je-suis-tm / quant-trading Goto Github PK

Hey Bro,

I love your works on git and want to share my ideas how to predict next market crash.
When you add external term on the brownian dynamic equation like mxdotdot + cxdot + kx = f(x) and solve it, you will get similar pattern. you can do curve fitting and getting c,k and also m and getting market dynamic contents and use this eq to predict next crash.

Check my paper fig.3.2.

https://drive.google.com/file/d/14XCTCn6in6tEab8ZJSCww6ISkawmg5Zq/view?usp=sharing

Jason

Parabolic SAR backtest.py:46: SettingWithCopyWarning:
A value is trying to be set on a copy of a slice from a DataFrame

Hi, I have just stumbled upon your project and in the GitHub project description there was mention of a CTA. Perhaps I missed it and that was meant to be the Oil Money Project or Smart Farmers project? Or is there a CTA.py missing? Thanks

hi,
Thanks for that great work!!
I am trying to generate just the buy/ sell signal (and not the pattern). I took the 5 years stock data however signals[signals['signals']!=0] is always empty irrespective of any stock I take. Appreciate any help.
Thank you!

Empty DataFrame
Columns: [Open, High, Low, Close, Adj Close, Volume, std, mid band, upper band, lower band, signals, cumsum, coordinates]
Index: []

Hi.First, thank you for the code. I am a beginner quant coder. I ve implemented another version of your code written by Sabir Jana.
https://github.com/sabirjana/blog/blob/master/Pair_Trading/sj_pairs_trading.ipynb

Now what I do is add the new daily close price and shift +1 in the notebook and re do the calculations . When I do this, the signals differ from the previous ones. So the problem is, if I do get a signal on that specific day, when I do the calculation again on the next day closing price, the signal from the previous day might not be there.

positions1=capital0//max(df1['asset1'])
positions2=capital0//max(df1['asset2'])
Hey, I think there is a possible forward bias in the pair trading code. When you size the positions in accordance to the maximum value an asset takes over the entire duration (including the time which has not been seen by the strategy), then the knowledge of maximum value over the duration peeps in.
A different approach can be to size based on the prices available at time 0.

Do you have any suggestions how I could take the MACD and pair it with the Awesome Oscillator to create a live trading script on Currency, not just backtesting?

Hi,

your github repo is really great and very clear, so i want to deal more with your tools. However, it takes a little bit of time to get used to it.

But for the beginning I would like to see some exemplary Jupyter notebooks with your explanations from the readme. In this way you were able to get directly into the topic.

Maybe you could create some notebooks, I would really like it!

Hey I hate doing it like this but your email ([email protected]) listed on the contact section of your website didn't work.

I'm sure you're very busy but still wanted to reach out because you only miss the shots you don't take, as i always say. Would you be interested in contributing to this momentum signal idea I have? Even if you could just grace us with a few hours per week or whenever you have ome downtime to spare. We would greatly appreciate.

We're just a small group of young developers (most high school) age working on a social trading app that mainly operates on Discord right now. I won't go into specifics right now as it's not relevant. But I do want to clarify that we can't compensate you for our time as we don't have the capital or the revenue to offer you a fair rate.

I know this is a long shot but just had to try, you seem to be full of knowledge with technical analysis and indicators so you would be perfect for this project.

Thanks for your time

Sent email to [email protected] , address was not found. what is the correct address?

Pair trading backtest.py Line 89:
signals['fitted']=np.mat(sm.add_constant(signals['asset2']))*np.mat(model.params).reshape(2,1)
Why assets2 became an independent variable at this time？But Can see that it is the dependent variable in regression model.
x=sm.add_constant(train['asset1']) y=train['asset2'] model=sm.OLS(y,x).fit()
I tried to exchange them but got worse results，I know your code is right,but i wanna know more abot this,why such two variable can be exchanged here?
THX!!!!!!!!!!!!!!!!!!!

I have download the gbp/usd data from the histdata.com, but I can't find the column's name.

could you give a example file?

thanks, this project is awsome, learn a lot.

When I run MACD Oscillator backtest.py for Apple Inc. (AAPL), I get a NameError.

ma1:11
ma2:29
start date in format yyyy-mm-dd:2019-03-18
end date in format yyyy-mm-dd:2020-03-18
ticker:AAPL
slicing:10
[*********************100%***********************]  1 of 1 completed
Traceback (most recent call last):
  File "C:/Users/25del/PycharmProjects/KrakenBot/MACD.py", line 134, in <module>
    main()
  File "C:/Users/25del/PycharmProjects/KrakenBot/MACD.py", line 124, in main
    new = signal_generation(df, macd)
  File "C:/Users/25del/PycharmProjects/KrakenBot/MACD.py", line 37, in signal_generation
    signals = method(df)
  File "C:/Users/25del/PycharmProjects/KrakenBot/MACD.py", line 21, in macd
    signals['ma1'] = signals['Close'].rolling(window=ma1, min_periods=1, center=False).mean()
NameError: name 'ma1' is not defined