The only interesting question is whether the strategic position of the jihadis -- al Qaeda and its cognates -- is improving or degrading. To extend your historical, the Germans killed many more Allied soldiers in 1942 than in 1940, but 1940 was a year of German ascendancy and 1942 saw German "turning point" defeats at El Alamien and Stalingrad.

As to the first question, it seems that al Qaeda's situation must be getting worse. It has no easy refuge, its command and control systems are under intensive surveillance, Muslim governments that were on the fence have flipped to our side (Pakistan and Saudi Arabia), and al Qaeda is burning soldiers and treasure in the Sunni triangle. We are vexing and exhausting al Qaeda, rather than the other way around.

This whole thing is a deeply flawed study and two surface observations make this whole idiocy quite clear:

1. They have Abu Nidal, the most infamous terrorist of the 1980s, dying in November rather than in August 2002.

2. They list only one of the two Russian planes that got blown up by suicide bombers back in August 2004 as international terrorism, and only then because there happened to be a Jewish passenger on board. This isn't flawed methodology, it's just stupid.

And to think I argued for the damn thing's release ...

Thank you Dan

Reference my post in NCTC Terrorism Study: Huh??

A.L.
I can understand why you are feeling under the weather. What I can't understand is why you would even consider working with flawed data.

I can understand why you are feeling under the weather. What I can't understand is why you would even consider working with flawed data.

There's no such thing as a perfect dataset, and you work with what you've got. Nearly all data has mistakes and flaws, and you just hope they're not too great and that they aren't systematic. Besides, although there may be omissions and mistakes that require fixing, it's probably not that far off. Studies produced by other sources have suggested similar results.

The thing is, the metric itself means almost nothing in isolation. 4 out of 5 attacks were in just 2 countries? And as A.L. says, not just two countries but two small geographic regions, mostly. Kashmir and the Baghdad Metro Area.

It means the fight is concentrated, and we're up against a determined enemy who is losing ground slowly and making the good guys pay a high price. And it also probably means that, well I don't know. My guess (and Dan can comment on this) is that it appears that a segment of the Jihadis, especially those in Pakistan and some who evacuated Afghanistan recently, may have targeted Kashmir, either as a potential Islamist State, or as a powder keg to ignite a larger struggle between India and Pakistan so that they can get their mitts on a nuke.

But that's not my area of expertise, and I don't think the Sikhs would just lie down. The Vale is one of the most beautiful places on Earth, and the fact that there's so much strife there is a huge tragedy.

Demosophist

"There's no such thing as a perfect dataset, and you work with what you've got. Nearly all data has mistakes and flaws, and you just hope they're not too great and that they aren't systematic. Besides, although there may be omissions and mistakes that require fixing, it's probably not that far off. Studies produced by other sources have suggested similar results."

My problem with the A.L.'s analysis of the data is the report explicitly states the following:

"The methodology used to generate this list of incidents, however, does not accurately capture the totality of terrorist incidents worldwide and could lead to anomalous and potentially inaccurate results."

A.L. does not inform the reader of the potential flaws in his analysis due to the report. Given the statement contained in the report, to me the error in using this data for analysis is far greater than the studies by other sources you refer to.

IOW GIGO - Garbage in garbage out. An old but true adage in the computing industry.

I've responded to Larry's rather discourteous counter-strike over on Jawa, in the comments section. But at least he's not some aloof old poll. Kudos for engaging.

USMC:

I'm not sure the data is "garbage," but any conclusions from it should certainly be tempered. On the other hand it's bound to be analyzed by the other side in the advocacy game anyway, good data or garbage, so conducting this analysis is like pointing out that even their analysis of bad data is colored by their advocacy position. And beyond the fact that the war has concentrated and intensified, I don't think there's much this data can tell us anyway. Again, we need more qualitative measures, or rather more measures of quality.

By the way, although he doesn't state it explicitly, I think what Larry is worried about is the possibility that a large number of small but lethal attacks now is an indication that there are a lot of people in training to produce bigger attacks later. He could be right about that worry. Dan would know more about that though. And what Larry doesn't seem to understand is that there's nothing about these data that make that concern more likely (even though they don't rule it out).

USMC:

One more thing to note that's true of dealing with any less-than-perfect database is what direction the errors and inaccuracies are likely to be. In this case the odds are much greater that the number of terrorist incidents was undercounted, than overcounted. It's just that there may be a trend to that lack of reliability, and past assessments probably undercounted too. In fact they probably undercounted more.

But again, it's probably not a good idea to make too much out of any analysis until someone has "cleaned the data" properly.

Tracking terrorism by "numbers of attacks" seems to assume that terrorism is a sort of spontanteous phenomenon, like road rage incidents in hot weather. 10 such incidents are cause for concern, but one is no big deal.

It's particularly poorly suited to evaluating Al Qaeda, which seems to be interested in the quality of their violence, not the quantity. There was ONE terrorist attack in the US in September 2001 - just ONE - which was preceded by years of planning.

Of course, this is supposed to be just one perspective on terrorism data, hopefully to be supplemented by many others. But this bean-counting seems to be exactly the sort of thing that middle-management bureaucratic drones love to seize on, to the exclusion of everything else: "Look, sales are up 50% from the the last Meaningless Period, and litigation costs are down 75%! Great job, everybody!"

The other analogy that occurs to me are the sad people who count acts of violence in Looney Tunes. "The coyote's dynamite goes off prematurely, causing him to fly up and hit his head on the rock ledge, after which he falls to the bottom of the canyon. Then the ledge breaks off and lands on him. That's four acts of violence in fifteen seconds ..."

A.L.:

Division of Labor

I'll do two months. The last two unless you have another preference. That leaves 8, I think.

There are good reasons to do this analysis, even though the data is far from ideal. Note how lethal the Iraq attacks are, compared to those in Kashmir. Gosh, I wonder why?

I don't see why Kashmir is counted and the Northern Caucussus isn't. I think you could even claim that there is no international terrorisme in Israel & the occupied territories as it is ruled under one goverment. The Hawks of both groups also claim ownership of the whole

I also wonder why the van Gogh's murder isn't counted as the killer had also a Marocan passport and he belonged to a international group.

Iraqi attacks are more lethal because they have more weapons and the number of attacks are undercounted