Parallel simulation involves running the same data through two systems and comparing the results1. In this case, the bank's data would be processed using both the modified interest calculation program and an audit software. The results from both systems would then be compared to check for discrepancies1. This technique provides strong evidence of the correctness of interest calculations as it directly tests the program's output against a known and trusted output1. While source code review23, manual verification of a sample of results4567, and review of QA test results8910 can also provide valuable insights, they do not offer the same level of direct, comparative evidence as parallel simulation1.
References:
Parallel simulation in IT testing - Universal CPA Review
5 code review best practices - Work Life by Atlassian
How to Make Good Code Reviews Better - Stack Overflow
Guidelines for the validation and verification of quantitative and qualitative test methods - Mathematics LibreTexts Method Validation and Verification - University of Utah Sample Procedure for Method Validation - NIST Method validation and verification - CFS Good Practices for Quality Assurance Reviewers: Assessing Evidence of Supervisory Review - IGNET How do quality assurance engineers test calculations? - Software Quality Assurance & Testing Stack Exchange Quality Assurance/Quality Control (QA/QC) Plan and Procedures - UNFCCC