Build a pairs trading model.
The algorithmic trading model is to be developed in Python. The code should be as efficient as possible; therefore, the programmer should be conscientious of latency in their coding.
Milestones:
1- Parameters for screening
Develop a program to test for the cointegration of 2 stocks within the following universe; S&P500, Nasdaq, Russell 2000 and 3000. Perform the Cointegrated Augmented Dickey-Fuller test on the universe (as detailed above) for mean reversion. Determine the optimal hedge ratio by performing a linear regression against the two-time series and then testing for stationarity under the linear combination.
2- Optimization for trading
The programmer will need to then develop the code needed to filter for profitable and actionable trades within a given time period (reversion to mean within x amount of days - TBD). The pair we are searching for must be statistically significant, therefore we will accept only pairs with a p-value (test statistic) smaller than the 5% critical value. The next step is to limit the pairs to those that on average revert back to the mean within a period of x days. Next the signal shall be established when the pairs extend beyond the one standard deviation mark from the mean.
3- Order system
To be discussed on the possible strategies in which we can enter and exits trades and the most efficient price possible (open to suggestions).
About the recuiterMember since Aug 26, 2017 Alvaro Garcia
from Barcelona, Spain