- ISRAEL: Not A Good Sign
- SUPPORT: MOUT For The 21st Century
- ATTRITION: Internet Geeks Have More Choices
- ON POINT: Spy Novels and Whodunnit: North Korea's Criminal Reality Is Intolerable
- PHOTO: Over The Philippine Sea
- BOOK REVIEW: The Campaigns of Sargon II, King of Assyria, 721-705 B.C. (Campaigns and Commanders Series)
- IRAN: Pride, Prejudice and Persecution
- AIR DEFENSE: No Quick Fix For SHORAD
- SPECIAL OPERATIONS: Benghazi Aftermath
- PHOTO: Birds Of A Feather Flock Together
- KOREA: Purging The Dynasty
- INFANTRY: Tech Takes its Toll
- INFORMATION WARFARE: HVIs Wanted Dead Or Alive
- CIC: The Duel of the Two Men, the Two Horses, and the Two Dogs
- PHOTO: Old And New Friends
- BOOK REVIEW: Franklin D. Roosevelt. Vol. II, The War Years, 1939-1945
- BOOK REVIEW: Franklin D. Roosevel, Vol I, Road to the New Deal, 1882-1939
Even before September 11, 2001, counter-terrorism experts sought to use statistical techniques to predict where the next big terror attack would occur. Similar tools had been used to predict criminal activity, and assist in identifying and catching criminals. But in the West, terror attacks were so rare that these predictive techniques were ineffective. That's because these predictive techniques require a critical mass (statistically relevant amount) of data before they become accurate enough to be useful. There were barely enough pro-terrorist individuals, at least in the U.S., to enable some prediction of their activities (forming new cells and planning attacks), but not enough actual attacks to predict when the next one would occur.
There are far more pro-terrorist individuals and small groups out there, than there are those who will actually carry out an attack. This was different in Europe, where there are ten times as many Moslems, and a larger percentage of them are hostile to the West. In Europe, the predictive techniques led to many planned attacks being detected and disrupted. This was particularly true for attacks that were highly likely to be carried out. It took the Europeans a few years to realize how serious the threat was, and adjust their statistical models accordingly. There were several attacks in the meantime, which provided the motivation to apply the statistical techniques in a more serious, and practical, manner.
In the United States, these techniques still suffer from a shortage of data (terrorists.) With enough data, you can test your model by successfully predicting the past, and then turn it on the future. But with insufficient data, you have to rely on human judgment. This is subject to other factors, like the political atmosphere. An example of this was the recent terror attack in Fort Hood, Texas. There, a Moslem army officer, shouting "God Is Great", murdered 13 soldiers and civilians, and wounded over thirty others. The major had previously been detected by the counter-terror intelligence system (both via emails to known terrorists and his public calls for attacks on non-Moslems.) When the FBI (which handles counter-terror intelligence inside the U.S.) urged the army to do something, the army declined. The FBI did not press the matter. One can imagine army commanders, confronting what the FBI described as a "potential" terrorist, realizing that in the current political climate, disciplining (or discharging) a Moslem army officer would endanger the careers of the generals involved in such a decision. So nothing was done, until the terrorist made his move.