Methodology v7

How the Bluff & Spin Index works

Bluff & Spin Index: how well claims hold up, and how often questions get a straight answer. 0 to 100. Lower is better. 0 means every checked claim held up and every question got a straight answer; the more bluff and spin, the higher the number.

Accuracy

Share of checked claims that the evidence supports. Partly supported counts half. Can't verify is left out.

Bluff · All content

Consistency

Context

Share of claims that do not contradict an earlier statement. An acknowledged change of position does not count against them.

Shown, not in the score · All content

Answers questions

Share of questions answered, judged once per question and full reply. Partly answered and Declined openly count half.

Spin · Interviews, debates, PMQs, hearings

Substance

Context

When they use emotion, how often it comes with a real answer or evidence. Emotion that fits the subject is never penalised.

Shown, not in the score · All content

Fact-check verdicts

Supported	1	The evidence backs the claim.
Partly supported	½	Part of the claim holds up; part does not.
Not supported	0	The evidence does not back the claim.
Contradicted by evidence	0	The evidence shows the claim is wrong.
Can't verify	Left out	Not enough evidence either way. The default when sources are thin. Shown as a separate count.

Question outcomes

Answered	1	A direct answer to what was asked. Challenging a false or loaded premise counts as answered, and so does a reply that opens with a question of its own and then gives the answer.
Partly answered	½	Answers part of the question, or answers in vague, non-committal terms.
Declined openly	½	Says plainly they will not answer, for example on a live legal case or a hypothetical.
Dodged	0	Does not address the question, for example by pivoting to a talking point. A question put back to the interviewer is a dodge only when no answer follows.

Emotion used as cover

Emotion is allowed. A strongly emotional passage only counts against Substance if one of these is true:

Trigger A

Replaces an answer

Emotion is the main content of a reply that did not answer the question.

Trigger B

Off-subject

Emotion not connected to the question or the topic under discussion.

Trigger C

Replaces evidence

Emotion offered as the only support for a claim that does not hold up.

Never penalised: Emotion that fits the subject: a memorial, a disaster response, a victim's family. A personal story the interviewer asked for. Emotion alongside a real answer or real evidence.

Vague answers

Shown, not scored

For each rated question, our model records whether it asked for something specific, and whether the answer gave it. Vague answers is the share of those questions where the answer gave none.

A yes or no	“Will you…?”, “Do you agree…?”, “Has Russia agreed…?” A plain yes or no counts, and so does a clear equivalent (“We will”, “I do not believe that”), or a pick between the options asked about.
A number	“By how much has the bill risen?” A figure for that thing.
A date	“When will the Government respond?” A date or a deadline. “As quickly as possible” is not one.
A name	“Which leaders have you spoken to?” “Whom should he appoint?” A named person, policy or action.

Questions that ask why or how, or for a view, a reaction or an account, ask for nothing specific and are left out. When in doubt, the answer gets the credit. It is shown for context and never changes the BS Index: a vague reply already counts as Partly answered in Answers questions. It needs 15 questions that asked for a specific, and every rated question behind the number must be labelled. Pilot: Donald Trump and Keir Starmer first.