Fixed a mistake in the pairs trading lecture

[quantopiancal]

An error was pointed out by community member Emilio via helpscout:

Hi, I found a mistake in this lecture. In the moving average section at the end, there is this piece of code:

# Calculate rolling beta coefficient
rolling_beta = pd.ols(y=S1, x=S2, window_type='rolling', window=30)
spread = S2 - rolling_beta.beta['x'] * S1

I noticed that in the rolling beta calculation y=S1 and x=S2. However, when calculating the spread S2 now becomes the dependent variable Y and S1 is the independent variable X Thanks, Emilio

Rene confirmed it was a mistake and offered a fix via slack:

That fix is reflected in the code change.

[quantopiancal]

@richafrank Jamie asked me to test the change for truthiness. Will re-run to change the cell output and test for truthiness.

[richafrank]

Nice. Ok, new request for you: Would it be possible to run this in an environment where we aren't generating all the FutureWarnings or saving them in the output for everyone?

[quantopiancal]

Hey Rich, I did my best to cut down on those error messages - there is one error left.

I did my best to alter the code while keeping the results exactly the same, but don't have the know how to convert pd.ols to whatever the statsmodels equivalent is.

I would have just run the code in a notebook with the original pandas version (0.17.1), but the notebook requires using get_pricing(), which is unavailable locally.

[richafrank]

Ah, I didn't realize the warnings were from 0.18 - I had assumed you were running this locally with a newer version! I wouldn't worry about the remaining warnings then. Definitely stick with what's on the platform.

Thanks for updating to Series.rolling.

1. Does std_30 = spread.rolling(window=30).std() need center=False, like the others?
2. Since this will be copied by lots of folks, could you add a space before the center=False args that you added, so we conform to PEP8?