I need a robot to find cointegrated pairs which are off 2 standard deviations.

my idea is to create a robot which finds cointegrated pairs using simple and  multiple regression, I will give the variables.

The project is: A Robot which will scan a list of stocks or forex to find cointegrated pairs according to the augmented dickey-fuller test,

the robot will scan and then ONLY show the pairs which are out of 2 standard deviations.

The Robot also will indicate the ADF score "-4,32";  ADF%: "90%" OR "95%" OR "99%" 

also will indicate the half-life of the operation 

indicates the Beta of the pair and the Deviation of the residual.

Simple Regression: E = Y - (Beta.X) - A

Multiple Regression: E = Y - (Beta.X) - (Timecoef.Time) - a    (which time is the number of periods)  

Multiple Regression: E = Y - (Beta.X) - (Timecoef.Time^2) - a

Multiple Regression: E = Y - (Beta.X) - (Timecoef.Time) - (Timecoef2.Time^2) - a

half-life : based on ORNSTEIN-UHLENBACK

I have an excel sheet which all calculations to help develop the robot.

I have the calculations of the regressions and the half-life I can support with that

Here's a example of what I want but this indicator only searches for the best period.