The Controversial Crime-Fighting Program That Changed Big-City Policing Forever


In celebration of New York Magazine’s 50th anniversary, this series, which will continue through October 2018, tells the stories behind key moments that shaped the city’s culture.

Dermot Shea has a name and a face straight out of a stereotypical New York City cop TV show. Adding to Shea’s throwback vibe is the world-weary rasp in his voice, a product of his blue-collar Queens roots and a thousand late nights in squad rooms.

Shea’s 48-year-old head, though, is stuffed with algorithms. He knows the changes in grand larceny rates by precinct and in shooting incidents by time of day, plus the recidivism metrics for robbery suspects, broken down by Bronx blocks, updated as of an hour ago. Yet Shea, the NYPD’s chief of crime-control strategies, isn’t simply some numbers geek. He is also immersed in the subtleties of how to question a drug dealer and the genealogy of the Untouchable Gorilla Stone Nation gang. His look and sound is classic patrol guide. But Shea’s mind has been molded by CompStat.

The management system was created in April 1994. Captains, lieutenants, and other unit heads from individual boroughs travel, on a rotating schedule, to police headquarters each week and are quizzed, in granular detail, about crime trends and the plans to combat them. The NYPD brass asking the questions is armed with reams of statistics, which are analyzed and mapped and projected onto multiple screens. The three-hour sessions can be so confrontational that anxious commanders sometimes vomit the night before appearing on the CompStat docket.

No New York invention, arguably, has saved more lives in the past 24 years. CompStat has helped drive down the city’s crime rates to historic lows and revolutionized policing around the world: Los Angeles, London, and Paris use a form of CompStat. Baltimore has CitiStat; New Orleans has BlightStat. Burlington, Vermont, runs CommunityStat, to battle the opioid epidemic.

CompStat has plenty of detractors, too, who say it helped fuel the stop, question, and frisk harassment of hundreds of thousands of black and brown New Yorkers. There is also considerable debate on just how much credit CompStat, and the NYPD in general, deserves for the crime decline. Cities including Houston and Phoenix saw similar declines and attributed them mostly to economic development and community policing. “Two decades of an expanding economy, and mass incarceration, have contributed the most to the crime drop,” says Rick Rosenfeld, a professor of criminology at the University of Missouri, St. Louis. “Smarter policing, informed by digitized crime data, has contributed in a number of places, including New York City. But we don’t know how much.”

CompStat’s first era dealt with the crime crisis of the ’90s. Its second chapter became notorious for a surge in the use of stop-and-frisk in the ’00s and ’10s. Now, with Shea at the helm, the system is being retooled to try to keep crime low while incorporating something far harder to accurately measure than homicides — New Yorkers’ feelings about their police department.

At this morning’s mid-January CompStat meeting, however, it is the feelings of Shea and chief of department Terence Monahan, 56, that are front and center in a conference room on the eighth floor of One Police Plaza. And those feelings are not pretty. “That’s unacceptable,” Monahan barks, and the room of 200 high-ranking officers, most of them in sharp dress blues, goes uncomfortably silent. Shea, Monahan, and several other top chiefs sit at the closed end of a U-shaped table and spend three hours grilling the leaders of their troops. Yes, there’s been plenty of great news recently: The city ended 2017 with 290 homicides, the lowest count in nearly 70 years. Shootings dropped 21 percent. Shea is deeply proud of these achievements — and he doesn’t waste more than a few minutes praising those assembled in front of him for making it happen. “It is important to reward people and tell them that they are doing great work,” he had told me earlier. “But we will use other forums for that. CompStat is for fixing what’s still wrong.”

Inside a CompStat meeting. Photo: Bryan R. Smith/The New York Times/Redux

At the moment what’s wrong is the December holiday-season robbery wave that hit bodegas (or “bogies,” in cop slang) in the Bronx. “Squad, who is looking at this? This is a commercial gunpoint robbery!” Monahan snaps. Massive video screens flash maps of the precinct, seemingly covered in measles: dots representing recent robberies. Other screens display photos of suspects, alongside their criminal records, their nicknames, and their recent addresses. The precinct commanders scramble to list random facts, then make pledges of renewed diligence.

