I suggest you this one... url

The Paradox System

i get a headache now after reading this tread.. :-)

The Paradox System

Why not try to automate this ? Peak detection : url Find peak for min and max in candle. Then find peak of Oscillator. Then compare..

my grail system

Did you try to introduce a low volatility filter to avoid open position in range market ? something like : Take position only when ADXR is over 25. This to limit to number of open position.

Expanding Grid

Grid : How to risk 1 M$ to make 10 cents ! :-)

Expanding Grid

I try several form of grid trading. Each time, i get out of my money after variable time. Maybe "Expanding grid" could minimize that ? But i'm sceptikcal.. You always can also be trapped in a "expension" channel depending of the trend itself. After ...

Expanding Grid

waste of time.. ? i think no ! I prefer have fun to make EA than watching tv !

Who makes real money from forex?

I found something interesting.. The ratio between D and SL is constant.. D/SL = Y Finding the good Y, you can move D and SL to optimize the slope. Each pair have its own Y depending of the pair caracteristic.

Machine Learning with algoTraderJo

With 6 pairs mix R = .9731 Maybe we can improve again with half hour timeframe ?

Machine Learning with algoTraderJo

My stat with my 6 pairs version: Formula: Y ~ <x> + <intercept> Number of Observations: 8058 Number of Degrees of Freedom: 2 R-squared: 0.9731 Adj R-squared: 0.9731 Rmse: 4816.8535 F-stat (1, 8056): 291372.0210, p-value: 0.0000 Degrees of Freedom: ...

Machine Learning with algoTraderJo

3 pairs mix EURUSD, EURJPY and USDCHF each have it's own parameter

Machine Learning with algoTraderJo

I have something very strange, i get same result than you but 12h later. Your 8h is for me 20H. 8AM in GMT+1 is the end of Sydney session, this not a very decisive time for EURUSD.. i think. ? My 20h seems to have more sens, this after the end of ...

Machine Learning with algoTraderJo

when you say A=8 .. This means A = 8PM in us system (20h in HH24) or 8AM ? thanks sam

Machine Learning with algoTraderJo

Every day on A hour, how your machine know to not entrer in a trade ? In my understanding, to enter long, you use long>0 and short<0 but did you use minimun return value to enter ? If the idea is to predict the outcome of a trailing stop, why you ...

Machine Learning with algoTraderJo

I get something not too far from your resul (trailing sl version). I also force a issued for the output at the end of the horizon if i get only a SL or TP. Without that i dont have enough data (2 years of data before first prediction!) But my result ...

Machine Learning with algoTraderJo

For a Long trade, if you get a very straightforward upside tendency within you horizon, You will not get any SL for your Long and you will get a SL for the short. This example will be skipped by the algo because you need to have both target reach in ...

Machine Learning with algoTraderJo

Real word vs simulation The close day of the trade is not same in two figure but i think i still interesting August to October this with the first version of the SVC image image

Machine Learning with algoTraderJo

Another question for you Jo, you enter long if : the long return is >0 and short return<0 or long return > short return ? (they can be both negative) Thanks sam

Machine Learning with algoTraderJo

F4 is c++ style framework, i respect coder who use this but i prefer 4 generation language like python or ruby. With python pandas dont need any framework juste use build it function and avoid many errors with array manipulation. I also prefer to ...

Machine Learning with algoTraderJo