to the FF community

PS: The enclosed backtest (EURUSD M15 with news filter) was achieved using Mashup Recovery 1.0 that is not available anymore and had a different entry logic based on RSI/RSIMA crossovers. I will in the near future integrate this strategy as an option to Mashup Recovery 2.
Strategy Tester_ Mashup-Recovery.pdf
Strategy Tester_STR(5)RdnSets.pdf
Mashup-Recovery_2.2.ex4

Pat494 Dec 18, 2015 2:04pm | Post# 2

Mash up recovery won't load.

Edorenta Dec 18, 2015 2:40pm | Post# 3

Mash up recovery won't load.
which build is your MT4? Should work

rabbit68 Dec 18, 2015 2:56pm | Post# 4

It is loading fine on my MT4 build 920. Will test on Monday.

AstronautFX Dec 18, 2015 5:16pm | Post# 5

1 Attachment(s)
thanks for this ea

My back test fails. Is this martingail?
Click to Enlarge

Name: test.png
Size: 47 KB

DerBerliner Dec 18, 2015 5:26pm | Post# 6

Thx for your EA !

It Looks like a very simple indi-based EA. On the one hand that's good - it's KISS !

But on the other Hand: usually these simple indi based strategies hardly work well in flat markets

Hmmm .... so a longer running forward test could provide more inside .... perhaps ....

Edorenta Dec 19, 2015 7:01am | Post# 7

thanks for this ea My back test fails. Is this martingail? {image}
Martingale is On by default but you can turn it off in the sets!
Btw your test quality is even worse than mine and starts 5 years ago, unless you have really good computing power I'm afraid none of us will be able to run a good set optimisation for the past 5 years.

Thx for your EA ! It Looks like a very simple indi-based EA. On the one hand that's good - it's KISS ! But on the other Hand: usually these simple indi based strategies hardly work well in flat markets Hmmm .... so a longer running forward test could provide more inside .... perhaps ....
Hey mate, well yes, I didn't integrate a volatility filter in this one, however the RSIMA is barely used is wide-spread EAs and I had great results scalping with it.
I'm quite sure this can be used to trade what you call flat markets, all a matter of optimization!
I'll work on a better version and sets when I'm home! Abroad atm

evildex Dec 19, 2015 8:23am | Post# 8

1 Attachment(s)
u just need the right settings
Click to Enlarge

Name: Screenshot_17.png
Size: 22 KB

Rijfie Dec 19, 2015 8:24am | Post# 9

1 Attachment(s)
Hi

Have been testing with martingale false on the eur/usd 15min chart for last year.
I have a little loss at the end so not bad at all.
Maybey someone have a better setting.
Go testing on several pairs.
Click to Enlarge

Name: Screenshot1.png
Size: 32 KB

Rijfie Dec 19, 2015 9:00am | Post# 10

1 Attachment(s)
Hi again

Same setting but now Takeprofit 0 and trail 45.
Just getting better
Click to Enlarge

Name: Screenshot1.png
Size: 30 KB

DerBerliner Dec 19, 2015 10:24am | Post# 11

Guys ! I hope you know, that all these BTs with the MT4 are dubious actions. Someone even would say they are BS, maybe complete BS.
Hmmm - especially in the case of poor test quality lower then at least 90% no serious conclusions can be made. The best is to have a quality of 99-99.9% using tick data.
But in any case I think you'll get from the BTs nothing more then a proof for the right way your EAs is following the strategy it is based on. It is mandatory to run it in a forward test or even better on a small live account (if you are willing to risk real Money).
Ok, yes if the EA failed in the BT it wont work in the forward test and even less in the live trading account. So to some extent the BT could perhaps help finding possible appropriate parameters + setfiles - nothing more. These setfiles obtained from BTs must show their truth in a following Forward test !

