York City police: ShotSpotter 60 percent accurate in recent test

Amanda Dolasinski
The York Dispatch

This article originally was published on Nov. 30, 2010.

The city's $600,000 gunfire detection system is still not performing as it should be, according to the latest live-fire test results.

York City Police tested the system for an hour earlier this month and reported about 60 percent accuracy, said Sgt. Troy Bankert. An earlier York Dispatch review of police data found that in 2009 the system detected only two of nine homicides involving gunfire; authorities said two of the homicides occurred in locations outside ShotSpotter's range.

ShotSpotter officials said the system should be operating at 80 percent accuracy or better. Police simply want the system to distinguish between gunfire and other noises, Bankert said.

MORE:ShotSpotter document reveals key human role in gunshot tech

MORE:'Horrific': Led by York City, county records most homicides ever in 2022

FILE - ShotSpotter equipment overlooks the intersection of South Stony Island Avenue and East 63rd Street in Chicago on Tuesday, Aug. 10, 2021. In more than 140 cities across the United States in 2023, ShotSpotter’s artificial intelligence algorithm and its intricate network of microphones evaluate hundreds of thousands of sounds a year to determine if they are gunfire, generating data now being used in criminal cases nationwide. (AP Photo/Charles Rex Arbogast, File)

The system is made up of 40 rooftop sensors designed to detect gunshots, map their location and notify the York County 911 center within seconds. Eight cameras connected to the system automatically turn in the direction of the shots.

In last week's live-fire test, officers shot firearms in six locations around the city. The system was tested for single and multiple gunfire from .22-caliber, 9 mm and 40-caliber firearms.

Company's response: Gregg Rowland, senior vice president for ShotSpotter, said an operations team was on site as the

live-fire test was conducted. He declined to comment on ShotSpotter's performance.

"Unfortunately, nobody in our company can confirm any accuracy data," he said.

Rowland said ShotSpotter should be performing to 80 percent accuracy or better in its coverage area, which is about 82 feet from the point of gunfire. He said engineers will analyze problems and fix them on site, but declined to say what problems York's system has.

"I'm sure our engineers look into the system from time to time, but I couldn't tell you what adjustments happened in York," he said. "Whatever problems were reported in York, they were probably repaired or will be repaired as part of our maintenance and support."

Issue with cameras: Bankert, who took over running York's ShotSpotter in January 2009, said he is not satisfied with the live-fire test results. He said he specifically wanted to test how the eight cameras connected to the ShotSpotter sensors responded.

"Some cameras didn't respond correctly to the gunfire," Bankert said. "They went the opposite direction."

The cameras are the most helpful aspect of ShotSpotter, Bankert said. The cameras are supposed to automatically turn toward the direction of gunfire about five seconds after it occurs.

Footage the cameras capture is sometimes used as corroborative evidence. Although they do not necessarily film the shooting, they can potentially catch people leaving the scene, and police can pull license plate numbers from vehicles leaving, he said.

Next step: After reviewing the live-fire test report, Bankert said he is not sure how else to tweak the system to record more gunfire incidents.

"I think they can change the sensitivity," Bankert said. "But then there's issues with sensitivity. We tried that before, but there's a lot more sounds that aren't gunfire."

In 2009, the system recorded 710 incidents in May, but most were determined to be construction work, and 685 incidents in July, which were largely determined to be firecrackers.

Last year, Bankert increased the sensors' sensitivity to weed out noises, such as those firecracker pops and construction work.

He also asked York County 911 operators to use discretion when dispatching "shots fired" calls. Dispatchers can replay the sounds that are captured by ShotSpotter sensors and determine if they are truly gunfire. Bankert said firecracker pops have a distinct whistle sound.

York City Police will continue to adjust the system, Bankert said.