Language:  ▾

Pair screening

Tab pair screening

This tab is to be used for screening of profitable stock pairs as per the parameters entered. The program automatically tests all pairs in the sectors entered and will display those that correspond to specified criterion. Pairs found can be further sent to detailed testing at the “Backtesting” tab.

Screening settings

  1. Go to the tab Screening
  2. Set the parameters in the left hand side of the screen
    1. Parameters of calculation model
    2. Performance parameters of the pair
    3. Sectors, in which the program will screen the pairs
  3. Click on the button “Selected sectors” to test selected sectors, or click on the button “All sectors” to test all defined sectors.

Method for pair screening

During screening, Stock Pair Builder will test both options of a pair, ie.: "A-B" and "B-A". A reason for this approach would be usage of statistic variable “Standard deviation” that can have different values. It is calculated from data series A-B or B-A. Essential fact is that the calculation equation for standard deviation from final set of data provides ESTIMATE of standard deviation and therefore the value calculated from data series A-B vs. B-A is different.

Duplicated pairs (“A-B” vs. “B-A”) can be filtered using PairID that will automatically identify pairs created by the same stocks without their order.

Stock Pair Trader enumerates pair in the same way as it is specified in the portfolio - including order of stocks.

Work with the results of screening

Results of screening (pairs table) can be sorted by clicking the heading of individual columns.

Unsuitable pairs can be deleted by clicking the button "Delete selected"

Screening results can be saved to the hard drive as a working portfolio using the button "Save results"

Sending of pairs to a detailed backtesting

Selected pairs can be send for a detailed backtesting using the button "Backtest selected". Pairs will be copied to the “Backtester” tab.

Warning

One pair calculation table is extensive and it covers several MB in the computer memory. Depending on the memory size it is possible to work simultaneously (backtest) at only 20 000 - 50 000 pairs!

Limitation is not related to pairs tested in frame of the screening (there can be unlimited amount of such pairs), but is related to number of pairs in the table of results of the screening, and backtester, respectively.

Screening parameters (for example setting of minimal daily volume, equity stability, profit, ...), or number of tested sectors need to be adjusted to this limitation.