Shea interrupts. “I want to go backwards here: We have a robbery and there’s a woman in the store. She happens to be on the phone, puts the phone down, opens the door, and the robber comes in?”

“That’s correct,” says Betania Nazario, a sergeant in the detective squad of the 42nd Precinct, which covers Morrisania and Crotona Park. “Do we believe she’s involved?” Shea asks. “Um, we’re still working on that,” Nazario replies. “Hold on,” Shea says, flipping through sheets of paper supplied by a research assistant. “Seven blocks away, a week earlier, we’ve got another bodega robbery in the 4-2, four perps, one of them a female. I don’t remember that many bodega robberies with female perps. Could she be part of this one too?” “Um, that’s a possibility,” Nazario says. Shea’s tone turns cutting: “Unless you’re telling me it’s common for females to be involved in bodega robberies.”

He pauses, sighs, pulls back to reinforce the big-picture message: Crime is down, arrests are down, so slow down and take your time piecing together connections. “Commercial robs are spiking significantly, all with perps casually walking to a robbery,” Shea says. “We’ve got to backtrack, draw that circle, find what building they went into, and ID them. This is time well spent, this is why we close shootings, because we put the resources in, we backtrack and we canvass.” Then he says it’s time for a coffee break, and you can hear the entire room exhale in relief.


It all started in a much different city — one suffering nearly 2,000 murders a year — in a highly improbable setting — over glasses of Champagne inside Elaine’s. Jack Maple was a 41-year-old former “cave cop,” a veteran of New York’s then-separate transit police department, who had developed an off-duty taste for homburgs, bow ties, two-tone spectator shoes, the Oak Bar at the Plaza, and Elaine’s.

While patrolling the decaying subway of the ’70s and early ’80s, Maple assembled what he called “Charts of the Future,” paper maps into which he stuck color-coded pins to track crime. It sounds simple, and fairly obvious. But no one in New York’s police departments was doing it until Maple, who soon was locking up dozens of gang members and getting promoted to lieutenant.

At the end of 1993, Maple had been hired as the NYPD’s top anti-crime strategist by the incoming police commissioner, Bill Bratton. At night, Maple parked himself at his prime table between the actors, pro athletes, and one priest at Elaine’s and watched closely as the restaurant’s namesake owner monitored every waiter and bar tab. Maple wrote four goals on a napkin: Accurate, timely intelligence. Rapid deployment. Effective tactics. Relentless follow-up and assessment. Those principles were soon disseminated through weekly meetings at police headquarters, called CompStat. Within weeks, crime starting plunging, and it hasn’t stopped.

Jack Maple, Bill Bratton, and Louis Anemone — NYPDers who played key roles in the birth of CompStat. Photo: Steffen Kaplan/The New York Times/Redux

That’s the beloved, mostly true myth, anyway. Maple’s friends admit that “the Jackster,” who died of colon cancer in August 2001, may have embellished a bit of the CompStat origin story. Bratton is quick to tell me that he was using wall charts to identify crime hot spots as a young sergeant in Boston in 1976. Later, hired as New York’s transit police chief under Mayor David Dinkins, Bratton recognized a like-minded ally in Maple. When Bratton returned to the city, as police commissioner for new mayor Rudy Giuliani, Maple was one of his first hires. “Maple’s concepts met Bratton’s systems, and that’s what this is actually all about,” says John Miller, Bratton’s press spokesman in 1994 and today the NYPD’s deputy commissioner for intelligence and counterterrorism.

Drawing up anti-crime principles was easy. When Maple and Bratton moved into police headquarters, though, they encountered an NYPD that compiled crime statistics every six months and had very little interest in changing the way it did business. Corruption scandals in the ’60s and ’70s had created an organization whose main motivation was staying out of trouble. Promotions were based primarily on seniority, not on improving safety.

