The article explores how to detect changes in multivariate Poisson sequences. Detecting changes in parameters is easier for independent sequences, but trickier for correlated ones. The study looks at how well detectors work under the assumption of independence when the data is actually correlated. The researchers found theoretical ways to measure detector performance when the change location is known. These measurements can be used as a benchmark for comparing other detectors. They also considered the more common scenario where the change location is unknown.