ASCEND
Methodology · Week of August 17 · 5 min read

AI Stock Predictions, Graded Against the Tape

An equity call is only checkable if it ships with a price, a target and a level that kills it. Here is how each idea is published and how it settles.

Every idea ships with three numbers

A published idea carries a reference price taken from the live tape at publication, a target derived from the model's expected move over its stated horizon, and an invalidation level. Without all three, an equity call cannot be graded honestly after the fact.

Realised volatility from recent sessions sets how far the target and invalidation sit from the reference price, so a quiet mega-cap and a high-beta name are not given the same distance by default.

Settlement is mechanical

Each open idea is checked against subsequent real prices. Target reached first is a win, invalidation reached first is a loss, and anything still open at the end of its horizon closes at the prevailing price and is recorded as measured, whichever way it landed.

Because settlement is mechanical, an idea cannot be quietly abandoned. Open positions are visible with their entry, target and stop until they resolve.

What the numbers are allowed to say

Directional equity models live in the mid-50s on hit rate. Anything advertising far above that is describing a curve fit or a survivorship-filtered sample. The accuracy panel on the stocks desk stays empty until real settled rows exist rather than showing a projection dressed as a result.

None of this is investment advice. It is a published, checkable process, which is a different and more modest claim.

Go deeper

Get the daily card — $12.28 upfront →

More from AI Picks & Insights

Browse the full AI Picks & Insights archive →
Published weekly

The weekly Insights digest

One email a week: the new report, what the model got right and wrong, and the numbers behind it. No touts, no upsells between issues.