Bratton wanted to instill a sense of urgency and a belief that the police could actually prevent crime, instead of just reacting after the fact. “I interviewed every precinct commander in the department twice: once under Lee Brown” — Dinkins’s first police commissioner — “and then again under Bratton, in ’94,” says John Linder, a consultant whose work became an integral part of the reforms. “And one of the precinct commanders said to me as CompStat was starting, ‘No one has ever lost a command because of crime going up. You lose it because you’ve got a dirty cop or because you’ve got a pissed-off community.’ Getting the organization to both believe it could fight crime and to focus on street crime was a massive culture change.”

But Bratton and Maple had not worked out a method for injecting their message into the NYPD’s daily operations. First they tried writing and distributing thousands of copies of booklets filled with new procedures tailored to reducing drug sales, car thefts, domestic violence. Many went unread. Maple ordered every precinct to create a crime map, and he would show up unannounced in the middle of the night to inspect the cops’ handiwork.

From here, memories diverge. In Miller’s telling, Maple’s fatigue produced the brainstorm to bring the precinct commanders to him, instead of the other way around. According to Gene Whyte, then a sergeant working closely with Maple, frustration at trying to schedule meetings out in the boroughs instigated the first gathering at headquarters. Regardless: On April 6, 1994, a group of very wary commanders from Brooklyn North shuffled into the press briefing room at One Police Plaza. One cop was selected, at random, to report on his precinct. “This poor fellow got up and started discussing his precinct as if he was talking to a community group: How many square miles it is. How many civic groups are in it,” Whyte says. “And Maple yells, ‘How many GLAs?’ That’s grand larcenies” — specifically, grand larceny auto. Louis Anemone, the squat, feisty chief of patrol installed by Bratton, cut in. “As soon as Maple put his hand down,” Whyte says, “Louie’s hand went up: ‘How about your robberies? Tell me about your robberies. How many patterns? What hours in a day do your guys work?’” The cop fumbled for answers. “Anemone stood up, waved his arms, and said, ‘That’s it! I’ve heard enough. You’re going to come back next week and we’re going to talk about crime.’”

Maple and Anemone were told there was no way crime statistics could be assembled daily. Yet the numbers soon started arriving — by fax, or driven downtown — under threat of transfers. Maple picked the brain of an underused administrative lieutenant, Billy Gorta, and drafted an eager junior staffer, John Yohe, to modify an existing computer program to analyze the data on an Hewlett Packard 386.

The sessions took on a structure gradually, with Maple and Anemone using the data to identify lagging precincts. But the gatherings quickly grew crowded as more personnel were required to attend, and in January 1995, moved to the eighth-floor command center. CompStat also became riveting, if uncomfortable, theater every Wednesday and Friday morning. Maple arranged for a slide of Pinocchio to be projected behind a commander whom he distrusted. “John Yohe would use the computer to draw a circle around four or five of these dots up on the screen in, say, the 44th Precinct,” Anemone says. “I’d say to the captain, ‘These are robberies.’ ‘Yessir, yes they are.’ ‘What can you tell us about them?’ ‘What do you mean?’ ‘Are they connected? Are they the same perpetrator or are they different perpetrators. What’s the attraction for this area of your precinct? What are you doing about this?’ ‘Well, I don’t have enough personnel,’ blah blah blah.” Maple and Anemone often knew more about what was going on inside individual precincts than the lieutenants and captains in charge of policing those neighborhoods. But embarrassing people wasn’t the point. “What CompStat brought was accountability,” Bratton says. “It was hated by managers and cops alike for the fact that they are being held accountable, when for 20 years prior they hadn’t been.”

