About five years ago, I began to write occasional articles that questioned the validity of New York City's crime statistics reported through CompStat, a vaunted crime-fighting tool that has been emulated across the country. I noted the discrepancy between publicly reported crimes on the FBI's index of seven major crimes and the comparable lesser crimes that received little attention - for example, serious assaults vs. misdemeanor assaults, or grand larceny vs. petit larceny. All the numbers dropped sharply the first two years after CompStat was introduced, but after that golden era, the "index" crimes kept dropping sharply while the non-index offenses leveled off. For example, from CompStat's first full year in 1995 to 2002, "index" or aggravated assaults dropped 33 percent, while non-index, or lesser, assaults stayed about level, dropping 3 percent. This was consistent with what the city's police union had charged in 2004: that crimes were being routinely downgraded. Crime was down, for sure, but not as much as the numbers showed. If the union was right, fraud was endemic. The numbers alone suggested the need for an independent investigation.After 2002, the NYPD refused to release the numbers for non-index crimes. I and others filed numerous Freedom of Information Law requests, but the department flouted the law. (My last appeal wasn't even answered.) But the New York Times recently noted that the numbers were being withheld and filed suit to get this data. Now, under cover of a blizzard, the NYPD has released the numbers.The new numbers for non-index crimes are in a different format from the past, making a precise comparison impossible. But there is still enough information to spot broad trends.The numbers show that the drop in major, publicly reported crimes has continued to outpace the drop in lesser offenses by a significant margin (although not as sharply as in the past). Aggravated assaults dropped 12 percent from 2003 to 2009, while lesser ones remained about level, falling 2 percent. Meanwhile, city hospital records have shown an increasing number of people hospitalized for assaults.Downgrading crimes is just part of the problem. Recent news coverage has shown how crime victims are systematically discouraged from reporting crimes - say, if a commanding officer insists that complaints will only be taken if the victim goes to the precinct station house. This was documented by a police officer in Brooklyn who secretly recorded the roll calls in his precinct. Newspaper reports detailed the scheme.We live in a society that places too much faith in data - school reading scores, earnings reports, and many other performance indicators. It's good to have the data, but when too much rides on the numbers, watch out.Often, the supposed watchdogs are those most likely to rely on questionable data. The editorial board of the New York Daily News is a prime example. It's often astute, but time and again, it blindly attacks those experts who have questioned the credibility of the police crime reports. Now, it claims that the trends in the new numbers just released are "more or less parallel" - that the 12 percent drop in serious assaults is pretty much the same as the 2 percent drop in lesser assaults.I don't think News owner Mort Zuckerman would view such a difference in his circulation figures the same way.