How Gang Members Behave Like Animals, and Maths Experts Are Now Predicting Where They Will Fight Rivals with 99% Accuracy
Amanda Williams, Daily Mail (London), March 26, 2013
Maths experts have used geometric equations learned from wild animals to predict the location of fights between rival gangs with almost 99 per cent accuracy.
Jeffrey Brantingham, an anthropologist at UCLA, in California, who uses statistics to study crime, has employed a theory devised by Alfred Lotka, an American statistician, and Vito Volterra, an Italian mathematician, in the 1920s.
The pair observed that similarly sized rival groups of a species – from lions to hyenas – claim territories whose boundaries form a perpendicular line halfway between each group’s home, be it a den or a beehive.
Their findings – called the Lotka- Volterra equations – have been long used as a staple of ecological theory.
Brantingham applied it to 13 equally sized criminal gangs from the Boyle Heights neighborhood of Los Angeles’ East Side.
He and his team, aided by police, identified an area or ‘anchor point’ which functioned as the gang’s home base and used the Lotka- Volterra equation to draw borders between the turfs, Smithsonian.com reports.
Brantingham said that according to the equation, if the gangs are equal in ability, the boundary between them was equi-distant and perpendicular between their anchor points. He added: ‘It’s a nice, simple, geometric organization.’
According to the equations, researchers then predicted where the violence between the rivals gangs was most likely to take place. They predicted 58.8 per cent would occur less than a fifth of a mile from the borders, 87.5 per cent within two-fifths of a mile and 99.8 per cent within a mile.
Analysis of 563 gang-related shootings in the area between 1999 and 2002, showed researchers their predictions were almost exactly accurate, with the location of real-life shootings being 58.2 per cent, 83.1 per cent and 97.7 per cent, respectively.
Brantingham, who said his mapping method better reflects criminal activity than other police methods because it is not dictated to by geography, is continuing to test the territory maps.
The team hope their model could provide a valuable tool for police departments in stemming gang violence and how best to locate police resources.
Brantingham said it could be employed when two gangs appear in an area for the first time.
He said: ‘Where should you put your police resources? This model does a relatively good job of figuring that out before any violence even happens.’
Brantingham said millions of years of evolution have created similar solutions to common problems, ‘regardless of what species you’re talking about.’