CompStat also became a means to implement a new theory: Broken Windows. Developed by criminologists James Q. Wilson and George Kelling, a Bratton adviser, and published in an article in The Atlantic in 1982, Broken Windows held that ignoring small violations of the law, or quality-of-life offenses, led to larger crimes and to increasing disorder. Maple didn’t read the academics’ theories until he’d left the NYPD, but he’d been practicing their precepts years earlier, in the subway. Cracking down on turnstile jumpers, he and Bratton found, had a multiplier effect, depressing robberies and assaults. Crucially, the gains were not simply a result of arresting larger numbers of fare-beaters, but in following up on the arrests and discovering that many of those in custody had outstanding warrants for more serious crimes.

When Bratton applied those ideas and tactics across the entire NYPD, and CompStat was launched, the results were dramatic: Major crimes dropped 39 percent in 27 months. How much credit should go to the cops and how much to coincidence has been argued ever since. “Arrest rates for less-serious offenses did contribute to the crime drop, but not more than 5 percent,” says Rosenfeld, the Missouri criminologist. “The larger contribution was Bratton’s and Maple’s insistence that commanders be held responsible for what was happening in their precincts.”

Linder, who later became Maple’s business partner in exporting CompStat to other cities, says that was the main idea from the outset. “Jack never saw a mystical connection between CompStat and Broken Windows,” he says. “And he took extreme exception to the idea that New York’s model was ‘zero tolerance.’ Because you weren’t just enforcing everything that was on the books — you were enforcing things that would lead to the reduction of violence.” The goal was to aggressively tackle quality-of-life crimes while using discretion.

The NYPD, newly energized by Bratton, Maple, and CompStat — and deploying the thousands of new cops whose hiring had begun under Dinkins — began a long winning streak. Murders went from 1,946 in 1993 to 673 in 2000; car thefts went from 111,611 to 35,422; rapes dropped from 2,813 to 2,068. The number of city residents in state prisons, meanwhile, crept to a peak of 47,315 in 1998 before steadily declining to about 23,000 in 2016. Misdemeanor arrests, particularly for drugs, rose sharply under Bratton and his immediate successor as commissioner, plunged for a few years, and then jumped again in the 2000s, from about 60,000 to 80,000 annually.

In his new book Uneasy Peace, NYU sociology department chairman Patrick Sharkey points to a more aggressive criminal justice system as one major factor in the crime drop. “But two elements have been underappreciated,” he says, “the expansion of private security efforts — from LoJack to business improvement districts and especially the explosion of nonprofit agencies confronting violence and focused on building stronger neighborhoods.” The cops have also been a significant piece of the equation,” Sharkey says. “The more directed and precise policing is, targeting places and people that account for a disproportionate share of violence, we have good and clear evidence it reduces crime. In New York policing was disorganized and scattershot before the emergence of CompStat.”

Success inflated egos and inflamed jealousies, however, with Giuliani pushing out Bratton in March 1996. Maple quit in solidarity. Anemone stayed, however, partly out of worry that CompStat and other department reforms, still new, would be rejected, like a transplanted organ, by the NYPD’s old guard and its new boss, the former fire commissioner. “I had to convince Howard Safir,” Anemone says. “He thought CompStat was smoke-and-mirrors. He certainly didn’t know very much about what we did as an agency. So he sat alongside me for five and a half weeks of CompStat meetings.”

Giuliani was a key ally: The mayor loved the drop in crime, if not Bratton, and was proud to tout CompStat as a breakthrough by his administration. When Vice-President Al Gore came to watch a CompStat session, Giuliani arrived early and shuffled the nameplates at the head table to place himself in a more prominent seat. “Childish,” Anemone says. But the meetings survived the politics.


The attacks of September 11, 2001, made Michael Bloomberg mayor and, in turn, Ray Kelly police commissioner for a second time (he was commissioner under Dinkins from ’92 to ’94). Kelly, by necessity, devoted a great deal of energy and attention to successfully ramping up the NYPD’s anti-terrorism capabilities. He also established the Real Time Crime Center, which uses surveillance cameras and the databases that feed CompStat to rapidly assemble information and distribute it to detectives immediately after a crime, or while a crime is still in progress.

