Karl Denninger points out that a statistical analysis of the Inspector General's report on the FBI investigation into President Trump reveals that it's fundamentally impossible.
The IG report, after reading through a good part of it, states that seventeen "errors" were made by the FBI.
May I remind you of an indisputable fact: Errors are randomly distributed.
That is, let's assume you intend to drive at 40mph. If you make an error you will operate your car some of the time at 38mph, and some of the time at 42mph. The errors, if they are actual errors, will be randomly distributed around the correct action. Some of the errors will place you inside the correct action and some of them will be place you outside of the correct action. Approximately an equal number of errors will fall on each side of the correct action; some will help you, some hurt, but given enough errors there should be an approximately equal number in each direction.
If your speed, as measured at 1 minute intervals, is 40, 42, 45, 43, 41, 40, 46, 50, 42 and 45 mph that is not "error." You are instead driving with a floor of 40mph; your intent is to drive at no slower than 40mph.
The probability of an error is 50% in each direction. Therefore half the errors should have been to Trump's benefit.
However, exactly zero of them were to Trump's benefit.
The odds of this being random chance can be computed.
The probability of the first error being to Trump's detriment is 0.5 (50%.)
Each successive "error" is also equally probable to be of either benefit or detriment. So when the second error occurs for them to both be to Trump's detriment by random chance -- that is, if it's an actual error without motivation or bias, is 0.50 * 0.50, or 1 in 4.
For seventeen "errors" to be all in the same direction is 0.5 ^ 17, or exactly 1 in 131,072.
This is not quite as bad of odds as winning the Powerball but it clowns the claim that such were errors and not intentional acts that evince a predetermined goal or desire.
Indeed in a criminal trial should odds of 1 in 131,072 be established that is almost certainly enough for you to be convicted and sent to prison. Remember that the standard in a criminal trial is not "beyond question" or "with absolute proof" -- it is beyond a reasonable doubt.
1 in 131,072 does not admit reasonable doubt.
The FBI and Department of Justice did, beyond reasonable doubt, intentionally target Donald Trump and his associates as a candidate and as President for surveillance and did, with corrupt intent beyond said reasonable doubt, intend to harm both him and his associates.
This did not occur due to error. The manifest weight of the evidence as proved by fundamental, middle-school mathematics, is that this campaign was intentional, it was malicious, and given that it involved knowingly false statements to a court by means of omission it was criminal.
There's more at the link. Highly recommended reading.
I'm not a professional statistician; however, I did study the subject as part of my Masters degree in business, for use in analysis and forecasting. I know enough about it to agree with Mr. Denninger's calculations, and the conclusions he draws from them.
From a statistical point of view, the conclusion is unmistakeable and unavoidable. The FBI did not make "mistakes" or "errors of judgment": it deliberately flouted the law and acted against the interests of candidate, then President, Trump at every turn.
If you were wondering what "insurance policy" FBI agent Peter Strzok discussed with FBI lawyer (and adulterous lover) Lisa Page, I don't think you need to look any further.
The question is, will they, and the FBI, be allowed to get away with it? The IG report suggests that the "deep state" would like to ensure that they do. Attorney General Barr's comments on that report suggest that they may not - a consummation devoutly to be wished by any and all Americans who value the Constitution and laws of this country, irrespective of their political views.