Recent advances in the field of wearable sensing has promoted the emergence of many health tracking devices, including heart rate monitors. Heart rate monitors are commonly either chest-based or wrist-based. Currently, it is unclear whether there is a substantial difference in the performance of these different heart rate monitors. To determine the difference in the performance, in this paper, we compare two chest-worn heart rate monitors and one wrist-worn heart rate monitor. Our initial results indicate that there is substantial difference between the devices - the root mean square error between devices can be above 10 beats per minute. However, even though there is difference in performance of different heart rate monitors, yet each of these devices are capable of detecting stress (using an machine learning model) with a F1-score of above 0.8. In this paper, we also introduce the idea of formally verifying the rules obtained from the machine learning classifier; such formal verification will enable improving the explainability and confidence of the outcome of the machine learning models.