Independent dual programming validation is very popular for analysis results in clinical trials. This raises the question, what is the reliability of independent dual programming validation? There is no research on the topic. This paper uses the Poisson process to build a process model of validation in independent dual programming, and computes probability of undetectable error in independent dual programming validation. By discussing some properties of the probability of undetectable error, this paper explains some phenomenon in validation process by independent dual programming. This paper also uses the number of discrepancies or detectable errors to estimate the expectation and variance of undetectable errors, and the reliability of analysis results from outsourcing.