Only recently, it has occurred to me that DVR (or MVR) could consistently be biased. By that, I mean that the statistic might tend to overstate (or understate) the likely outcome. So far, the results indicate that this is not a problem. Anyone who wants more information is invited to ask for it.