Why this lab exists

The problem this is trying to fix

Market commentary has a structural flaw: being right is remembered, being wrong is forgotten, and nobody keeps the books. The same person can look prescient year after year, because the misses are never dug up.

So the idea here is simple — lock a judgement the moment it is written, along with the date it settles and the data that decides it. When the date arrives, check the answer. The ones that hold stay; the ones that are overturned go to the very top. Not out of unusual honesty, but because without doing this I cannot tell which of my own judgements are actually worth anything.

What is here

Research notes are the bulk of it — usually starting from an episode or a report, then recomputing the claim against my own data to see whether it holds. The ledger holds calls that are locked and awaiting settlement (Chinese only for now). The lab has small interactive tools that turn a concept into something you can put your own numbers into. Projects collects code written for my own use that turned out to be useful to others.

What is not here, stated plainly: no stock picks, no price targets, no individual recommendations. This is about how to judge, not what to buy.

Who writes this

I am Bocky, an individual investor with an engineering background. The most useful thing that training left me is not any particular technique — it is a habit: ask first what would make this wrong. A claim that cannot say how it would fail usually cannot say why it holds either.

Public work at github.com/bockybocky — code, sources and issues are all there. Read it, copy it, open a PR. A curated list lives in Projects.

What the human does, what the AI does

This site is human-AI collaboration. It does not pretend to be entirely handwritten, nor entirely machine-generated.

The split: direction, judgement and final calls are made by the human. The AI handles research legwork, code drafts, bilingual rewriting, tests and first drafts. Everything is reviewed by a human before it ships — the AI does not decide, it clears the grunt work and forces the thinking to get sharper. The English edition is a rewrite rather than a literal translation, also human-reviewed.

Say something

Every post takes comments. If you think a call is wrong, send a counter-example — the challenge page (Chinese only for now) takes them directly. Being overturned is a good outcome here; it is the entire point of the place.

Licensing

Written content is licensed under CC BY-NC-SA 4.0: share and adapt freely with attribution, non-commercially, and share alike. Code is MIT licensed — take it.

Disclaimer

All investment-related content here is educational and methodological. It is not investment advice, contains no price targets, and does not address any current position. Investing carries risk; make your own decisions or consult a qualified professional.