R for Data Science – Register for interest in March 2024

National Bonds – A Statistical Analysis of Prize Bonds Candidate Numbers

This report provides an independent statistical analysis of the monthly candidate datasets generated by the draw procedure for the National Bonds Corporation PJSC.  The objective of this analysis is to detect and report, where possible, any non-random patterns in the datasets from which the winning bond numbers are drawn.

The Bond ID is a seven-digit number which signifies the owner of a bond or consecutive range of bonds.  It is drawn from the Candidate file and examined statistically.  In doing this, the relative weighting of each Bond ID from the Historic Record file is referenced.  This accounts for the difference in probability of each digit due to a differing number of bonds with different weights being associated with each Bond ID (related the length of time bonds are owned).

Throughout the report, all statistical tests are carried out at a significance level of α=5%.  In carrying out a large number of statistical tests, it is possible for a certain number of significant results to occur purely by chance (i.e. 5%).  A commentary of the results from all statistical tests is provided within the text.