Philip F. Lee, 5/25/03 (Rev. 9/22/08)
In the book Ballistic Imaging, Daniel L. Cork, John E. Rolph, Eugene S. Meieran, and Carol V. Petrie, Editors, Committee to Assess the Feasibility, Accuracy and Technical Capability of a National Ballistics Database, National Research Council, 2008, the authors concluded "A national reference ballistic image database of all new and imported guns is not advisable at this time" (see page 6 of the linked summary). The reasons for their conclusion in 2008 came from the same problems with automated matching in large databases identified by California.
So, California experimented with “ballistic fingerprinting” technology and found that it is not likely to be effective and they decided to not adopt such a program because of the expected poor performance. This performance expectation of the automated IBIS ballistic fingerprinting system has certainly been proven in practice by its failures in Maryland and New York. Of particular interest are the findings from an independent Dutch ballistics expert Attachment D employed by California to review its testing process and conclusions about the effectiveness of the IBIS system. This expert explains why IBIS technology will not be useful in automating the matching of criminal evidence in a large database of legally sold gun shell casings. (California test details and other bits of information are given in Appendices A through F which are available via the Internet Archive.)
Without first studying the problem and the capability of the technology, Maryland and New York adopted laws in 2000 requiring new handguns sold in retail to have a shell casing submitted to the state for ballistic fingerprinting with the hope that shell casings recovered at the crime scene can be linked to the criminal responsible. Unlike Maryland and New York, California decided to study the problem before passing a law.
The IBIS processing of collected shell casings in New York and Maryland have produced practically no results at a cost of millions of dollars which could have been more productively spent on additional police, prosecutors, or prison facilities instead. Automated “Ballistic fingerprinting” performance is low because the technology to make it practical doesn't exist. And, if new technology is developed tomorrow, the chances are that it will not be compatible with IBIS and the investments made in Maryland and NY will be wasted.
For “ballistic fingerprinting” of shell casings to identify the criminal user of a firearm, the criminal must fire the firearm and leave a shell casing to be recovered at the crime scene (revolvers don't automatically leave shell casings so criminal use of revolvers won't be threatened even by effective technology. This crime scene shell casing must then be linked by comparison to the gun in the legal retail sales data base, and the retail record must link to the user either directly or through investigations.
“Ballistic fingerprinting” programs will rarely aid investigation of shooting crimes, will not serve to deter criminal use of guns and will not be cost effective since:
1) The technology doesn't work even in ideal conditions! California reached that conclusion (see CAAGR), by test firing a total of 2,000 rounds from 790 different pistols of the same model. When the shell casings used with a particular gun came from the same manufacturer, computer matching failed 38 percent of the time. When the casings came from different manufacturers, the failure rate rose to 62 percent. Normal wear and tear was a minimum in these cases since the test rounds were fired immediately after the round used to create the “ballistic fingerprint. Normal wear of the mechanical parts making the marks used in "ballistic fingerprinting" will increase real world failure.” Nor was this failure limited to the computer matching tool – 16% of the shell cases could not be matched even by an expert examiner with more than 30 years experience and approximately 42% of the remaining cases had unfavorable markings which might have prevented matching by the expert.
2) Computer matching of ballistic fingerprints has a high false match rate requiring trained ballistic examiners to make the final determination from a number of candidates. False matching by the computer increases with the size of the database of ballistic fingerprints and increases the work for human examiners. That means more examiners will need to be trained and employed and that is a cost ignored by proponents of the technology. The California Attorney General concluded that current “ballistic fingerprint” technology is too costly and ineffective for comprehensive application on a database of legally sold firearms.
3) Criminals frequently alter firearm serial numbers to prevent tracing, and could alter the ballistic fingerprint of a gun in a few minutes (3 minutes according to CAAGR, Appendix F).
4) Criminals using revolvers do not normally leave behind shell casings at crime scenes. It is wasteful to collect such information from revolvers as Maryland does (we note that the number one gun used in crime in is a Smith & Wesson .38 revolver so this technology will be relatively ineffective against the gun most frequently used by criminals).
5) 500,000 guns are estimated to be stolen annually (see Comprehensive Firearms Tracing: Strategic and Investigative Uses of New Data on Firearms Markets, Philip Cook & Anthony Braga, University of Arizona Law Review 43, no. 2 (Summer 2001) and by definition these guns are available to the criminal market. Such guns cannot be traced back to the users by ballistic fingerprinting. Approximately 40% of guns used by criminals were stolen or bought on the illegal street market (especially criminal recidivists are more inclined to use this source -- see Firearm Use by Offenders, Bureau of Justice Statistics Special Report, November 2001, NCJ 189369).
6) Crimes are frequently committed with illegal firearms, that is, the firearms are unregistered, stolen, or modified in some way (see The Licensing and Registration Status of Firearms Used in Homicide, Jenny Mouzos, Public Policy Series, No. 151, Aust. Inst. of Criminology, Canberra). Hawaii, which has had registration and licensing of guns for several decades, has spent hundreds of thousands of man-hours administering these laws. This massive effort has not produced a single case with police claiming that licensing and registration were instrumental in identifying a criminal.
7) Maryland, despite its ballistic fingerprinting of handguns, has continued to have the highest rate of robbery of any state (see MarylandCrime7.doc), so suggesting that criminals are deterred by “ballistic fingerprinting” is to hope against evidence.
crime fighting technologies lack money. For example, a recent Rand report notes the problem with
“The results of DNA testing on violent crimes is
backlogged, causing delays in the freeing of suspects or prosecution. The delay in processing toxicology tests is
hindering benefactors from settling insurance claims and estates of loved ones
that have died. Without toxicology
reports, coroners cannot issue death certificates required by insurance
companies.” See Challenges and Choices
for Crime-Fighting Technology: Federal Support of State and Local Law
Enforcement, William Schwabe, Lois M. Davis, Brian A. Jackson, Rand, MR 1349.
Phil Lee on 7/22/10. Contact maryland_alert at Yahoo dot com (sorry for being obscure, but web mail address scavenge programs make this practice necessary).