### Author Topic: Co-integration  (Read 112 times)

#### Richard Feit

• Newbie
• Posts: 3
• Karma: +0/-0
##### Co-integration
« on: July 07, 2020, 03:16:10 am »
I am new here and would like to know which field I should look at to determine the level of cointegration? What number and where it is.
Thanks

• Sr. Member
• Posts: 270
• Karma: +10/-0
##### Re: Co-integration
« Reply #1 on: July 07, 2020, 03:16:45 am »

#### Richard Feit

• Newbie
• Posts: 3
• Karma: +0/-0
##### Re: Co-integration
« Reply #2 on: July 07, 2020, 04:51:23 pm »
OK, I read the part about cointegration and the results one should look at. I am still very confused.
Can you give me an example of a highly cointegrated pair and one that is only marginally so that I can reverse engineer the results to find out what to look for?  I am most interested in scanning for specific criteria for another type of strategy on your paid service.

Richard

• Sr. Member
• Posts: 270
• Karma: +10/-0
##### Re: Co-integration
« Reply #3 on: July 08, 2020, 04:52:50 am »
Cointegration is a binary feature. There is nothing like "marginal" or "high" amount of cointegration. PTL uses the standard ADF test with two confidence levels - 95% and 99% - to help you discover cointegrated pairs. It is up to you to decide which statistical confidence level is enough for you to decide if the pair is cointegrated.

https://en.wikipedia.org/wiki/Cointegration
https://en.wikipedia.org/wiki/Augmented_Dickey%E2%80%93Fuller_test

#### Richard Feit

• Newbie
• Posts: 3
• Karma: +0/-0
##### Re: Co-integration
« Reply #4 on: July 18, 2020, 11:04:27 am »
OK, I give up on finding how to judge cointegration.
What I am looking for is how to scan for particular criteria.
For example, can I scan for a 95% confidence level with less than 25% correlation, as an example?