CompStat itself, however, quietly chugged along basically unaltered — which seemed wise, as the city’s crime rate continued to decrease. Murders dropped from 587 in 2002 to 419 in 2012; rapes went from 2,144 to 1,445 and burglaries from 31,275 to 19,168 during the same period. Yet one seemingly minor change in the process helped to eventually produce toxic side effects.

It’s all about the numbers. Photo: New York Daily News Archive/NY Daily News via Getty Images

Cops had stopped, questioned, and frisked thousands of New Yorkers during Bratton’s run in the mid-’90s. But the numbers were not rigorously compiled, and they were not included on the CompStat menu. “We’d always avoided making stops a part of CompStat,” Bratton says. “That type of activity didn’t have a place in CompStat. It was encouraging the precinct commanders to feel that they wanted more numbers.”

Mike Farrell, who began his career under Bratton and rose to become a deputy commissioner under Kelly, scoffs at Bratton’s account as revisionist history. “It had nothing to do with a strategic choice,” he says. “Those numbers simply weren’t collected.” A lawsuit growing out of the fatal 1999 shooting of Amadou Diallo, however, forced the department to start keeping accurate count of stops, with the process kicking in during the first years of the Bloomberg administration. “The rationale to incorporate the stops statistics into CompStat,” Farrell says, “was, ‘What are all of the indicators that reflect enforcement activity in the command?’”

But over time the number of stops exploded. Partly by design — Kelly believed they were an effective means of crime-fighting — and partly by institutional momentum, they more than quadrupled from 2002 to 2006, from 97,296 to 506,491, took a brief dip, and then climbed to 685,724 in 2011. “The name comes from ‘compare statistics,’ and they were always comparing to the previous year and trying to beat it,” says John Eterno, a former NYPD officer turned Molloy College criminologist, who co-wrote the book The Crime Numbers Game: Management by Manipulation with Eli Silverman. “There was enormous pressure on the captains to meet quotas and cover themselves at CompStat. And stop-and-frisk was a key tool. If they got questioned about a robbery, they could say, ‘We did ten stop-and-frisks in that area.’”

Farrell rejects that analysis. “I spent 17 and a half years going to CompStat meetings, and from every chief of department there were constant admonitions, ‘This isn’t about the numbers,’” he says. “There was no drumbeat for ever-increasing numbers of stops or summonses or arrests. It was simply, ‘What do those numbers reflect for a given command at a given point in time?’ And commanders went with whatever tactics the crime situation they were confronting demanded.”

Fueling the growing outrage was who was being stopped: innocent people (around 80 percent of the time) and black and Latino New Yorkers (roughly 85 percent). Kelly has consistently maintained that there was no racial calculation involved: He simply sent cops to where crime was concentrated, and that happened to be in poorer neighborhoods where the suspects, as well as the victims, were predominantly people of color. Kelly also attributes what became the furor over stop-and-frisk to stricter record-keeping and to opportunism by political opponents of the Bloomberg administration. “Stops were always underreported when I was a cop,” Kelly says. “What we did was require forms to be made out. The highest number of stops amounted to less than one stop a week per patrol officer. Everything is bigger in New York and that number was never put in perspective. I think the issue was skillfully managed by the plaintiffs in those lawsuits against the city, and Judge Shira Scheindlin wanted the case. Her decision was disgraceful, and we know it would have been overturned.” In 2013 Scheindlin, a federal court judge, ruled that the city’s stop-and-frisk program was being applied unconstitutionally, calling it a “policy of indirect racial profiling.” The Bloomberg administration appealed and was granted a stay; de Blasio, in his first month as mayor, agreed to accept a federal monitor and other reforms. He later dropped the appeal.

In some ways CompStat had become a victim of its own success. “The prevailing wisdom had been that the police are minor actors, and socioeconomic conditions are the root causes of crime,” says Silverman, a John Jay College professor of criminal justice who spent years studying CompStat’s evolution. “But with Bratton and then Kelly, not only can the police control crime but now the police must control crime, because the political and the police leadership have committed themselves to it. So during the Kelly era there was so much pressure to get the right numbers, and an assumption that if cops reduced their ‘activity’ then crime would go up.”

