Discussion about this post

User's avatar
Kirk Watson's avatar

I believe Arthur Andersen went belly up on just the charging documents. Could we be so lucky with the SPLC?

I have empirically found that a small amount smoke (raveling thread) always leads to deeper problems. One has to tease them all out until the last problem is found. True about buildings, people and institutions.

BehaviorForecastsProbablyHard's avatar

Yeah, I read that BAM, and had thoughts and disagreements, mostly too boring to bring up with the man directly.

There's basically the narrow view of the SPLC the guy argues, that more or less gives them a win in accepting their history at face value, and focusing on the electioneering inference.

Stepping a bit further, there is a reason why the SPLC might be a unique 'intelligence' provider, and that is because it is not. This argument is that akshully, the SPLC's origin is in laundering traces of political violence, and that its 'electioneering' is directly a terrorist conspiracy. IE, it was originally a southern Democrat organ, and by the way it presented the KKK, whitewashing any institutional involvement or cover ups by state Democrat Parties. (There is also the current day argument that CRT is effectively a white nationalist ideology, and that BLM a white supremacist terrorist organization.) The conclusion of this argument is that the government only wanted to fund and sanction one conspiracy of 'intelligence providers' and activists, because if there were multiple intel organizations they might have more than one conclusion, and some conclusions are very inconvenient.

Intelligence analysis generally is a very uncertain, crazy making business, and it admit methods that make huge trades in false positive rate versus false negative rate.

In general, in statistics the detection hypothesis compares two distributions, and the math forces trades between true and false positives and negatives. Behavioral detections are inherently subjective, because everyone uses themself as a naive estimator, and there is no way to back that out to model distributions that are truly objective and portable between analysts. This is because every analyst is an individual who bring a unique set of priors. You can get similar analysts with overlapping priors, and overlapping estimates of distributions, and this can be a good thing, or a bad thing.

Intel organization selection and management winds up being an art, that is very easy to break. It is also 'easy' to infer when people are deliberately breaking the organization to give a predetermined and politically expedient answer. 'easy' is in scare quotes, because you get true and false positives.

Military intelligence, and criminal intelligence, wants certain sub organizations to have a high false positive rate, because they actions they use to fix things have delay and inertia, and and because they want a very low false negative rate. The high false rate of detection gets fixed a bit with the tracking organization, and with 'we are still waiting for the soviets to come through fulda gap' type wait states.

So, the minimal explanation of problem with what the SPLC is doing is that they inject noise to /cause/ their false detections. Which itself is not bad, unless we are very expensively trying to fix such detections, or to prevent mass murders that are not in fact possible.

However, there is a much more damaging explanation for what the SPLC in particular is doing, in more general combination with the universities.

BAM describes two groups of people, bankers and activists. Both are described as believing similar ideas, which makes sense as they were trained at universities with what would have largely been the same undergraduate course work on political terrorism in US history, the history the US government pays scholars to endorse.

US has three government levels of law enforcement, federal, state, and local, all relevant to domestic terrorist conspiracies. Certain historic domestic terrorist conspiracies, the mainstream government endorsed history more or less directly implicates state and local law enforcement, and selective 'ignorance'. The mainstream history attributes everything to a vague and inchoate racism, and does not consider whether specific politicians had specific goals in mind.

CAse one, a man is accused of rape, rioters burn neighborhoods, and then there is a long period of charges not being brought, and of official investigations not trying obtain testimony.

We can ask ourselves if an organization can be involved in any of these activities, and apply tests.

Were there arsons in 2020? I have not personally seen the sites of any fires, and I am not personally an arson investigator. I feel there probably were, but I do not directly know that, and am curious some number of possible arsons in 2020.

An arson hypothesis is 0, or 1, or 2 organizations. The mainstream narrative is either 0 (it is an idea, not an organization), or 2 (these other people did it).

The subjective statistical inference between 0 and 1 is based in uniformity of methods. Is there the variation expected of random individuals?

The subjective statistical inference between 1 and 2 is location and timing of events. Two actually independent organizations would not synchronize their events, because they would not be operating from a shared body of planning information.

The other statistical inference is the test of partisan correlation. Again, state and local law enforcement is relevant. So you basically get a list of incidents, and plot them in four boxes based on the mayor and governor. Uncorrelated, you would have a pretty even (1) division between the four boxes. If the Republicans are super racist, and did it, then there should be a lot of arsons in Republican governor/Republican mayor cities. But, if it is almost exclusively Democrat/Democrat cities, then basically the hypothesis of reliable intelligence from the SPLC is impeached.

The SPLC is left academic, and their ideas are that left Democrats are not driving anti-black domestic terrorism.

I have outlined how someone looking at the terrorism intelligence problem from first principles might come up with answers that are much less convenient.

Anyway, case two, a man is accused of murder, rioters burn neighborhoods, and there is a long period of charges not being brought, etc.

The charges against the SPLC are right on the edge of being too late for RICO.

(1) Okay, how does one normalize to what expected value? Still, if the Republicans are so bad, there are apparently a lot of blacks in Alabama and Mississippi, and I don't know of many arsons targeting them in 2020.

4 more comments...

No posts

Ready for more?