AI Stocks FAQ
What the equity engine actually does, how its edge is defined, and how the accuracy number on the desk is produced. Everything below reflects the published model — the factor stack, the settled-only grading and the posted price list.
Questions members ask first
What is the AI Stocks engine?
ASCEND-EQ 1.0 scores a screened universe of 30 US large caps, megacap technology names and liquid ETFs across 8 weighted factors, converts the composite into a calibrated probability, and publishes an idea only when that probability is materially better than what the tape already prices in.
How is the model's edge defined?
Edge is the gap between the model's calibrated probability of the move happening at the published invalidation distance and the probability implied by the current price and volatility. No gap, no publication — the engine would rather post fewer ideas than pad the board.
How is AI Stocks accuracy measured?
From settled ideas only. Each published idea carries an entry, a target and an invalidation level; when live prices resolve one of those, the idea is graded automatically and the accuracy figure on the desk moves. Open ideas are never counted, and nothing is back-dated.
What accuracy should I expect?
Blended validated accuracy sits in the mid-50s against roughly a 38.5% break-even at the published invalidation distance. Any equity model advertising 80% is showing you a curve fit — we publish the real figure with the sample size next to it.
Which factors does the model weigh?
Price Momentum (20/60d), Earnings Revision Breadth, Valuation vs Sector, Institutional Flow, Options Skew & Gamma, Macro Regime Sensitivity, Sector Breadth, Positioning & Sentiment. Weights are fixed for a model version, so two ideas from the same session are directly comparable.
Where do the prices come from?
Live market quotes, not a static seed file. The market board and every idea's entry, move and invalidation distance are priced off the current tape, and the board refreshes on its own about once a minute while a session is open.
How often is the board published?
Once per session, before the US open: six conviction calls, four swing setups and three volatility reads. Settlement then runs continuously through the session and after the close.
Does it cover options, futures or penny stocks?
No. Names must clear liquidity and spread floors so a published entry and invalidation are realistically fillable. That rules out thin small caps, and the engine publishes directional equity and ETF ideas rather than options structures.
What does access cost?
$50 for a month, $135 for a quarter or $480 for a year, paid upfront with no auto-renew. When a term ends, access ends unless you renew.
Is this financial advice?
No. Ascend Exponentially publishes quantitative research and is not a registered investment adviser or a broker. Every idea can lose. Size positions to a loss you can absorb.
Go deeper
Ascend publishes quantitative research, not investment advice. Past measured accuracy does not guarantee future results, and every published idea can lose.