Published forecasts

Every forecast we have published, in the order we published it. Each row carries a hash of itself and the row before, so changing a probability after a match has been played would break every hash after it. That is what stops us quietly revising a number and claiming we always said it — and you do not have to take our word for the check, it runs when you load this page.

The published forecasts for Philadelphia Union v CF Montreal

These are the forecasts we published for this fixture, hash-chained so a number we wrote before kick-off cannot be quietly changed afterwards.

← Back to the full record

Chain intact across all 2034 published forecasts.

Daily seals

At the end of each day we write that day's final hash to separate storage. The chain alone would catch a single altered forecast, but not a wholesale rebuild from edited rows — evidence kept in the same place as the thing it attests to proves less than it appears to. Note a seal today and you can hold us to it later.

#PublishedMatchH / D / AResultHash
10482026-09-04 02:06Philadelphia Union vCF Montreal65% / 20% / 15%our call landedwe called home 65% · home win44a93c2f6744…
7782026-09-03 05:06Philadelphia Union vCF Montreal69% / 19% / 12%our call landedwe called home 69% · home win2f4fae756b1f…
6192026-09-02 12:09Philadelphia Union vCF Montreal69% / 19% / 12%our call landedwe called home 69% · home winb7e0ae6d4b49…

The result column pairs the outcome we called (our most-likely of home / draw / away, with its probability) against what actually happened — a verdict, not a score: a 60% call missing is not a failure. Whether the probabilities are any good is a separate question, answered on the accuracy page. This page only shows that we have not changed our mind about them afterwards.