A bug / feature of human thinking is the tendency to discern meaningful patterns in random or unrelated phenomena - a suspicion, gut feeling, a hunch. The problem with suspicions is that some people in some circumstances have difficulty in separating suspicions from evidence or even proof of the thing that they were suspicious of in the first place. Instead of taking the suspicion as a hypotheses to be tested, they jump over that part and go straight to the conclusion.
Modern law enforcement agencies, doubtless including MI6, train investigators about the cognitive bias to help them to avoid the pitfalls of jumping from suspicion to conclusion.
An interesting article entitled
Criminal Investigative Failures: Avoiding the Pitfalls by D. Kim Rossmo, Ph. D. was published in two parts in the
FBI Law Enforcement Bulletin It is an easy read and goes over cognitive biases, probability errors, and organizational traps that can lead to errors in the investigation of crimes.
It is worth reading both parts in their entirety. Here are some points that struck me as being particularly relevant to AK/RS. They would, of course, given my own biases.

Part One - Cognitive Biases from
September 2006 Volume 75 Number 9
Part Two - Probability Errors from
October 2006 Volume 75 Number 10
PIP will like the Prosecutor's Fallacy:
PGP will enjoy the defence attorney's fallacy. The article goes on to use the OJ Simpson case as an example.
The article wraps up with the group dynamics or "Organizational Traps" that can make it difficult for LE to abandon an initial theory that is not supported by emerging evidence. There is a particularly interesting section on groupthink.
Why do you think they turned up? I thought they were returning some phones.