Methodology

Only real predictions.

A claim makes the board only if it is a genuine forecast about the future. It has to name a specific outcome you can check, so it can be proven right or wrong. It does not need a deadline: some calls name one and we hold them to it, others simply predict that something will happen, and we settle those when it does. Either way it is settled by a trusted, independent source, never on our word, and judged by the same formula as everyone else's.

Most of what someone says doesn't qualify. Opinions, hot takes, and vague hedges never get scored. We only count specific claims we can check against the public record.

The share of settled calls that came true.

Every settled prediction is either right or wrong. The score is the share you got right — right calls divided by settled calls. No weights, no adjustments, no fine print. Math you can check on your fingers.

One bet, one grade. Repeating a claim fifty times doesn't count fifty times: every restatement of the same bet collapses into one call, graded once on the final position held before the outcome was knowable. The call is dated from the first time it was said on the record.

How early a call was made — and what the betting markets thought at the time — is shown on every call, so you can judge how hard it was. It doesn't change the math. If difficulty ever enters the score, it ships as a new, numbered version of this formula, never a quiet edit.

Two of Hasan Piker's calls.

Both are real, both are settled, and both are scored the way every call is.

One right out of two settled is 50. That the shutdown call was made ten months out is right there on the call — it just doesn't move the number. A real score runs this same count across a forecaster's whole settled record.

The rules that keep it honest.

1. Two sources, or it doesn't count.One dated source for the claim, and a separate, trusted source for what actually happened. One source alone never puts a call on the board.
2. No credit after the fact.A call is judged on exactly what it predicted. If something similar happens later, that's a new call. We never use it to rescue a miss or to downgrade a hit.
3. No score until the record is there.We publish an overall score only once someone has ten settled calls. Before that you see the calls themselves, but no number.
4. Blank stays blank.If there's nothing sourced to score, we leave it empty. We don't pad it, guess it, or round up to look complete.