Karl Denke was arrested in Münsterberg in Silesia on December 21, 1924, after a lodger he had attacked survived and raised the alarm. He hanged himself in the police cell that night. What the police then found in his house was a ledger in which he had recorded his killings, and the number usually given for him, thirty or more and possibly over forty, comes from that ledger. It has never been tested by anything. No charge was put to him, no evidence was examined by a court, and nobody ever cross-examined the man who wrote the book.
Fifty-four entries in this archive describe someone against whom no conviction exists at all. Some died before they could be charged, some killed themselves in a cell within days of arrest, some were declared insane and committed, one escaped hours before he was due to enter a plea. They are scattered across four centuries and every continent, and the thing they share is not how they died. It is that a number is attached to each of them and no court ever fixed it.
That is worth stating plainly, because a trial is not only a mechanism for punishment. It is the only occasion on which a victim count is ever forced to be specific. Charges name particular people on particular dates. Evidence for each is examined separately, and counts that cannot be supported get severed, dropped or acquitted. The number that survives a trial has been through something. The number attached to a man who was never tried has been through nothing but the judgment of whoever was holding the file.
Dean Corll is the case that shows this is not an argument against the figures themselves. Corll was shot six times by his own accomplice in Houston on August 8, 1973, and there was no trial. The count for him is twenty-seven, and it is as solid as any convicted total in the archive, because twenty-seven bodies were recovered and identified. That number never depended on a court, so losing the court cost it nothing. It is the control case, and it tells you what to look for in the others: whether the figure rests on something that would have survived being questioned.
Mostly it does not, and the drift is usually upward. Elizabeth Bathory was arrested on December 30, 1610 and never tried. The figure supported by the testimony taken at the time is thirty-seven to fifty-one; the figure she is famous for, six hundred and fifty, rests on a single uncorroborated claim. Belle Gunness, who was never tried and whose own death has never been settled, has eleven victims recovered and identified in whole or part, while the sheriff's estimates at the time ran past forty and were never established. Joseph Ball, who shot himself in Texas in 1938 to avoid arrest, carries fourteen, and that figure rests on investigative attribution and local account rather than any court finding.
Confessions drift worse than estimates, because they sound like the best evidence there is. Manuel Delgado Villegas, never brought to trial after being diagnosed with a severe mental disorder, confessed to forty-eight killings. Seven are corroborated. Several of the others were shown to be impossible. Fred West was charged with twelve murders and confessed and retracted repeatedly before hanging himself at Winson Green on January 1, 1995; the figure usually attached to him is thirteen and there is no conviction against him for any of them. In both cases a trial would have done the obvious thing and made him account for each one separately, which is exactly what never happened.
Bruno Ludke is the case that stops this being a one-directional argument, and it is the most important entry in the set. He was never charged and never tried. He was declared criminally insane under a Nazi decree and handed to a medical facility in Vienna, where he died on April 8, 1944 after being injected as part of an experiment. Eighty-odd killings were attributed to him on the strength of confessions obtained in custody, and German historians working from the surviving files have since concluded that there was no evidence he had killed anyone at all. The drift runs both ways. A number set by whoever held the file can be far too high, and it can be a fiction from end to end.
When there is no trial the process does not simply stop. It frequently lands on somebody else. Four of Elizabeth Bathory's servants were tried and three were executed while she was confined at Csejte Castle by royal order and never charged: the servants got the trial, the noblewoman got house arrest. In West Jordan, Utah, on November 17, 1991, Margaret Kastanis died the same day as her three children and no charge was ever brought against her. Her husband Sam was arrested and tried for all four deaths. At his trial a forensic pathologist examined the bodies and concluded that the most likely sequence was that she had killed the three children and then stabbed herself, noting that she was the only one of the four without defensive wounds and the only one with the blood of all three children on her clothing. Sam Kastanis was acquitted in July 1993, having stood trial for killings the medical evidence attributed to his wife.
A striking number of these cases end in custody, within days of the arrest, which means the state had the man and then lost him. Denke hanged himself the night he was brought in. Vlado Taneski, a journalist who had reported on the killings he was accused of, was arrested in Kicevo on June 22, 2008 and found dead in a cell at Tetovo prison the following day with his head in a bucket of water, recorded as suicide. Itzcoatl Ocampo killed himself in the Orange County jail on November 27, 2013, before his trial on six charged deaths, so no court ever ruled on any of them and none was tested in evidence.
Masten Wanjala, who was never charged and never tried, is the same failure without the death. Arrested in Nairobi on July 14, 2021, he was produced in court within days, but plea taking was deferred twice at the investigators' request. Thirteen counts of murder were prepared and never put to him. He escaped on October 13, 2021, hours before he was finally due to enter a plea, and three officers on duty were arraigned over it. His entry carries zero convictions and a police attribution of ten to thirteen killings of children, and there is now no mechanism by which that will ever be resolved.
Where the record fails it can fail completely. Charles Floyd was arrested in Tulsa on November 22, 1949, declared insane, committed to an asylum and never tried for any of the killings. What became of him afterwards is not recorded in any source: no release, no transfer, no death has been traced. Neither his birth year nor his death year is known. The strangulations attributed to him in Tulsa between 1942 and 1948 rest on police attribution and nothing else, and the archive cannot even say when the man they are attributed to was born.
Albert DeSalvo is the best known name in the set and shows how durable an untested figure can be. He was never tried for any murder. His conviction, in January 1967, was for the Green Man robberies and sexual assaults, and he was sentenced to life for those. He was killed by another inmate at Walpole on November 25, 1973, and nobody was convicted of that either. Up to thirteen killings are suspected and none was ever established against him in court, which has not prevented the figure being repeated for sixty years as though it had been.
Enriqueta Marti Ripolles died in the Reina Amalia prison in Barcelona on May 12, 1913, reportedly killed by other inmates, though the prison records give illness. She was never tried. Ten or more deaths are attributed to her and none is established. Her case has the additional feature that the account which made her famous was substantially built by the press of the day, and the trial that would have separated the reporting from the evidence never took place.
What the fifty-four say together is narrower than the sensational reading and harder to dismiss. It is not that these people were innocent, and in several cases the physical evidence is overwhelming. It is that the figure beside each name was set by an investigator, a coroner, a newspaper or the man himself, and then nothing ever happened to it. A trial is an adversarial procedure whose main product, for an archive like this one, is a number somebody had to defend. Remove it and what is left is the first estimate anybody made, which is why these counts range from twenty-seven bodies recovered and identified to a man who probably killed nobody, all filed under the same heading.
Every one of these entries now says on its own page that no court ever ruled, and why. That is the only honest form for them to take. The number stays, because the deaths are real and the cases belong in the record, but it is no longer presented as something established, and the reader is told who established it instead.