Edorenta Dec 19, 2015 1:08pm | Post# 12

So to some extent the BT could perhaps help finding possible appropriate parameters + setfiles - nothing more. These setfiles obtained from BTs must show their truth in a following Forward test !
You're right buddy, hence as I said I usually backtest using tickstory's data to get the 99.9% modelling plus raw spread. The trick to avoid here is is curve fitting, but optimisation can be done proper and people claiming backtests are full BS are the real bulshitters out there.
If something has done great for the past 5 years there is no guarantee it will keep on doing great, but it is possible. Backtest is at least a way of getting rid of the worst sets combinations (fat tails) and focus on the good ones

DerBerliner Dec 19, 2015 1:38pm | Post# 13

So it would be good to have a look at some tick data based BTs running in a broader range of param's numbers to see whats going/went on during the last months/years.
I dont have the tick data and I usually do not provide BTs, did it only in the past more often.

So, someone else can enlighten us here ?

Otherwise I could perhaps start a forward demo test ... like this KISS approach somehow.

(Hmmm .... on the other hand: let's remember a very, very simple fact: there are zillions of EAs out in FX world - guess why ? )

DerBerliner Dec 19, 2015 2:33pm | Post# 14

It seems to me the best work time for this EA are the LO/NY sessions, otherwise it could run into flat markets presumably not profitable for this bot

(So - a time filter would be a good addition in the code running the EA fully automatically)

Edorenta Dec 19, 2015 2:46pm | Post# 15

It seems to me the best work time for this EA are the LO/NY sessions, otherwise it could run into flat markets presumably not profitable for this bot (So - a time filter would be a good addition in the code running the EA fully automatically)
I'll do it tomorrow with additional volatility filter as well, so you can either trade sessions or high volatility.

DerBerliner Dec 19, 2015 2:46pm | Post# 16

@Edorenta

Ok, I couldn't resist. I put your EA on one of my test platforms for a demo forward run.

M15 with EURJPY, EURUSD and GBPUSD, no martingale, all other params very similar to yours.

Broker is Global Prime.

https://www.myfxbook.com/members/TGT...-gp-22/1456491

P.S.: Ah ... didn't catch your answer above ! So - ok I'll change tomorrow to your newer EA version with filters !

rabbit68 Dec 19, 2015 11:25pm | Post# 17

does the ea compute rsima internally ? is there a custom indicator for this to visualise ?

{quote} I'll do it tomorrow with additional volatility filter as well, so you can either trade sessions or high volatility.

Edorenta Dec 20, 2015 6:33am | Post# 18

does the ea compute rsima internally ? is there a custom indicator for this to visualise ? {quote}
I didn't code any custom indi to visualise, it is all internal (I set the RSI last values as an array and calculate the MA on this array) however if you want to visualise it, simply add an MA to an RSI window and you'll have it

DerBerliner Dec 20, 2015 7:09am | Post# 19

Hi Edorenta !

Two requests about the new version.

Could you put in the local computer time instead of GMT, it seems to me easier for the users, otherwise in dependence of the broker time we have to recalculate anew which numbers we have to use in the params.

Do you have any recomendations for the numbers to use for the volatility ? I have no experience using these vola indis. Yes, it's easy to say roughly: ok, the vola is usually higher during LO + NY or their overlapping period - presumably.

Edorenta Dec 27, 2015 6:21am | Post# 20

1 Attachment(s)
Hey guys, merry christmas to you all

I've worked a lot on this EA lately and was not able to upload a stable version before.
You will now be able to chose:
-The main strategy (Entry on stochastic crossover/double MA slope change/volatility hike);
-The Takeprofit, Stoploss and Trailing calculation type (On ATR, Momentum, or classic fixed pips);
-More money management options: Classic / Martingale / Anti-Martingale.
-A dynamic risk calculation mode: Classic = fixed lots, Dynamic = % of your equity risked every trade.

I've done a few backtests using Tickstory 99% data modeling, and I believe there is a true potential in this ea, as the worst chainloss I had in my backtests was about 10, which makes the recovery mode really interesting. If you can optimize the sets a little you can get to a W/R close to 60% and no more than 7 losses in a row over the last 2 years.
I'll try to run some backtests at work and upload them here as my computer seems not to be powerful enough.

I'd be happy to hear you guys on any bug/advice to work it up!
Mashup-Recovery.ex4


© Forex Factory