People who write predictions get scored. People who say them don't.

Half the archive can be settled and half cannot. 63 of 126 records carry a deadline; the rest name no date at all, so there is no moment at which anyone can say whether they came true.

The split does not fall where you would expect. It is not the careful against the careless. It is writing against speaking.

The medium does the work

A written prediction is usually made in a form that demands a date — a year-ahead post, a newsletter, a numbered list. The form supplies the deadline whether or not the author was thinking about being scored.

A spoken one has no such pressure. On a podcast nobody stops the conversation to ask which quarter you meant, and the sentence moves on. Jakob Nielsen, who writes his, has a deadline on six of six. Elon Musk, who mostly says his out loud, has one on one of seven.

Why it matters

The people quoted most often are frequently the least scoreable, because they talk more than they publish. A track record assembled from what someone said is therefore not a record of how often they were right — it is a record of how often they happened to be specific.

Nothing here fixes that. What it does is show it: a claim with no deadline sits permanently unresolved, which is the honest answer rather than a missing one.