Regression Report is a good source for selecting stocks for pair trading. It uses Orthogonal Regression to determine what is the NORMAL RELATIONSHIP between two stocks and calculates how far they are away from the NORMAL RELATIONSHIP (the regression lines).

Each row represents a possible trading opportunity. Any pair can appear more than once because different number of data points used.

How to Use Regression Report:

The Regression report give you ideas about which pairs are good for pair trading. A good pair has the following characteristics:

1. Highly correlated, usually 0.8 and above.
2. The larger the distance is better.
3. If Distance is posiitive, Buy Inst1 and Sell Inst2, otherwise the reverse.
4. The shares ratio of Inst1 to Inst2 is also on the Report.

The Columns:

• Inst1 & Inst2: The two instruments in pair trade.
• Distance: The distance between current price and the NORMAL RELATIONSHIP, in terms of times of standard deviation of distances of all data points. The larger the number, the further away from the NORMAL RELATIONSHIP and the more likely it will back to the NORMAL RELATIONSHIP.
  • if distance is positive: the Inst2 price is higher than it should be under the NORMAL RELATIONSHIP, you can buy Inst1 and sell Inst2;
  • if distance is negative: the Inst2 price is lower than it should be under the NORMAL RELATIONSHIP, you can sell Inst1 and buy Inst2.
• Correlation: The correlation between the two instruments.
• Points: The number of data points used to calculate correlation and regression. Our reports use 21 points for one months, 42 points for two months, ... and 252 points for one year.
• Ratio(I1:I2): The ratio of quantity of I1:I2. If the ratio is 2, the right amount to buy/sell is 2 Inst1 per 1 Inst2.
• Name1 & Name2: Names of Inst1 and Inst2, just for reference.

Why Orthogonal Regression?

The Ordinary Linear Regression calculates a line which makes the squared sum of distances of all points along Y-axis (the black lines in the graph), the Orthogonal Regression calculates a line which makes the squared sum of distances orthogonally to the line (the dotted lines). Orthogonal Regression is good for both X-Y variables is independent. All our Regression Reports are using Orthogonal Regression.

Subscribe to our Regression Report

We provide the following Regression Reports:

• Free Edition: This report contains 1,000 regression lines with pairs of same industry and distance under 1. It is available at here. You can subscribe to get it via email.
• Regular Edition: This Report contains 10,000 regresion lines with pairs of same industry and largest distance. You can subscribe to get it via email.
• Custom Edition: Custom Report can be designed to meet your requirement. You can choose any sectors, industries, capital size, volatility, dividend payout, financial ratios, and any other criteria to filter stocks and regression lines you want. Please contact us if you want to have your own Custom Report.

Please check here to subscribe our report.