MARI - My Automated Revenue Invention EA

ok guys not to worry, thanks anyway for your help especially u ejohnston.
anyway i'm trying it with fxpro.

watch this space...

thank u again

sam

Hello. Wanted to inform TCO of an change that I have been trying out. I added a 4H confirmation trend to the code, requiring the MACD 4H to agree with the trade by confirming the trend direction before setting BuyReady1 and SellReady1. This was an attempt to filter out the "double dip" on the 1H MACD that has been causing killer losses. I have been running it on 15 pairs forward demo since Dec 2. It has taken 36 trades, all winners for 1380 pips. It took 2 sets of trades on the USDJPY and three sets of trades on the EURJPY. These were the only trades taken. USE_RSI_Filter = false. I am going to open it up a little by setting Buys_Below_Zero_Only = false and Sell_Above_Zero_Only = false and see if it will take more trades on more pairs.

I don't totally understand the logic behind the opening multiple trades, and increasing the lot size. Is this a Martingale attempt to recover losses? If it is, it isn't working. I do not think that I will ever be comfortable with your great EA with this feature in there. With the 4H confirmation trend addition, I did see it take 10 winning trades on three occasions. In the past, the only time it took ten trades was when it was losing (I have Max Trades set to 10). I agree with ej5 when he suggests that we find a way to stop the EA from taking multiple trades when it is under water. But most of the time when it takes winning trades, it is in drawdown when it opens the additional trades. I have thought about adding a test that would not open additional trades unles the open trades were running a profit. But this will probably limit most of the winning multiple trades because they start out in a loss before going to profit.
Perhaps this note will give you some ideas to share.

Any body trading this system?
TX.

I am running MARI V2. I have had it running on 4 pairs on a demo account for about 3 weeks. I have some conservative setting with Take Profit, Stop Loss and Max Trades.

So far I am at even steven on the trades. It has made several winners, but has lost everything it made. Right now I am up $5.00, with 3 trades open each with a loss as of right now.

That is an interesting idea. Thanks!

TCO could you post the new version you was testing already? I think many of us have concluded that in the current state it would be unfeasible due to its greatest strength also being its greatest weekend.

The lot multiplier and multiple trades was not designed to martingale or double down to try and recover losses. This was orginally intended to amplify winning trades. The problem is it does a better job at throwing lots of trades at already losing trades. There should be a only add another position after x pips of profit.

I really like the idea of the even longer term check to block the double dipping/whipsaw effect we see sometimes.

I have been running Mari live for about 3 weeks on four pair with 60/15 settings and two pair with 15/1. So far I am at break even, even after a significant number of winning trades. A losing pyramid of trades Knocks me down after I work my way up. I am in agreement at this stage that the bot will not go any farther until we address the multiple losing trades and whipsaw effect.

Easyrider, could you post your code modified version of Mari that reduces the whipsaw effect. I would like to run some tests with it as I think it is on the right track.

TCO gave up on us? lol.

Great Thread TCO...hats off to everyone pulling together here.I have 2 suggestions to limit the big losses and wiping out accounts.I have some of my Ea's where a "close all at dollar amount" is an option.If I want to bank lets say $100 and I have it set on the EU then no matter how many trades it takes to get the job done once it overall reaches $100 for that pair it closes all trades and then starts over.You can set different amounts based on the pair also so the beast GJ you can set at $200 and another quiet pair at $75.00
The second suggestion is an Equity protect.I have this also in some martingales and it begins at 50% equity protect then as the account increases in profit I move the equity protect up also thus protecting profits.So if there is a collapse instead of loosing all gained it will close all trades to save at the required equity protect percent.They usually also shut down for a set amount of hours that you have control of until the dust settles.
I hope these suggestions become an option.

ej5, I have attached the version of the MARI code that I modified to provide for confirmation of the trend on a higher timeframe in an attempt to filter out the times when the MACD on the LTF "double dips". Go through the code to see what I did -- pretty simple, really. You will now see three timeframes in the external parameters box: the original MARI LargeTimeFrame and SmallTimeFrame and a "ConfirmationTimeFrame" that I added. I have been running with 60 - 5 - 240 for these timeframes. You will also see an external parameter called "MacdConfirmMinSlope = 0.00001" This is the minimun MACD slope that must be sensed for trend to be established on the Confirmation Timeframe. The slope is the change in MACD bar from the previous bar to the current bar. You will recall that the MACD parameter are 68, 136, 12, so the MA's wiil be very smooth because they are so long. So the slope will be quite accurate when determined simply by subtracting previous from current. The goal in implementing a minimum slope on the Confirmation Timeframe was to keep the EA out of the market when the 240 MACD was flat and/or changing trend on the largest timeframe. The problem that I encountered is that the MACD varies in magnatude from currency pair to currency pair. So you must optimize this parameter for each currency pair. I have been thinking about trying to "normalize" the MACD values so that they are the same from one pair to the next, but I first wanted to confirm the merit of the approach. The lack of a rapid backtest facility with this EA is a problem. Note that if you set MacdConfirmMinSlope = 0.0, the EA will function as if the minimum slope feature was not present.
TCO is working on his modification, and we must be patient. It takes time to test and gain confidence in a revised model. If it is released too soon, than one becomes burdened with responding to the thread with questions that may be unnecessary if testing takes place first. My opinion. I simply that TCO for an EA that still has great potential.
I am going to add some code to implement your (ej5) suggestion to open additional position only when the EA is running a profit. I have concerns that this will also limit winning trades, because they all start out negative. We will see.

still no trades, default settings etc etc

sam

In alpari.uk. for me also no trades.

Sami, do you have Use_RSI_Filter = false??

easy

ur ea still attach at tf5m ???

EasyRyder,

Thanks for posting this as I feel the EA as it currently stands is stuck. I know TCO is working on an updated version but he should use us as his legs to do a ton of testing. We will find bugs and fixes that may take a single person a long time to find. I agree if he is still trying out basic concepts you want to wait until the bot is mostly water tight so you don't spend time on distracting stuff instead of the meat of the EA.

I have seen cases where the original thread starter gets distracted or life calls and the thread just kind of dwindles. I have also seen cases where the creator finds the final fix and never comes back to the forum. I'm not saying TCO is any of these but we need a few people such as yourself to keep the momentum and bot moving forward.

Thanks so much for sharing.

Yes, When the timeframes are 60, 5, and 240 I attach to the 5m.

hi easy

yes rsi filter is false

sami

Here is my 2 day result which unfortunately my computer isn't letting the statement come up in account history so I took a pic of my results and attached the sceenshot.The demo account is 50K (i know not like I trade..now)using 60/5 with RSI false.I have one trade still in play usdcad at -28.Lots are .1.My broker is Think Forex and minimum lot size is .05 but it did nothing.As soon as I upped the lots to .1 then it started trading.

Spiritfriends, which version and timeframe are you using? Is everything else set at default?