Ironically, after seizing on stop-and-frisk as a campaign issue and winning a major upset in the Democratic mayoral primary in 2013, de Blasio’s first big hire was Bratton, whose systems had in some respects created the problems he ran on. In Bratton’s second tour as commissioner the challenge wasn’t reducing crime so much as keeping it down, and making nice with the neighborhoods. Again, though, he used CompStat as a primary vehicle in reforming the NYPD. He elevated James O’Neill, who had been stuck in middle management during the Kelly years, to run the meetings. “I think we, as a department, did lose our way for a while,” O’Neill says. “We started talking more about the quantity of enforcement instead of quality, which damaged us in the community. I tried to bring more focus to the violence. It wasn’t just about how many summonses you wrote or how many people you arrested. Who are the people involved in crime and violence? That was an evolution that took place when Dermot and I started running CompStat.”

The number of stops plunged in Kelly’s final two years as commissioner and have kept falling since de Blasio became mayor, with just 12,404 recorded in 2016, the last full year for which statistics are available (though in December 2017, a federal monitor said cops appear to be underreporting street stops). Maybe conditions have changed and they aren’t necessary anymore — or maybe the continuing decline in crime is an admission they were never necessary in such high volume.


Kelly had increased the focus on gangs with an initiative called “Operation Crew Cut.” Bratton expanded the idea, in part by moving the NYPD toward “predictive policing,” a concept he’d started developing while chief of the Los Angeles police department. “It’s an outgrowth of CompStat into the 21st century,” Bratton says. “With the development of algorithms, we’re able to look at all this crime information in addition to a visual representation on the map. Now, with the technology, we have the ability to say, ‘Cops need to be in this area during this time, there’s a good likelihood a burglary is going to be committed, and here’s who is likely to be committing it.’” Could cops soon be using the data to preemptively remove people from the streets? “The ACLU types are worried about the use of artificial intelligence,” Bratton says. “By being able to identify the 5,000 prolific criminals, we can focus attention on them and not be bothering 700,000 innocents.” He’s even more dismissive of research that claims demographics and the economy, not the NYPD, are responsible for the reduction in crime. “The academics arguing that it was lead paint, or — who’s the character that wrote Freakonomics — that it was abortion that solved the crime problem?” he says. “Well, if that was the case, how come birth control, which came in 20 years before, didn’t solve crime? No. It was the essentiality of CompStat and all of its appendages.”

The abortion and lead-paint theories have indeed gotten mixed reviews from criminologists. Yet there is a fairly solid consensus that factors including the end of the crack boom, the Clinton-era addition of 100,000 cops to the streets, and gentrification have been significant factors in reducing crime rates in multiple cities. “The urgency now is to develop a new model to keep violence low,” says Sharkey, the NYU sociologist. He is encouraged by CompStat’s latest local spinoff. “Under the mayor’s office of criminal justice, NeighborhoodStat brings in the PD, the housing authority, sanitation, community groups, with the same model as CompStat — putting data on the big screen and figuring out who needs to step up,” Sharkey says. “It’s the next stage of CompStat, and maybe the most important shift in policing.”


O’Neill succeeded Bratton as commissioner in September 2016, and he turned CompStat over to Shea and Monahan (who was recently promoted to chief of department). Part of their push to update the system is a “sentiment meter,” based on polling conducted by Linder, Maple’s one-time business partner. “The idea is to be able to get an indication, in real time, sector by sector, of how people feel about their safety and how they feel about the NYPD,” O’Neill says. “It’s in beta right now. It’s not fully developed.” Kelly, for one, is highly skeptical that the data will ever be dependable or relevant. “You don’t want hostility between police and average citizens. But this is a unique business,” the former commissioner says. “You’re going to have some hostility, because of what society asks police to do. They are the bearers of bad news. They write traffic summonses. They are authorized to use deadly force. Sometimes they kill people. This is not a job description that makes people love you.”

