Keeping it simple is crucial since we don't want to filter ourselves out of everything that could be traded. I will try some new things soon but need some clear cut direction as to the strategies we would like to try. I can program several into the Jugulator and then allow the user to choose in the parameters what strategy they would like to use.

What value for risk is good? is the value in percent? is the default value of 12 high? how about 0.1?

Thanks

Hi,

I used this ea, apperantly it will close all the same currency when margin call or single close. How to avoid this happen? how to change the parameter to make only 1 position to be noly close??

I have lost 3 position when the margin call but if the only 1 position to be close, the other 2 can safe my current margin.

I think we need to step back and see what is it that the EA is trying to accomplish. It is simply a martingale system. That means the probability of success on the first signal is only 50/50 without a signal or trigger. This trade will either close with a profit or a new trade twice the size will be placed to make up for the loss and make a preset profit for one lot, and so on with additional trades. The probability remains the same for all the following trades, just like a coin toss. In 10points3, however, the first trade is placed based on a signal from some indicators so there is some improvement in probability of success for the first trade. If we had a bad signal and a loosing first trade progressive trades are placed in the same bad direction of the first trade as money management and compounding our risk until all trades are stopped out (dynamic stop) or the margin is depleted. Assuming a MaxTrade=6 (a total of 6.3 lots risk for a 0.1 lot initial size), the probability of having 6 trades in the wrong direction and being stopped out is 1 in 64 (I think) with a big loss. This loss will wipe out a small account and not so good for a larger account. You should expect this to happen an average of 5-6 times a year (1D chart). I am not sure if any account will recover from this in the long term, especially if it happens early in the run. Martingale is a game of luck (remember I live in Vegas). That is why that none of the versions with different or multiple indicators has significantly improved the results. Only the MM tweaks have made a difference for trades other than the first trade in the series. An early big loss and a big drawdown will have negated this improvement.

With the current code, the probability of success is improved only for the first trade and other progressive trades are only for money management and are risky trades. However, the probability of success can be improved if every trade is initiated based on a signal from some system indicator, regardless of the direction.

So, here is an idea to tinker with:

1. Pick a simple one or two indicator system (MACD, RSI, etc.) that is responsive to both trending and range bound markets
2. Use the signals from the system to initiate all trades
3. Use Martingale to double up the size of the next trade if the first trade is a looser and continue to double up, regardless of the direction of the next signal. You can set a MaxTrades to control a total wipeout
4. Use a dynamic stop-loss such as ATR/2, preferably for the next time frame
5. Use fibo levels or pivots for profit taking

This is not quite a martingale since the sizes of wins and losses are not multiples or the same size. But, since we are using consistent SL and TP targets, in the long-term they will average out.

Regards,

Go Gators!

Matt, glad to see you back. I've been doing some more thinking and I feel this EA is only about half done. It only makes money (big money) when the market moves against you, which is ok, but it would be better if you could ride a trend for larger profits and build on them as the market moves with you, just like you have to increase your lot sizes as it moves against you. So the way I see it is you have protected the equity in the acct. with what you currently have built and makes profits at the same time with quite a bit of risk. The risk is ok but you should be able to get it started in the right direction better than it is currently doing and eliminate some of that risk. The newer version should add to the positions and have an option of moving a trailing stop to protect profits at each pipstep level. Another thing could be that if so many pips are recorded rapidly on a 5 min bar (+) trailing stops are moved agressively up to protect profits and (-) positions are closed before equity can be drawn down much. I have a couple more ideas but maybe I'll wait to see how these are recieved. :>) These additions would not take away from what is there now but add more profit potential. Thanks for reading terry

I think this sounds reasonable and will begin to look in this direction along with some of the other ideas on this thread. Thanks for the input.

The ideas of moving SL up if the position goes good for you and increasing TP as well along the way sounds like a good idea to me. There have been a number of times I have seen the TP almost achieved and then falls back where we may be able to take a different approach and take advantage of that somehow. I would be interested to hear all you have to offer on these subjects and will think thru some ideas about implementing this as well. Thanks for the input.

I was thinking of something like that as well but to have a setting that says if price is within 1% of TP price and the market goes in the wrong direction then close with a profit anyway. Not sure how many time I've lost a trade because TP was maybe off by one pip. That would suck. Need some kind of TP margin variable.

I tryed this one with these settings:

TakeProfit=25;
Lots=3;
InitialStop=80;
TrailingStop=20;
MaxTrades=3;
Pips=5;
SecureProfit=10;
risk=24;
TF=Daily;
Pair=GBP/USD;

50k to 850k in 2006 (attached). Also a big profit in 2005. 90% modeling quality.
Forward testing are showing a risky style (attached).
Maybe we can try to improve?
Cheers

Attached Images

backtest.jpg (161.8 KB, 249 views)

Attached Files

FWTest.htm (35.8 KB, 136 views)

maybe can use profitmultiple coded by autofx in goblin EA