Case studies.
How a thesis was built and how it actually resolved — specific setups, historical moves and multi-year themes, written after the outcome is known and scored against what was said at the time.
Eight studies scoped, none finished.
A case study is written backwards from a known outcome, which is exactly why it takes longer than the thesis it examines. None of these are published yet.
One study, part by part.
A case study is assembled backwards from a known outcome, which is why it is the slowest format here and the only one that can score itself honestly.
The call as published
Quoted with its date, exactly as it went out. No hindsight edits, no quietly improved reasoning.
What actually happened
The price path, marked — including the part where it went the wrong way first.
The mechanism
Flows, an earnings surprise, a policy change. What caused the move, rather than a tidy rule reverse-engineered from it.
The score
Was the reasoning right, was the outcome lucky, or both. These are separate findings and the study says which.
What changes
What would be done differently next time — and, just as explicitly, what would not.
Written after the outcome, including the failures.
Every category here is written to the same four rules — a trigger, an invalidation, a size frame and a review. These three are specific to this one.
No hindsight edits
The original call is quoted as it was published, not as it should have been written. A case study that improves its own past reasoning teaches nothing.
The failures are included
A study of only the trades that worked is a highlight reel. The ones that invalidated are the ones that carry the lesson, and they get the same word count.
The mechanism, not the moral
Each study names what actually caused the move — flows, an earnings surprise, a policy change — rather than reverse-engineering a tidy rule out of one outcome.