I agree.. bugs can be very difficult to find and can introduce bias. In my case my code in live trading is the same as my backest. My vm machine is use for both.You also need to run your code in practice account during a couple of month/year for ...

Machine Learning with algoTraderJo

ok thanks l'm creating my dataset now... 2000-2015 half of the work is done..

Machine Learning with algoTraderJo

If you dont reach the SL within the horizon you skip the day ?The training set could have very different date , is not a problem i guess ?

Machine Learning with algoTraderJo

What you mean by "more favorable" ? less loss ? Let use a long example for understanding.. D = 5 Start with H0 : open 1.0000 SL = 100pip (0.9900) next hour H1 : 1.0050 SL = (0.9950) change next hour H2 : 0.9980 SL = (0.9950) keep next hour H3 : ...

Machine Learning with algoTraderJo

For the outputs evolve a long and short trade forward a (D) number of bars. This means moving the virtual SL for the trade as specified in the system rules above. Record the trade exit for both and assign each to a separate output. Each example has ...

Machine Learning with algoTraderJo

Another question, if only one hit the target.. what value are you using for the other one. i.e We hit the SL for the long and nothing for the short, then we have [TP,?]

Machine Learning with algoTraderJo

I dont use anymore mysql , too heavy for my needs.. I use read_csv function in pandas and then i pickle it on hard drive. After it take few seconds to load with readpickle. I always work with tick and after depend of my need i create with resampling ...

Python Trader code and skills sharing

I dont try those package. Do your own is not more effort than learn new package but its more adaptative in the long term

Python Trader code and skills sharing

I always use stoploss or takeprofit so you need to find which is come firs after the opening of the position. For that i verify every tick after the trade open. If stoploss is hit before i loss the stoploss pip value If takeprofit is hit first i ...

Python Trader code and skills sharing

I use Pandas for backtesting, very powefull lib with everything you need. Last 6 month, i work on backtesting system on a virtual machine with VMBOX on Lubuntu. work very well with python, pandas, scikitlearn, talib.

Python Trader code and skills sharing

Nothing is sure in fx.. i agree

EURUSD

wait for London Open..

EURUSD

vmbox

Can I control my main PC using a 2nd and 3rd computer?

zzzzzzzzzzzzzZZZZZZZZzzzzzz i dont trust dude in tie.. :-)

Machine Learning with algoTraderJo

50 first testing trades with Raccoon SVC... 62.7% win rate. My backtesting is at 58%, l'm over my simulation.. I need to lost more.. :-) url

Machine Learning with algoTraderJo

Before now, i test running learning machine on raspberrypi. but..it to slow for my need. Many librairies like scikit Learn, Pandas, Talib is not well support on arm. Now i use Lubuntu on virtual machine to run my setup, it is working very well. The ...

Machine Learning with algoTraderJo

I use python Pandas with scikit Learn.

Machine Learning with algoTraderJo

For the target: if the SL and TP is not reach at the end of horizon for both, what value are you using ?

Machine Learning with algoTraderJo

That totally clear ! thanks i try this

Machine Learning with algoTraderJo