Consensus

Why it was paused

The scraper behind these pages was sound. It queried Covers.com for real contest picks, contained no invented numbers, and stopped with an error rather than filling in blanks when a feed was unreachable. That part we would stand behind.

The presentation was not sound. Each pick carried a count badge — “8x”, “6x” and so on — that readers would reasonably take to mean the number of handicappers on that side. In fact a scraped public-money percentage was being bucketed into an integer and added to the same counter. A card reading “8x” could mean eight handicappers agreed, or it could mean public money on that side was above 75%. Those are different claims and they were being presented identically.

A second problem sat alongside it: a separate script had been overwriting the daily board with a frozen slate of 100 picks captured in early March, restamped with the current date. Twenty-five different dates ended up carrying byte-identical pick data, including boards published in July showing basketball and hockey games.

What happens now

The archived boards below are left in place and unchanged. We have not deleted them, hidden them from search engines or redirected them, because quietly erasing what was published is not a correction. Read them knowing the count badge means what is described above and not what it appears to mean.

Consensus returns only if the count can be made to mean one unambiguous thing. If it cannot, it does not come back.

Something more useful in the meantime

A consensus figure tells you where money went, not whether the people behind it are any good. Trust My Record tracks individual handicappers with picks timestamped before the game and graded automatically afterwards, which answers the question consensus never could.