Search

MesserMonitor

Methodology

Inclusion criteria, source levels, review path, coverage limits and corrections in context.

Scope of recording

MesserMonitor’s recording scope covers publicly supported incidents in Germany from 1 January 2026 up to the relevant processing date. The decisive requirement is that a knife or another stabbing or bladed implement was used against a person, its use was immediately attempted, or a specific threat was made with it. With suitable source support, this includes machetes and swords, among other weapons and implements.

A discovery is not published immediately. It must meet the inclusion rules, be supported by an admissible source and pass the factual and privacy checks.

Included

  • physical attacks involving knives or other stabbing or bladed implements, especially where stabbing or cutting injuries are documented,
  • immediately attempted attacks involving such objects without a confirmed injury,
  • specific threats involving a knife or another object capable of stabbing or cutting,
  • recognisable handling of a knife or another object capable of stabbing or cutting during a specific interpersonal confrontation, where the object is in the immediate proximity of another person who is present and creates a concrete threat; deliberate brandishing or a separate verbal threat is not required,
  • robbery, attempted robbery and extortion where such an object is visibly kept ready at the scene or used directly and the incident is directed at people,
  • supported follow-up reports and corrections to recorded incidents.

Excluded

  • mere possession, carrying, discovery or seizure of a knife or another stabbing or bladed implement,
  • checks and violations of carry restrictions without a specific threat or attack,
  • self-harm, accidents and property damage without an attack on another person,
  • general disputes in which a knife is neither directed at another person nor used as a threat,
  • arrests or investigations in which a knife or another stabbing or bladed implement is mentioned only indirectly,
  • incidents outside Germany, even if an arrest or later investigation takes place in Germany,
  • general debate, commentary or statistical reports without a specific incident.

Borderline cases and uncertain information

Unclear cases are not made to fit. They remain unpublished or are reassessed only after a reliable follow-up report.

  • If a reliable source expressly reports that a person was possibly, probably or allegedly injured with a knife or a specifically named object capable of stabbing or cutting, the remaining uncertainty does not automatically exclude the incident. The uncertainty is retained in the published description.
  • The description “sharp object” or “unspecified weapon” is not sufficient without an account of how it was used. An incident may be included when the source documents a threat directed at a person, an immediately attempted use, a stabbing or cutting action, or an injury caused by it.
  • If the information does not show that the action was directed at a person, the case is not published.
  • When accounts conflict, duplicates remain unresolved or core details are unsupported, the case remains unpublished until those points are resolved.

Sources and traceability

At least one reliable, publicly accessible source must support the stabbing, cutting or threatening action directed at a person and the classification of the incident. Sources may include publicly accessible releases from police and investigative authorities, releases from other public authorities, reliable regional and national media reports, publicly accessible follow-up reports and corrections, and other publicly verifiable sources where the relevant details are sufficiently traceable.

Sources are weighted according to their origin, proximity to the incident and the traceability of the information they contain. Original official releases generally carry particular weight for factual details. Journalistic sources may provide additional confirmed information. Multiple sources may complement an incident and support the plausibility of its classification.

A search result alone is not a reliable source. Reposts, automatically generated result pages or contributions without independently verifiable details are not automatically sufficient as the sole basis for a case.

Each published case identifies its publicly accessible evidence. MesserMonitor does not reproduce complete third-party texts or images; it summarises the supported information in its own words.

Automated and AI-assisted processing

MesserMonitor uses automated and AI-assisted methods to find publicly accessible sources, structure information, link reports to incidents and summarise supported details. These methods support the processing of larger numbers of reports.

The methods must not add or invent unknown facts. Unsupported claims are not presented as established facts. The publicly accessible sources identified for each case remain decisive.

MesserMonitor sets the substantive and editorial rules for processing and publication. Published content is subject to editorial control and may be reviewed, changed, rejected or corrected later on the basis of the available sources. MesserMonitor remains editorially responsible for the published content.

How a case is processed

  1. Discovery and automated preprocessing: Publicly accessible leads are found, structured and screened to determine whether they describe a specific action with a knife or another object capable of stabbing or cutting, directed at a person, and fall within the inclusion rules.
  2. Source review: The stabbing, cutting or threatening action and the classification of the incident must be supported by at least one reliable, publicly accessible source. Where several sources are available, their details are compared.
  3. Recording: The date, location, incident type and other material details are structured and summarised from the sources. Unknown or conflicting details are not added or presented as certain.
  4. Reconciliation: Different reports may be linked to the same incident. Possible duplicates, follow-up reports and conflicting source details are compared so that an incident is not counted several times merely because several reports were published.
  5. Publication or rejection: A case appears only when the inclusion rules, source basis, privacy and public presentation meet the applicable requirements. Insufficiently supported or unresolved cases remain unpublished or are rejected. Published content may later be corrected or withdrawn.

Follow-up reports, corrections and withdrawals

Different publications may be linked to the same incident. A published record is intended to represent the incident rather than the number of reports published about it.

Later reports may clarify the incident date, location, injuries, case type or other details and may support a supplement or correction. They are not automatically counted as a new case.

If an entry is no longer sufficiently supported or must no longer remain public, it is corrected or removed from public display. The basis for the change remains traceable internally.

Presentation, location data and privacy

Only information required to understand the incident is made public. Names of private individuals, contact details and exact private addresses are not published. Public coordinates are permitted only at state, district or municipality level; finer or otherwise inadmissible coordinates are withheld.

Where known, an incident is assigned to its actual occurrence date. The source publication date and later updates are handled separately. If no reliable occurrence date is available or the reported dates conflict, the publication date of the principal source is used as an approximation and labelled accordingly. Unknown or approximate dates are not presented as exact.

Figures refer to the published cases found in the reviewed sources. They do not replace police statistics or assess the safety of individual places.

Coverage limits and the dark figure

MesserMonitor can document only incidents that become public and meet the inclusion rules. Incidents that are not reported, not made public or described too vaguely are absent. Differences in publication practices also affect which regions and periods are more visible.

No extrapolation: The number of published cases cannot be used to calculate the true number of incidents or a completeness rate.

Comparisons over time are meaningful only when the recording scope, available sources and inclusion rules are taken into account. If the source coverage or rules change, values from different periods are not directly comparable.

Difference from police crime statistics

Police crime statistics count offences known to the police and fully processed by them under their own counting rules. MesserMonitor documents publicly described incidents. One incident may involve several people affected, suspects, offences and publications. The timing also differs: MesserMonitor uses the occurrence date where known, whereas police crime statistics generally assign a case to the year in which police processing was completed.

Report a correction or addition

Notices are checked against the source provided. Where a change is required, the entry is corrected and the change is recorded. You can also use the correction form to submit a missing source or follow-up report; name and email address are optional.

Report a correction