Public sentiment may someday become central to the NYPD’s operations. At the moment, Shea uses CompStat to hammer home the importance of communication and coordination, of not just mapping the dots but making connections between them, in order to build strong cases that will lock up the 5,000 or so criminals the NYPD believes are responsible for the bulk of the city’s violence.

So on this January morning when Shea orders the 45th Precinct, which covers Throgs Neck and City Island, to the podium, he’s pleased to see seven people — representing everything from the warrants to the narcotics to the public housing squads — crowding around a single microphone, with one cop picking up where the previous leaves off. “You weren’t expecting to get called on today, were you?” Shea says mischievously. “We’re ready!” the precinct’s executive officer says.

“This gang that was supposed to be inactive — I’m not so sure,” Shea says. “What are we seeing — slashings, shots fired?” Deputy inspector Julie Morrill jumps in. “There was a recent arrest made up in Maine of a couple of our guys,” she says. They’d gone north to sell crack and led state police on a 44-mile car chase. The precinct’s detective squad sergeant chimes in with the good news that an associate of the interstate crack dealers has been arrested on an outstanding warrant and returned to Brooklyn. More troubling is a report of shots fired outside a day-care center. “Double down on the debriefings,” Shea says. “Real nice analysis, but there is a lot of smoke here. There are some bad people involved here.”

He sends everyone back to the Bronx with words of praise and thanks. Shea is genuinely pleased by how far crime has fallen, and optimistic that CompStat can take advantage of the peace dividend to improve community relations. But something from two hours ago keeps gnawing at him: the mention of a shooting suspect who has been arrested 16 times and is currently out on probation. “There’s so much pressure, at every level, not to incarcerate people,” he tells me afterward. “To close Rikers, to put people in programs instead of in jail. Fine. But then what? There’s 2,200 people out on probation right now and no one knows where they are. We had two shot the other night in the 9-0, in Brooklyn. One dead, one not dead. The guy that’s killed, he’d been arrested within the last six months for a shooting. Uncooperative, no prosecution. The guy that’s hit and doesn’t die was arrested last year for shooting into a crowd. Pleads guilty to possession of a gun, put on probation. No jail. He’s shot last night. He knows exactly who shot him. We’ll work with him and we’ll work with the prosecutors. But this is the cycle. A huge percentage of shootings are done by people on parole and probation.” Which is why, for all the time he spent dissecting cases with cops today, Shea’s angriest words were often intended to be heard by two other audiences: the bosses of the probation department, who watched the meeting on video upstairs, and a lawyer from the Bronx district attorney’s office, whom Shea had invited to sit in on this CompStat session. “They love me in the DA’s office,” Shea says with a sly smile. “And I love them. But we need help.”

Mounted on a shelf directly behind him is a display: Jack Maple’s homburg and one of his bow ties. Framed life-size photos of Maple — with a subway anti-gang squad; surrounded by crime-scene tape — are hung on the conference room walls. Above the podium where groups of cops were grilled are Maple’s words, inscribed in wood in letters two feet high: “We will be relentless until New York is in fact the safest city in America.”

Maple launched this meeting in 1994 in part because he couldn’t get quick access to crime information. Cops now have instant access, through their smartphones, to most of the data and analysis Shea just spent three hours dissecting. Which makes it seem anachronistic and inefficient to continue dragging 200 top cops into headquarters every week. Is the day coming when CompStat will be obsolete? “I don’t see it,” Shea says. “We’re still in the human business. And there’s nothing like looking people in the eye to get the message across.”

Order Highbrow, Lowbrow, Brilliant, Despicable: 50 Years of New York, a celebratory book chronicling the magazine’s history with powerful images and behind-the-scenes stories from staff and subjects.

The Crime-Fighting Program That Changed New York Forever