The morning sweep
Every weekday, before the opening bell, we compare about sixteen hundred liquid US-listed names against a reference set of past chart shapes and write down the two hundred that look most like them. This page is that list, in alphabetical order, published before the session it covers and never revised afterwards. It is a place to start looking, not a list of recommendations.
Educational and entertainment statistics only. Not investment advice. This is a research model showing its work, for learning and paper trading. It does not know you or your situation, and nothing here is a recommendation to buy or sell.
How to read this
This is a list of names to look at, not names to buy. Two hundred out of about sixteen hundred is roughly one name in eight — a wide net, not a shortlist. Most of these will do nothing at all. None of them are positions, and nothing here is being scored, watched or gated on your behalf.
The order is alphabetical, and that is a decision rather than a formatting choice. The comparison does produce an internal order, and we do not publish it. The first reason is not secrecy: when we tested whether that order carries information, the result did not reach a standard we are willing to publish a ranking on. Publishing one anyway would assert a claim our own evidence does not support — the mistake we documented and withdrew on the method note. A name’s position on this page means nothing.
Everything here was decided before the market opened. The list reads bars that closed the previous session — that cutoff is printed above — and it is written down before the opening bell, so no part of it can have been chosen with the benefit of the day’s trading. The timestamp is the point of the page. A list made in advance and dated is the only kind that can be scored honestly afterwards, and we intend to score it in public.
There are no numbers next to the names, and there will not be. The readings behind the sweep are the private part of the model. Publishing them, or the order they produce, would hand over the model itself rather than its output. What we can publish honestly is who made the list, the moment it was fixed, and — once enough sessions have passed — how the whole group behaved.
Market data comes from a single exchange’s feed rather than the consolidated tape. That is consistent from day to day, which is what matters for comparing names against each other, but it is not the whole market.