PredPol maps cut property crime per patrol minute in an LAPD trial

In a 2011-2013 LAPD trial rated Promising by CrimeSolutions, patrols sent by the ETAS model saw 7.4 percent less crime per 1,000 patrol minutes than patrols sent by analysts.

  • 2 society

    Patrols guided by the model saw an average 7.4 percent less crime volume per thousand patrol minutes than patrols guided by analysts.

Depth runs from 1, a convenience or time saved, to 5, a life or a livelihood preserved.

Status
Reported benefit.
Evidence

foundstudiedreportedclaimed

An independent assessment report supports the benefit.

Compared with

Patrols sent to boxes chosen by LAPD crime analysts. Measured.

The system’s part
Days were randomly assigned to the model's boxes or to an analyst's boxes, with the same officers on both and blind to the condition. The comparison isolates the model's choice of places from the analysts' choice.
Limits
Not yet checked.
Costs and harms
Not yet checked.
Later
Withdrawn. CrimeSolutions notes that the LAPD no longer has a contract with PredPol, citing a Guardian report of 8 November 2021 on the end of the LAPD's predictive policing programs.

Download data Cite Suggest a correction

Benefit type
Reduced Violence
System
Epidemic-type aftershock sequence (ETAS) crime forecasting model, known as PredPol (later Geolitica)
Made by
PredPol (later Geolitica), from the ETAS model of George Mohler and colleagues
Used by
Los Angeles Police Department (Foothill, North Hollywood and Southwest divisions)
Where
Los Angeles, California, United States
Stage
In a pilot
Date
Kind
Aggregate result
Broad benefit
Analyzing Data or Information, Organizing, Planning, and Prioritizing Work
Collections
Scrutiny and findings
Last revised
31 October 2026

Sources

  1. Assessment report, by authors with no tie to the system.

    CrimeSolutions, National Institute of Justice, U.S. Department of Justice. Program Profile: Predictive Policing Model in Los Angeles, Calif. Rated Promising. Posted 28 November 2022.

    This program involved the use of a crime forecasting model to direct police patrol to dynamic hot spots to reduce crime. The program is rated Promising.

Details

AI system

Workflow

The model was re-estimated each day from the previous 365 days of crime reports. Patrol officers were directed to spend available time in the top 20 boxes on each shift.

Follow-up

Conditions for replication

Recent crime reports reach the model from the department's live records system. Officers have free time on shift to patrol the boxes.

Entry details

Sector

Policing.

Purpose

George Mohler and colleagues built the ETAS model, later sold as PredPol, to help police departments decide where to send patrols to reduce crime.

Not yet checked: model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.

Cite this record

Peter H. Vartanian. (2026). PredPol maps cut property crime per patrol minute in an LAPD trial (AIBD-0725). Artificial Intelligence Benefits Database, edition of 31 October 2026. https://petervartanian.xyz/aibd/records/AIBD-0725.html (archived at https://doi.org/10.7910/DVN/AADSII, version 4.0)

BibTeX Data Field definitions