About this project
How the reconstruction is built — and how far to trust it.
How the reconstructed portfolio is built
This site estimates a portfolio for @aleabitoreddit by reading public posts chronologically and inferring position changes. It is an estimate, not disclosed holdings.
Inputs
We read only public posts. No private data, no brokerage data, and no contact with the account. Posts are raw INPUT; the site never republishes their text.
From posts to positions
For each post, a language model classifies what it implies for each ticker: OPEN, ADD, TRIM, CLOSE, REAFFIRM, or just a MENTION. Most posts are mentions or reaffirmations, not buys or sells. A deterministic ledger then folds these events in time order into a running set of approximate holdings.
Sizing is inferred, not stated
Dollar sizes are almost never disclosed, so weight buckets (CORE, MEDIUM, STARTER, WATCH) are qualitative inferences from conviction tone, emphasis, and how often a name recurs — not money.
Decay
When a name stops appearing, its estimated conviction decays toward WATCH and its confidence badge drops. Silence is treated as fading conviction, which may not be true.
Accuracy & limitations
What this reconstruction can and cannot tell you, stated plainly.
Current coverage
As of 2026-06-16, the reconstruction has processed 1231 posts into 84 tracked names (76 currently estimated open).
Known sources of error
Misclassified actions, names mentioned as customers/suppliers being mistaken for holdings, missing size information, and time lag between a real trade and a post about it. Each position carries a confidence badge (LOW/MEDIUM/HIGH) reflecting how much evidence supports it.
Not advice
Nothing here is investment advice or a recommendation, and the site is not affiliated with the account it describes.