What Are the Odds

Sooner or later every conversation about this lands on the same question. What are the odds? It is the right question, and also the most dangerous one, because a large improbability is the most abused number in the long history of scripture numerology. So I want to take it in the right order. First how the figure is built. Then, and this matters more, how to read it.

How the number is built

Of the thirteen letter groups, eight are divisible by nineteen with no counting choices to make at all. Treat each one as roughly a one in nineteen event, and eight of them coming up together is already on the order of ten billion to one. The remaining five groups involve choices about what to include. By the project's own count there were on the order of three hundred million possible configurations of those choices, and four structural constraints cut that field down to sixty four survivors. Put the independent pieces side by side and the project's estimate for the whole structure arising by accident sits somewhere around one in a trillion.

Why that number is not, by itself, a proof

Now the hard part, the part the numerologists always skip. A probability is only as honest as the rules that were fixed before anyone looked. If I am allowed to choose what to count, and how to count it, and when to stop counting, all after seeing the text, then I can manufacture a one in a trillion out of almost anything. The whole graveyard of miracle numerologies that came before this one was built exactly that way: wander the data, try scheme after scheme until one pays off, then report the odds of the winner as though it had been called in advance. The improbability is real and it is meaningless, because the search that found it was never constrained.

What makes an improbability mean something

A long shot earns its weight only when the procedure that produced it was fixed ahead of time and disclosed in full, so anyone can confirm there was no fishing. That is the entire reason this project publishes its counting rules before its results, ranks every judgment call by how contested it is, and ships code that reproduces the totals on any machine. The number is downstream of the discipline. Without the discipline it is theatre. With it, the number becomes a fair question rather than a trick. The rules are laid out here, and the code is here.

Where I land

So I will not ask you to believe the structure is meaningful because the odds are long. Long odds are where the cons begin, not where the truth ends. I would rather you check that the rules were fixed before the result, check that the code runs, check that the totals are what they claim to be, and only then decide for yourself what a procedure that clean, returning a result that unlikely, is actually worth. The odds are the last thing to look at, not the first.

That is the rule the whole project is built on, turned back on its own most impressive number. Trust, but verify. Don't believe me. Count.

Next in the seriesThe Discovery Framework →