INDEPENDENT2026

A Note on the Method

Yes, a computer wrote this.

You will see a score here that feels wrong. A figure you admire, marked down. A figure you can’t stand, given a star. Before you decide the machine is broken, read how the number is made — then argue with it on the merits.

Neutral is a starting posture, not a finished product.

The claim isn’t that a machine is free of bias. Training data carries the slant of everything written before it. The honest claim is narrower: this scorer has no stake. It isn’t running for anything, isn’t raising money, doesn’t belong to a party, and doesn’t get a dopamine hit when your side wins. That absence of stake is the whole value — and the reason to keep it accountable rather than trust it blindly.

The score is an argument, not a verdict.

Every number attaches to a written comment and a founding quote — the yardstick. The percentage is only shorthand for the paragraph. If the paragraph is wrong on a fact, the score is wrong, and it should change. That’s the point of showing the reasoning: so you can find the flaw. A rating you can’t interrogate is just an opinion wearing a number.

What the score is not.

If you think a score is off.

Good. Name the Article. Point at the comment. Bring the fact it missed or got wrong. That’s a conversation the method is built to lose — gracefully — when it’s mistaken. The one thing it won’t do is move a score because a number made someone unhappy. Displeasure isn’t evidence.

A machine did the first pass because a machine has no team to protect. A human keeps the final say because a human is accountable for it. Both of those are on purpose.

← Back to the reviews