Every number in this archive comes from a court. That is the rule the roster runs on, and it is the right rule. It also means the numbers record something narrower than what happened.
Twenty-two entries here show the gap directly. In each one a man or woman was charged with a series and convicted of part of it, and the part that fell away is still on the file somewhere, unresolved.
Maoupa Maake was charged with thirty-five murders in South Africa and convicted of twenty-seven. Sararat Rangsiwuthaporn has been linked by police to fourteen deaths between 2015 and 2023 and convicted of two, sentenced to death twice. Michaela Roeder was charged with seventeen deaths on a German ward and convicted of a smaller number, sentenced to eleven years. Gerald Stano confessed to forty-one, was convicted of ten, and was sentenced to death for three.
Those are not the same kind of gap, and the difference matters. Maake's twenty-seven convictions and Rangsiwuthaporn's two both rest on evidence a court accepted. Stano's forty-one is a number he produced himself.
Sometimes the shortfall is procedural rather than evidential. Milton Johnson was tried on separate indictments for separate episodes, as Illinois practice then required, and convicted of four of ten murders. James William Miller was convicted of six of the seven Truro killings; his co-defendant Christopher Worrell had died in a car crash before any of it came to court, which is why Miller was tried on a joint-enterprise basis for deaths he said he had only waited through.
Sometimes the state simply stopped. Marcus Shrader was convicted of one of four murders, never tried for the others, and died in prison in 2007. Robert Dean Weeks was convicted of two deaths and never charged with the other two. Curtis Don Brown was convicted of one killing in 1989 and only charged with two more twenty years later, after cold-case DNA.
Two entries in this group have a conviction count of zero.
Robin Gecht is associated with eighteen or more killings in the Chicago Ripper case and has been convicted of none of them. He is in prison for the attempted murder, rape and aggravated battery of a woman who survived, sentenced to 120 years in 1984. Bartolomeo Gagliano was tried for three killings, convicted of none, escaped custody five times and hanged himself in prison in 2015. Both men appear in an archive of serial murder without a murder conviction between them, which the entries say plainly.
The most uncomfortable cases are the ones where the missing convictions belong to someone else.
Alexander Bychkov was convicted of nine killings; another man had already been convicted of one of them. Altemio Sanchez raped and killed in Buffalo for two decades while Anthony Capozzi, a man with schizophrenia, served twenty-two years for two of the attacks. The Frankford Slasher case has eight or nine deaths, of which one man was convicted of a single killing on disputed evidence, and the rest have never been solved.
John Paul Phillips is the sharpest example of what happens when the record gets tidied. A newspaper reported that he had been convicted of two killings after DNA linked the cases. He was convicted of one, and the DNA arrived fifteen years after he died.
Read together, the twenty-two say something about what a victim count is. It is not a measure of how many people someone killed. It is a measure of how many killings a prosecutor could prove, in a particular jurisdiction, with the evidence available at the time, against a defendant who was alive and fit to stand trial.
That is why this archive carries both figures wherever the record supports it, the convicted count and the attributed one, and refuses to average them. The gap between the two is not a rounding error. It is a list of cases nobody closed.